minix/drivers/amddev/amddev.c

475 lines
10 KiB
C
Raw Normal View History

2008-02-25 13:07:19 +01:00
/*
amddev.c
Driver for the AMD Device Exclusion Vector (DEV)
*/
#define _SYSTEM
#define _MINIX
#include <minix/config.h>
#include <minix/type.h>
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sys/vm_i386.h>
2008-02-25 13:07:19 +01:00
#include <minix/com.h>
#include <minix/const.h>
#include <minix/ipc.h>
#include <minix/syslib.h>
#include <minix/sysutil.h>
#include <minix/endpoint.h>
2008-02-25 13:07:19 +01:00
#include <ibm/pci.h>
/* Offsets from capability pointer */
#define DEV_OP 4 /* Selects control/status register to access */
#define DEV_OP_FUNC_SHIFT 8 /* Function part in OP reg. */
#define DEV_DATA 8 /* Read/write to access reg. selected */
/* Functions */
#define DEVF_BASE_LO 0
#define DEVF_BASE_HI 1
#define DEVF_MAP 2
#define DEVF_CAP 3
#define DEVF_CAP_MAPS_MASK 0x00ff0000
#define DEVF_CAP_MAPS_SHIFT 16
#define DEVF_CAP_DOMS_MASK 0x0000ff00
#define DEVF_CAP_DOMS_SHIFT 8
#define DEVF_CAP_REV_MASK 0x000000ff
#define DEVF_CAP_REV_SHIFT 0
#define DEVF_CR 4
#define DEVF_ERR_STATUS 5
#define DEVF_ERR_ADDR_LO 6
#define DEVF_ERR_ADDR_HI 7
static int dev_devind;
static u8_t dev_capptr;
static u8_t *table;
static int find_dev(int *devindp, u8_t *capaddrp);
static u32_t read_reg(int function, int index);
static void write_reg(int function, int index, u32_t value);
static void init_domain(int index);
static void init_map(unsigned int ix);
2008-02-25 13:07:19 +01:00
static int do_add4pci(message *m);
static void add_range(u32_t busaddr, u32_t size);
static void del_range(u32_t busaddr, u32_t size);
2010-01-11 15:22:29 +01:00
static void do_pm_notify(message *m);
2008-02-25 13:07:19 +01:00
static void report_exceptions(void);
Basic System Event Framework (SEF) with ping and live update. SYSLIB CHANGES: - SEF must be used by every system process and is thereby part of the system library. - The framework provides a receive() interface (sef_receive) for system processes to automatically catch known system even messages and process them. - SEF provides a default behavior for each type of system event, but allows system processes to register callbacks to override the default behavior. - Custom (local to the process) or predefined (provided by SEF) callback implementations can be registered to SEF. - SEF currently includes support for 2 types of system events: 1. SEF Ping. The event occurs every time RS sends a ping to figure out whether a system process is still alive. The default callback implementation provided by SEF is to notify RS back to let it know the process is alive and kicking. 2. SEF Live update. The event occurs every time RS sends a prepare to update message to let a system process know an update is available and to prepare for it. The live update support is very basic for now. SEF only deals with verifying if the prepare state can be supported by the process, dumping the state for debugging purposes, and providing an event-driven programming model to the process to react to state changes check-in when ready to update. - SEF should be extended in the future to integrate support for more types of system events. Ideally, all the cross-cutting concerns should be integrated into SEF to avoid duplicating code and ease extensibility. Examples include: * PM notify messages primarily used at shutdown. * SYSTEM notify messages primarily used for signals. * CLOCK notify messages used for system alarms. * Debug messages. IS could still be in charge of fkey handling but would forward the debug message to the target process (e.g. PM, if the user requested debug information about PM). SEF would then catch the message and do nothing unless the process has registered an appropriate callback to deal with the event. This simplifies the programming model to print debug information, avoids duplicating code, and reduces the effort to print debug information. SYSTEM PROCESSES CHANGES: - Every system process registers SEF callbacks it needs to override the default system behavior and calls sef_startup() right after being started. - sef_startup() does almost nothing now, but will be extended in the future to support callbacks of its own to let RS control and synchronize with every system process at initialization time. - Every system process calls sef_receive() now rather than receive() directly, to let SEF handle predefined system events. RS CHANGES: - RS supports a basic single-component live update protocol now, as follows: * When an update command is issued (via "service update *"), RS notifies the target system process to prepare for a specific update state. * If the process doesn't respond back in time, the update is aborted. * When the process responds back, RS kills it and marks it for refreshing. * The process is then automatically restarted as for a buggy process and can start running again. * Live update is currently prototyped as a controlled failure.
