gint/src/tmu/tmu.c

475 lines
11 KiB
C
Raw Normal View History

//---
// gint:tmu - Timer operation
//---
#include <gint/timer.h>
#include <gint/drivers.h>
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
#include <gint/drivers/states.h>
#include <gint/clock.h>
#include <gint/intc.h>
#include <gint/mpu/tmu.h>
#include <stdarg.h>
#undef timer_setup
//---
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
// Driver storage
//---
/* inth_data_t - data storage inside interrupt handlers */
typedef struct
{
void *function; /* User-provided callback */
uint32_t arg; /* Argument for function */
volatile void *TCR; /* TCR address for TMU */
} GPACKED(4) inth_data_t;
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
/* This array references the storage areas of all timer handlers */
static inth_data_t *timers[9] = { NULL };
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
/* Arrays of standard and extra timers */
static tmu_t *TMU = SH7305_TMU.TMU;
static etmu_t *ETMU = SH7305_ETMU;
/* TSTR register for standard timers */
static volatile uint8_t *TSTR = &SH7305_TMU.TSTR;
//---
// Local functions
//---
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
/* conf(): Configure a fixed timer */
static void conf(int id, uint32_t delay, int clock, void *f, uint32_t arg)
{
if(id < 3)
{
/* Refuse to setup timers that are already in use */
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
tmu_t *T = &TMU[id];
if(T->TCR.UNIE || *TSTR & (1 << id)) return;
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
/* Configure the counter, clear interrupt flag*/
T->TCOR = delay;
T->TCNT = delay;
T->TCR.TPSC = clock;
do T->TCR.UNF = 0;
while(T->TCR.UNF);
/* Enable interrupt and count on rising edge (SH7705) */
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
T->TCR.UNIE = 1;
T->TCR.CKEG = 0;
}
else
{
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
etmu_t *T = &ETMU[id-3];
if(T->TCR.UNIE) return;
/* No clock input and clock edge here. But TCR and TCNT need
some time to execute the write */
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
do T->TCR.UNF = 0;
while(T->TCR.UNF);
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
do T->TCOR = delay;
while(T->TCOR != delay);
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
do T->TCNT = delay;
while(T->TCNT != delay);
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
T->TCR.UNIE = 1;
}
timers[id]->function = f;
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
timers[id]->arg = arg;
}
/* matches(): Check if a timer matches the provided specification and delay */
static int matches(int id, int spec, uint32_t delay)
{
/* A specific idea only matches the exact timer */
if(spec >= 0) return id == (spec & 0xf);
/* TIMER_ANY always matches ETMU only for delays at least 100 µs */
if(spec == TIMER_ANY) return (id < 3 || delay >= 100);
/* TIMER_TMU and TIMER_ETMU match as you'd expect */
if(spec == TIMER_TMU) return (id < 3);
if(spec == TIMER_ETMU) return (id >= 3);
/* Default is not matching */
return 0;
}
/* available(): Check if a timer is available (UNIE cleared, not running) */
static int available(int id)
{
/* The timer should also be installed... */
if(!timers[id]) return 0;
if(id < 3)
{
tmu_t *T = &TMU[id];
return !T->TCR.UNIE && !(*TSTR & (1 << id));
}
else
{
etmu_t *T = &ETMU[id-3];
return !T->TCR.UNIE && !T->TSTR;
}
}
/* stop_callback(): Empty callback that stops the timer */
static int stop_callback(void)
{
return TIMER_STOP;
}
//---
// Timer API
//---
/* timer_setup(): Reserve and configure a timer */
int timer_setup(int spec, uint64_t delay, timer_callback_t function, ...)