2009-12-21 15:12:21 +01:00
/* SEF functions and variables. */
FORWARD _PROTOTYPE( void sef_local_startup, (void) );
Initialization protocol for system services. SYSLIB CHANGES: - SEF framework now supports a new SEF Init request type from RS. 3 different callbacks are available (init_fresh, init_lu, init_restart) to specify initialization code when a service starts fresh, starts after a live update, or restarts. SYSTEM SERVICE CHANGES: - Initialization code for system services is now enclosed in a callback SEF will automatically call at init time. The return code of the callback will tell RS whether the initialization completed successfully. - Each init callback can access information passed by RS to initialize. As of now, each system service has access to the public entries of RS's system process table to gather all the information required to initialize. This design eliminates many existing or potential races at boot time and provides a uniform initialization interface to system services. The same interface will be reused for the upcoming publish/subscribe model to handle dynamic registration / deregistration of system services. VM CHANGES: - Uniform privilege management for all system services. Every service uses the same call mask format. For boot services, VM copies the call mask from init data. For dynamic services, VM still receives the call mask via rs_set_priv call that will be soon replaced by the upcoming publish/subscribe model. RS CHANGES: - The system process table has been reorganized and split into private entries and public entries. Only the latter ones are exposed to system services. - VM call masks are now entirely configured in rs/table.c - RS has now its own slot in the system process table. Only kernel tasks and user processes not included in the boot image are now left out from the system process table. - RS implements the initialization protocol for system services. - For services in the boot image, RS blocks till initialization is complete and panics when failure is reported back. Services are initialized in their order of appearance in the boot image priv table and RS blocks to implements synchronous initialization for every system service having the flag SF_SYNCH_BOOT set. - For services started dynamically, the initialization protocol is implemented as though it were the first ping for the service. In this case, if the system service fails to report back (or reports failure), RS brings the service down rather than trying to restart it.
2010-01-08 02:20:42 +01:00
FORWARD _PROTOTYPE( int sef_cb_init_fresh, (int type, sef_init_info_t *info) );
Basic System Event Framework (SEF) with ping and live update. SYSLIB CHANGES: - SEF must be used by every system process and is thereby part of the system library. - The framework provides a receive() interface (sef_receive) for system processes to automatically catch known system even messages and process them. - SEF provides a default behavior for each type of system event, but allows system processes to register callbacks to override the default behavior. - Custom (local to the process) or predefined (provided by SEF) callback implementations can be registered to SEF. - SEF currently includes support for 2 types of system events: 1. SEF Ping. The event occurs every time RS sends a ping to figure out whether a system process is still alive. The default callback implementation provided by SEF is to notify RS back to let it know the process is alive and kicking. 2. SEF Live update. The event occurs every time RS sends a prepare to update message to let a system process know an update is available and to prepare for it. The live update support is very basic for now. SEF only deals with verifying if the prepare state can be supported by the process, dumping the state for debugging purposes, and providing an event-driven programming model to the process to react to state changes check-in when ready to update. - SEF should be extended in the future to integrate support for more types of system events. Ideally, all the cross-cutting concerns should be integrated into SEF to avoid duplicating code and ease extensibility. Examples include: * PM notify messages primarily used at shutdown. * SYSTEM notify messages primarily used for signals. * CLOCK notify messages used for system alarms. * Debug messages. IS could still be in charge of fkey handling but would forward the debug message to the target process (e.g. PM, if the user requested debug information about PM). SEF would then catch the message and do nothing unless the process has registered an appropriate callback to deal with the event. This simplifies the programming model to print debug information, avoids duplicating code, and reduces the effort to print debug information. SYSTEM PROCESSES CHANGES: - Every system process registers SEF callbacks it needs to override the default system behavior and calls sef_startup() right after being started. - sef_startup() does almost nothing now, but will be extended in the future to support callbacks of its own to let RS control and synchronize with every system process at initialization time. - Every system process calls sef_receive() now rather than receive() directly, to let SEF handle predefined system events. RS CHANGES: - RS supports a basic single-component live update protocol now, as follows: * When an update command is issued (via "service update *"), RS notifies the target system process to prepare for a specific update state. * If the process doesn't respond back in time, the update is aborted. * When the process responds back, RS kills it and marks it for refreshing. * The process is then automatically restarted as for a buggy process and can start running again. * Live update is currently prototyped as a controlled failure.
2009-12-21 15:12:21 +01:00
2008-02-25 13:07:19 +01:00
int main(void)
{
int r;
message m;
Basic System Event Framework (SEF) with ping and live update. SYSLIB CHANGES: - SEF must be used by every system process and is thereby part of the system library. - The framework provides a receive() interface (sef_receive) for system processes to automatically catch known system even messages and process them. - SEF provides a default behavior for each type of system event, but allows system processes to register callbacks to override the default behavior. - Custom (local to the process) or predefined (provided by SEF) callback implementations can be registered to SEF. - SEF currently includes support for 2 types of system events: 1. SEF Ping. The event occurs every time RS sends a ping to figure out whether a system process is still alive. The default callback implementation provided by SEF is to notify RS back to let it know the process is alive and kicking. 2. SEF Live update. The event occurs every time RS sends a prepare to update message to let a system process know an update is available and to prepare for it. The live update support is very basic for now. SEF only deals with verifying if the prepare state can be supported by the process, dumping the state for debugging purposes, and providing an event-driven programming model to the process to react to state changes check-in when ready to update. - SEF should be extended in the future to integrate support for more types of system events. Ideally, all the cross-cutting concerns should be integrated into SEF to avoid duplicating code and ease extensibility. Examples include: * PM notify messages primarily used at shutdown. * SYSTEM notify messages primarily used for signals. * CLOCK notify messages used for system alarms. * Debug messages. IS could still be in charge of fkey handling but would forward the debug message to the target process (e.g. PM, if the user requested debug information about PM). SEF would then catch the message and do nothing unless the process has registered an appropriate callback to deal with the event. This simplifies the programming model to print debug information, avoids duplicating code, and reduces the effort to print debug information. SYSTEM PROCESSES CHANGES: - Every system process registers SEF callbacks it needs to override the default system behavior and calls sef_startup() right after being started. - sef_startup() does almost nothing now, but will be extended in the future to support callbacks of its own to let RS control and synchronize with every system process at initialization time. - Every system process calls sef_receive() now rather than receive() directly, to let SEF handle predefined system events. RS CHANGES: - RS supports a basic single-component live update protocol now, as follows: * When an update command is issued (via "service update *"), RS notifies the target system process to prepare for a specific update state. * If the process doesn't respond back in time, the update is aborted. * When the process responds back, RS kills it and marks it for refreshing. * The process is then automatically restarted as for a buggy process and can start running again. * Live update is currently prototyped as a controlled failure.