{
int clock = 0;
/* Get the optional argument */
va_list va;
va_start(va, function);
uint32_t arg = va_arg(va, uint32_t);
va_end(va);
/* Default value for the callback */
if(!function.v) function.v = stop_callback;
/* Find a matching timer, starting from the slowest timers with the
smallest interrupt priorities all the way up to TMU0 */
for(int id = timer_count() - 1; id >= 0; id--)
{
if(!matches(id, spec, delay) || !available(id)) continue;
/* If ID is a TMU, choose a timer prescaler. Assuming the worst
running Pphi of ~48 MHz, select the finest resolution that
allows the requested delay to be represented. */
if(id < 3 && spec >= 0)
{
/* Explicit timers with clock in the specification */
clock = (spec >> 4) & 0xf;
}
else if(id < 3)
{
uint64_t sec = 1000000;
/* Pphi/4 until 350 seconds */
if(delay <= 350 * sec) clock = TIMER_Pphi_4;
/* Pphi/16 until 1430 seconds */
else if(delay <= 1430 * sec) clock = TIMER_Pphi_16;
/* Pphi/64 until 5720 seconds */
else if(delay <= 5720 * sec) clock = TIMER_Pphi_64;
/* Pphi/256 otherwise */
else clock = TIMER_Pphi_256;
}
/* Find the delay constant for that timer and clock */
if(spec < 0) delay = timer_delay(id, delay, clock);
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
conf(id, delay, clock, function.v, arg);
return id;
}
return -1;
}
/* timer_delay() - compute a delay constant from a duration in seconds */
uint32_t timer_delay(int id, uint64_t delay_us, int clock)
{
uint64_t freq;
if(id < 3)
{
const clock_frequency_t *cpg = clock_freq();
freq = cpg->Pphi_f;
if(clock == TIMER_Pphi_4) freq >>= 2;
if(clock == TIMER_Pphi_16) freq >>= 4;
if(clock == TIMER_Pphi_64) freq >>= 6;
if(clock == TIMER_Pphi_256) freq >>= 8;
}
else
{
/* ETMU all run on TCLK at 32768 Hz */
freq = 32768;
}
/* fxcg50: Calculated = 29491200 but it's too low */
/* TODO: Account for down spread spectrum in the CPG */
// uint64_t freq = 29020000 >> 2;
uint64_t product = freq * delay_us;
return product / 1000000;
}
/* timer_control() - start or stop a timer
@id Timer ID to configure
@state 0 to start the timer, 1 to stop it (nothing else!) */
static void timer_control(int id, int state)
{
if(id < 3) *TSTR = (*TSTR | (1 << id)) ^ (state << id);
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
else ETMU[id-3].TSTR = state ^ 1;
}
/* timer_start() - start a configured timer */
void timer_start(int id)
{
timer_control(id, 0);
}
/* timer_reload() - change a timer's delay constant for next interrupts */
void timer_reload(int id, uint32_t delay)
{
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
if(id < 3) TMU[id].TCOR = delay;
else ETMU[id-3].TCOR = delay;
}
/* timer_pause() - stop a running timer */
void timer_pause(int id)
{
timer_control(id, 1);
}
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
/* timer_stop() - stop and free a timer */
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
void timer_stop(int id)
{
/* Stop the timer and disable UNIE to indicate that it's free */
timer_pause(id);
if(id < 3)
{
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
TMU[id].TCR.UNIE = 0;
TMU[id].TCOR = 0xffffffff;
TMU[id].TCNT = 0xffffffff;
}
else
{
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
etmu_t *T = &ETMU[id-3];
T->TCR.UNIE = 0;
/* Also clear TCOR and TCNT to avoid spurious interrupts */
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
do T->TCOR = 0xffffffff;
while(T->TCOR + 1);
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
do T->TCNT = 0xffffffff;
while(T->TCNT + 1);
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
do T->TCR.UNF = 0;
while(T->TCR.UNF);
}
}
/* timer_wait(): Wait for a timer to stop */
void timer_wait(int id)
{
if(id < 3)
{
tmu_t *T = &TMU[id];
/* Sleep only if an interrupt will be there to wake us up */
while(*TSTR & (1 << id)) if(T->TCR.UNIE) sleep();
}
else
{
etmu_t *T = &ETMU[id-3];
while(T->TSTR) if(T->TCR.UNIE) sleep();
}
}
/* timer_spinwait(): Actively wait for a timer to raise UNF */
void timer_spinwait(int id)
{
if(id < 3)
{
tmu_t *T = &TMU[id];
while(!T->TCR.UNF) {}
}
else
{
etmu_t *T = &ETMU[id-3];
while(!T->TCR.UNF) {}
}
}
//---
// Predefined timer callbacks
//---
/* timer_timeout() - callback that sets a flag and halts the timer */
int timer_timeout(void volatile *arg)
{
int volatile *x = arg;
if(x) (*x)++;
return TIMER_STOP;
}
//---
// Driver initialization
//---
/* Interrupt handlers for standard timers (4 gates) */
extern void inth_tmu(void);
/* Interrupt handlers for extra timers */
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
extern void inth_etmu4(void);
extern void inth_etmux(void);
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
static void constructor(void)
{
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
if(isSH3())
{
TMU = SH7705_TMU.TMU;
ETMU = SH7705_ETMU;
TSTR = &SH7705_TMU.TSTR;
}
}
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
static void configure(void)
{
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
uint16_t etmu_event[6] = { 0x9e0, 0xc20, 0xc40, 0x900, 0xd00, 0xfa0 };
*TSTR = 0;
/* Install the standard TMU's interrupt handlers */
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
void *h = intc_handler(0x400, inth_tmu, 128);
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
timers[0] = h + 84;
timers[1] = h + 104;
timers[2] = h + 116;
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
/* Clear every timer to avoid surprises */
for(int id = 0; id < 3; id++)
{
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
do TMU[id].TCR.word = 0;
while(TMU[id].TCR.word);
TMU[id].TCOR = 0xffffffff;
TMU[id].TCNT = 0xffffffff;
/* Standard timers: TCR is provided to the interrupt handler */
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
timers[id]->TCR = &TMU[id].TCR;
}
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
for(int id = 0; id < timer_count()-3; id++)
{
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
etmu_t *T = &ETMU[id];
/* Extra timers seem to generate interrupts as long as TCNT=0,
regardless of TSTR. I'm not entirely sure about this weird
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
behaviour, but for safety I'll set TCOR/TCNT to non-zero. */
T->TSTR = 0;
do T->TCOR = 0xffffffff;
while(T->TCOR + 1);
/* Also TCNT and TCR take some time to record changes */
do T->TCNT = 0xffffffff;
while(T->TCNT + 1);
do T->TCR.byte = 0;
while(T->TCR.byte);
}
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
/* Install the extra timers. On SH3, only ETMU0 is available */
for(int i = 3; i < timer_count(); i++) if(i != 7)
{
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
void *h = intc_handler(etmu_event[i-3], inth_etmux, 32);
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
timers[i] = h + 20;
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
/* On SH3, the ETMU handler is not at an offset of 0x900 (event
code 0xd00) but at an offset of 0xa0 */
uint32_t *etmu_offset = h + 16;
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
if(isSH3()) *etmu_offset = *etmu_offset - 0xf40 + 0x2a0;
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
uint16_t *data_id = h + 14;
*data_id = i;
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
uint32_t *TCR = h + 28;
*TCR = (uint32_t)&ETMU[i-3].TCR;
}
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
/* Also install ETMU4, even on SH3, because it contains common code */
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
h = intc_handler(etmu_event[4], inth_etmu4, 96);
kernel: dynamic loading of GMAPPED functions to user RAM This commit introduces a large architectural change. Unlike previous models of the fx-9860G series, the G-III models have a new user RAM address different from 8801c000. The purpose of this change is to dynamically load GMAPPED functions to this address by querying the TLB, and call them through a function pointer whose address is determined when loading. Because of the overhead of using a function pointer in both assembly and C code, changes have been made to avoid GMAPPED functions altogether. Current, only cpu_setVBR() and gint_inth_callback() are left, the second being used specifically to enable TLB misses when needed. * Add a .gint.mappedrel section for the function pointers holding addresses to GMAPPED functions; add function pointers for cpu_setVBR() and gint_inth_callback() * Move rram to address 0 instead of the hardcoded 0x8801c000 * Load GMAPPED functions at their linked address + the physical address user RAM is mapped, to and compute their function pointers * Remove the GMAPPED macro since no user function needs it anymore * Add section flags "ax" (code) or "aw" (data) to every custom .section in assembler code, as they default to unpredictable values that can cause the section to be marked NOLOAD by the linker * Update the main kernel, TMU, ETMU and RTC interrupt handlers to use the new indirect calling method This is made possible by new MMU functions giving direct access to the physical area behind any virtualized page. * Add an mmu_translate() function to query the TLB * Add an mmu_uram() function to access user RAM from P1 The exception catching mechanism has been modified to avoid the use of GMAPPED functions altogether. * Set SR.BL=0 and SR.IMASK=15 before calling exception catchers * Move gint_exc_skip() to normal text ROM * Also fix registers not being popped off the stack before a panic The timer drivers have also been modified to avoid GMAPPED functions. * Invoke timer_stop() through gint_inth_callback() and move it to ROM * Move and expand the ETMU driver to span 3 blocks at 0xd00 (ETMU4) * Remove the timer_clear() function by inlining it into the ETMU handler (TCR is provided within the storage block of each timer) * Also split src/timer/inth.s into src/timer/inth-{tmu,etmu}.s Additionally, VBR addresses are now determined at runtime to further reduce hardcoded memory layout addresses in the linker script. * Determine fx-9860G VBR addresses dynamically from mmu_uram() * Determine fx-CG 50 VBR addresses dynamically from mmu_uram() * Remove linker symbols for VBR addresses Comments and documentation have been updated throughout the code to reflect the changes.