2009-12-21 15:12:21 +01:00
/* SEF local startup. */
sef_local_startup();
2008-02-25 13:07:19 +01:00
for(;;)
{
report_exceptions();
Basic System Event Framework (SEF) with ping and live update. SYSLIB CHANGES: - SEF must be used by every system process and is thereby part of the system library. - The framework provides a receive() interface (sef_receive) for system processes to automatically catch known system even messages and process them. - SEF provides a default behavior for each type of system event, but allows system processes to register callbacks to override the default behavior. - Custom (local to the process) or predefined (provided by SEF) callback implementations can be registered to SEF. - SEF currently includes support for 2 types of system events: 1. SEF Ping. The event occurs every time RS sends a ping to figure out whether a system process is still alive. The default callback implementation provided by SEF is to notify RS back to let it know the process is alive and kicking. 2. SEF Live update. The event occurs every time RS sends a prepare to update message to let a system process know an update is available and to prepare for it. The live update support is very basic for now. SEF only deals with verifying if the prepare state can be supported by the process, dumping the state for debugging purposes, and providing an event-driven programming model to the process to react to state changes check-in when ready to update. - SEF should be extended in the future to integrate support for more types of system events. Ideally, all the cross-cutting concerns should be integrated into SEF to avoid duplicating code and ease extensibility. Examples include: * PM notify messages primarily used at shutdown. * SYSTEM notify messages primarily used for signals. * CLOCK notify messages used for system alarms. * Debug messages. IS could still be in charge of fkey handling but would forward the debug message to the target process (e.g. PM, if the user requested debug information about PM). SEF would then catch the message and do nothing unless the process has registered an appropriate callback to deal with the event. This simplifies the programming model to print debug information, avoids duplicating code, and reduces the effort to print debug information. SYSTEM PROCESSES CHANGES: - Every system process registers SEF callbacks it needs to override the default system behavior and calls sef_startup() right after being started. - sef_startup() does almost nothing now, but will be extended in the future to support callbacks of its own to let RS control and synchronize with every system process at initialization time. - Every system process calls sef_receive() now rather than receive() directly, to let SEF handle predefined system events. RS CHANGES: - RS supports a basic single-component live update protocol now, as follows: * When an update command is issued (via "service update *"), RS notifies the target system process to prepare for a specific update state. * If the process doesn't respond back in time, the update is aborted. * When the process responds back, RS kills it and marks it for refreshing. * The process is then automatically restarted as for a buggy process and can start running again. * Live update is currently prototyped as a controlled failure.
2009-12-21 15:12:21 +01:00
r= sef_receive(ANY, &m);
2008-02-25 13:07:19 +01:00
if (r != OK)
Basic System Event Framework (SEF) with ping and live update. SYSLIB CHANGES: - SEF must be used by every system process and is thereby part of the system library. - The framework provides a receive() interface (sef_receive) for system processes to automatically catch known system even messages and process them. - SEF provides a default behavior for each type of system event, but allows system processes to register callbacks to override the default behavior. - Custom (local to the process) or predefined (provided by SEF) callback implementations can be registered to SEF. - SEF currently includes support for 2 types of system events: 1. SEF Ping. The event occurs every time RS sends a ping to figure out whether a system process is still alive. The default callback implementation provided by SEF is to notify RS back to let it know the process is alive and kicking. 2. SEF Live update. The event occurs every time RS sends a prepare to update message to let a system process know an update is available and to prepare for it. The live update support is very basic for now. SEF only deals with verifying if the prepare state can be supported by the process, dumping the state for debugging purposes, and providing an event-driven programming model to the process to react to state changes check-in when ready to update. - SEF should be extended in the future to integrate support for more types of system events. Ideally, all the cross-cutting concerns should be integrated into SEF to avoid duplicating code and ease extensibility. Examples include: * PM notify messages primarily used at shutdown. * SYSTEM notify messages primarily used for signals. * CLOCK notify messages used for system alarms. * Debug messages. IS could still be in charge of fkey handling but would forward the debug message to the target process (e.g. PM, if the user requested debug information about PM). SEF would then catch the message and do nothing unless the process has registered an appropriate callback to deal with the event. This simplifies the programming model to print debug information, avoids duplicating code, and reduces the effort to print debug information. SYSTEM PROCESSES CHANGES: - Every system process registers SEF callbacks it needs to override the default system behavior and calls sef_startup() right after being started. - sef_startup() does almost nothing now, but will be extended in the future to support callbacks of its own to let RS control and synchronize with every system process at initialization time. - Every system process calls sef_receive() now rather than receive() directly, to let SEF handle predefined system events. RS CHANGES: - RS supports a basic single-component live update protocol now, as follows: * When an update command is issued (via "service update *"), RS notifies the target system process to prepare for a specific update state. * If the process doesn't respond back in time, the update is aborted. * When the process responds back, RS kills it and marks it for refreshing. * The process is then automatically restarted as for a buggy process and can start running again. * Live update is currently prototyped as a controlled failure.
2009-12-21 15:12:21 +01:00
panic(__FILE__, "sef_receive failed", r);
if (is_notify(m.m_type)) {
if (_ENDPOINT_P(m.m_source) == PM_PROC_NR) {
do_pm_notify(&m);
continue;
}
}
else if (m.m_type == IOMMU_MAP) {
2008-02-25 13:07:19 +01:00
r= do_add4pci(&m);
m.m_type= r;
send(m.m_source, &m);
continue;
}
printf("amddev: got message from %d\n", m.m_source);
}
}
Basic System Event Framework (SEF) with ping and live update. SYSLIB CHANGES: - SEF must be used by every system process and is thereby part of the system library. - The framework provides a receive() interface (sef_receive) for system processes to automatically catch known system even messages and process them. - SEF provides a default behavior for each type of system event, but allows system processes to register callbacks to override the default behavior. - Custom (local to the process) or predefined (provided by SEF) callback implementations can be registered to SEF. - SEF currently includes support for 2 types of system events: 1. SEF Ping. The event occurs every time RS sends a ping to figure out whether a system process is still alive. The default callback implementation provided by SEF is to notify RS back to let it know the process is alive and kicking. 2. SEF Live update. The event occurs every time RS sends a prepare to update message to let a system process know an update is available and to prepare for it. The live update support is very basic for now. SEF only deals with verifying if the prepare state can be supported by the process, dumping the state for debugging purposes, and providing an event-driven programming model to the process to react to state changes check-in when ready to update. - SEF should be extended in the future to integrate support for more types of system events. Ideally, all the cross-cutting concerns should be integrated into SEF to avoid duplicating code and ease extensibility. Examples include: * PM notify messages primarily used at shutdown. * SYSTEM notify messages primarily used for signals. * CLOCK notify messages used for system alarms. * Debug messages. IS could still be in charge of fkey handling but would forward the debug message to the target process (e.g. PM, if the user requested debug information about PM). SEF would then catch the message and do nothing unless the process has registered an appropriate callback to deal with the event. This simplifies the programming model to print debug information, avoids duplicating code, and reduces the effort to print debug information. SYSTEM PROCESSES CHANGES: - Every system process registers SEF callbacks it needs to override the default system behavior and calls sef_startup() right after being started. - sef_startup() does almost nothing now, but will be extended in the future to support callbacks of its own to let RS control and synchronize with every system process at initialization time. - Every system process calls sef_receive() now rather than receive() directly, to let SEF handle predefined system events. RS CHANGES: - RS supports a basic single-component live update protocol now, as follows: * When an update command is issued (via "service update *"), RS notifies the target system process to prepare for a specific update state. * If the process doesn't respond back in time, the update is aborted. * When the process responds back, RS kills it and marks it for refreshing. * The process is then automatically restarted as for a buggy process and can start running again. * Live update is currently prototyped as a controlled failure.