2020-09-17 14:48:54 +02:00
timers[7] = h + 84;
*(uint32_t *)(h + 92) = (uint32_t)&ETMU[4].TCR;
/* Enable TMU0 at level 13, TMU1 at level 11, TMU2 at level 9 */
intc_priority(INTC_TMU_TUNI0, 13);
intc_priority(INTC_TMU_TUNI1, 11);
intc_priority(INTC_TMU_TUNI2, 9);
/* Enable the extra TMUs at level 7 */
intc_priority(INTC_ETMU_TUNI0, 7);
if(isSH4())
{
intc_priority(INTC_ETMU_TUNI1, 7);
intc_priority(INTC_ETMU_TUNI2, 7);
intc_priority(INTC_ETMU_TUNI3, 7);
intc_priority(INTC_ETMU_TUNI4, 7);
intc_priority(INTC_ETMU_TUNI5, 7);
}
}
//---
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
// State and driver metadata
//---
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
static void hsave(tmu_state_t *s)
{
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
s->TSTR = *TSTR;
for(int i = 0; i < 3; i++)
{
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
s->t[i].TCOR = TMU[i].TCOR;
s->t[i].TCNT = TMU[i].TCNT;
s->t[i].TCR = TMU[i].TCR.word;
}
for(int i = 3; i < timer_count(); i++)
{
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
struct tmu_state_stored_timer *c = &s->t[i];
etmu_t *T = &ETMU[i-3];
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
/* Don't snapshot an interrupt state, because the timer state
is sometimes garbage protected by a masked interrupt. */
c->TCOR = T->TCOR ? T->TCOR : 0xffffffff;
c->TCNT = T->TCNT ? T->TCNT : c->TCOR;
c->TCR = T->TCR.byte & 0xd;
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
c->TSTR = T->TSTR;
}
}
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
static void hrestore(tmu_state_t const *s)
{
*TSTR = 0;
for(int i = 0; i < 3; i++)
{
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
TMU[i].TCOR = s->t[i].TCOR;
TMU[i].TCNT = s->t[i].TCNT;
TMU[i].TCR.word = s->t[i].TCR;
}
for(int i = 3; i < timer_count(); i++)
{
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
struct tmu_state_stored_timer const *c = &s->t[i];
etmu_t *T = &ETMU[i-3];
core, tmu: add gint_switch(), return to menu, and improve timer code * Add the gint_switch() function which executes user-provided code from the system (CASIOWIN) context. * Added interrupt masks to the core context (should have been there long ago). * Added the gint_osmenu() function that switches out of gint to invoke GetKeyWait() and inject KEY_CTRL_MENU to trigger the main menu. This uses many CASIOWIN syscalls, but we don't care because gint is unloaded. Trickery is used to catch the key following the return in the add-in and/or display a new application frame before GetKeyWait() even finishes after coming back. This is only available on fx9860g for now. * Removed any public syscall definition to clear up interfaces. * Patched the DMA interruption problem in a weird way on fxcg50, a driver function will be used to do that properly eventually. * Changed the driver model to save driver contexts in preallocated spaces instead of on the stack for overall less risk. * Enabled return-to-menu with the MENU key on fx9860g in getkey(). * Changed the keyboard driver to emit releases before presses, as a return-to-menu acts as a press+release of different keys in a single driver frame, which confuses getkey(). * Fixed a really stupid bug in memcpy() that made the function really not work. Improvements in the timer driver: * Expose ETMU modules as SH7705_TMU and SH7305_TMU in <gint/mpu/tmu.h>. * Remove the timer_t structures, using SH*_ETMU and SH*_TMU instead. Only interrupt gate entries are left hardcoded. * Discovered that not only every write to the TCNT or TCR of an ETMU takes about 1/32k of a second (hinting at registers being powered by the same clock as the timer), but every write occuring while a previous write is pending is *lost*. This led to terrible bugs when switching ETMU contexts too fast in gint_switch(). * Removed an internal timer_address() function. * Overall simplified the handling of timers and the initialization step.
2020-05-10 14:03:41 +02:00
do T->TCOR = c->TCOR;
while(T->TCOR != c->TCOR);
T->TSTR = c->TSTR;
do T->TCNT = c->TCNT;
while(T->TCNT != c->TCNT);
do T->TCR.byte = c->TCR;
while(T->TCR.byte != c->TCR);
}
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
*TSTR = s->TSTR;
}
gint_driver_t drv_tmu = {
.name = "TMU",
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
.constructor = constructor,
.configure = configure,
.hsave = (void *)hsave,
.hrestore = (void *)hrestore,
.state_size = sizeof(tmu_state_t),
};
kernel: driver and world system overhaul Changes in the driver and world system: * Rewrite driver logic to include more advanced concepts. The notion of binding a driver to a device is introduced to formalize wait(); power management is now built-in instead of being handled by the drivers (for instance DMA). The new driver model is described in great detail in <gint/drivers.h> * Formalized the concept of "world switch" where the hardware state is saved and later restored. As a tool, the world switch turns out to be very stable, and allows a lot of hardware manipulation that would be edgy at best when running in the OS world. * Added a GINT_DRV_SHARED flag for drivers to specify that their state is shared between worlds and not saved/restored. This has a couple of uses. * Exposed a lot more of the internal driver/world system as their is no particular downside to it. This includes stuff in <gint/drivers.h> and the driver's state structures in <gint/drivers/states.h>. This is useful for debugging and for cracked concepts, but there is no API stability guarantee. * Added a more flexible driver level system that allows any 2-digit level to be used. Feature changes: * Added a CPU driver that provides the VBR change as its state save. Because the whole context switch relied on interrupts being disabled anyway, there is no longer an inversion of control when setting the VBR; this is just part of the CPU driver's configuration. The CPU driver may also support other features such as XYRAM block transfer in the future. * Moved gint_inthandler() to the INTC driver under the name intc_handler(), pairing up again with intc_priority(). * Added a reentrant atomic lock based on the test-and-set primitive. Interrupts are disabled with IMASK=15 for the duration of atomic operations. * Enabled the DMA driver on SH7305-based fx-9860G. The DMA provides little benefit on this platform because the RAM is generally faster and buffers are ultimately small. The DMA is still not available on SH3-based fx-9860G models. * Solved an extremely obnoxious bug in timer_spin_wait() where the timer is not freed, causing the callback to be called when interrupts are re-enabled. This increments a random value on the stack. As a consequence of the change, removed the long delays in the USB driver since they are not actually needed. Minor changes: * Deprecated some of the elements in <gint/hardware.h>. There really is no good way to "enumerate" devices yet. * Deprecated gint_switch() in favor of a new function gint_world_switch() which uses the GINT_CALL abstraction. * Made the fx-9860G VRAM 32-aligned so that it can be used for tests with the DMA. Some features of the driver and world systems have not been implemented yet, but may be in the future: * Some driver flags should be per-world in order to create multiple gint worlds. This would be useful in Yatis' hypervisor. * A GINT_DRV_LAZY flag would be useful for drivers that don't want to be started up automatically during a world switch. This is relevant for drivers that have a slow start/stop sequence. However, this is tricky to do correctly as it requires dynamic start/stop and also tracking which world the current hardware state belongs to.
2021-04-23 18:50:20 +02:00
GINT_DECLARE_DRIVER(13, drv_tmu);