2009-12-21 15:12:21 +01:00
/*===========================================================================*
* sef_local_startup *
*===========================================================================*/
PRIVATE void sef_local_startup()
{
Initialization protocol for system services. SYSLIB CHANGES: - SEF framework now supports a new SEF Init request type from RS. 3 different callbacks are available (init_fresh, init_lu, init_restart) to specify initialization code when a service starts fresh, starts after a live update, or restarts. SYSTEM SERVICE CHANGES: - Initialization code for system services is now enclosed in a callback SEF will automatically call at init time. The return code of the callback will tell RS whether the initialization completed successfully. - Each init callback can access information passed by RS to initialize. As of now, each system service has access to the public entries of RS's system process table to gather all the information required to initialize. This design eliminates many existing or potential races at boot time and provides a uniform initialization interface to system services. The same interface will be reused for the upcoming publish/subscribe model to handle dynamic registration / deregistration of system services. VM CHANGES: - Uniform privilege management for all system services. Every service uses the same call mask format. For boot services, VM copies the call mask from init data. For dynamic services, VM still receives the call mask via rs_set_priv call that will be soon replaced by the upcoming publish/subscribe model. RS CHANGES: - The system process table has been reorganized and split into private entries and public entries. Only the latter ones are exposed to system services. - VM call masks are now entirely configured in rs/table.c - RS has now its own slot in the system process table. Only kernel tasks and user processes not included in the boot image are now left out from the system process table. - RS implements the initialization protocol for system services. - For services in the boot image, RS blocks till initialization is complete and panics when failure is reported back. Services are initialized in their order of appearance in the boot image priv table and RS blocks to implements synchronous initialization for every system service having the flag SF_SYNCH_BOOT set. - For services started dynamically, the initialization protocol is implemented as though it were the first ping for the service. In this case, if the system service fails to report back (or reports failure), RS brings the service down rather than trying to restart it.
2010-01-08 02:20:42 +01:00
/* Register init callbacks. */
sef_setcb_init_fresh(sef_cb_init_fresh);
sef_setcb_init_lu(sef_cb_init_fresh);
sef_setcb_init_restart(sef_cb_init_fresh);
Basic System Event Framework (SEF) with ping and live update. SYSLIB CHANGES: - SEF must be used by every system process and is thereby part of the system library. - The framework provides a receive() interface (sef_receive) for system processes to automatically catch known system even messages and process them. - SEF provides a default behavior for each type of system event, but allows system processes to register callbacks to override the default behavior. - Custom (local to the process) or predefined (provided by SEF) callback implementations can be registered to SEF. - SEF currently includes support for 2 types of system events: 1. SEF Ping. The event occurs every time RS sends a ping to figure out whether a system process is still alive. The default callback implementation provided by SEF is to notify RS back to let it know the process is alive and kicking. 2. SEF Live update. The event occurs every time RS sends a prepare to update message to let a system process know an update is available and to prepare for it. The live update support is very basic for now. SEF only deals with verifying if the prepare state can be supported by the process, dumping the state for debugging purposes, and providing an event-driven programming model to the process to react to state changes check-in when ready to update. - SEF should be extended in the future to integrate support for more types of system events. Ideally, all the cross-cutting concerns should be integrated into SEF to avoid duplicating code and ease extensibility. Examples include: * PM notify messages primarily used at shutdown. * SYSTEM notify messages primarily used for signals. * CLOCK notify messages used for system alarms. * Debug messages. IS could still be in charge of fkey handling but would forward the debug message to the target process (e.g. PM, if the user requested debug information about PM). SEF would then catch the message and do nothing unless the process has registered an appropriate callback to deal with the event. This simplifies the programming model to print debug information, avoids duplicating code, and reduces the effort to print debug information. SYSTEM PROCESSES CHANGES: - Every system process registers SEF callbacks it needs to override the default system behavior and calls sef_startup() right after being started. - sef_startup() does almost nothing now, but will be extended in the future to support callbacks of its own to let RS control and synchronize with every system process at initialization time. - Every system process calls sef_receive() now rather than receive() directly, to let SEF handle predefined system events. RS CHANGES: - RS supports a basic single-component live update protocol now, as follows: * When an update command is issued (via "service update *"), RS notifies the target system process to prepare for a specific update state. * If the process doesn't respond back in time, the update is aborted. * When the process responds back, RS kills it and marks it for refreshing. * The process is then automatically restarted as for a buggy process and can start running again. * Live update is currently prototyped as a controlled failure.
2009-12-21 15:12:21 +01:00
/* Register live update callbacks. */
sef_setcb_lu_prepare(sef_cb_lu_prepare_always_ready);
sef_setcb_lu_state_isvalid(sef_cb_lu_state_isvalid_standard);
/* Let SEF perform startup. */
sef_startup();
}
Initialization protocol for system services. SYSLIB CHANGES: - SEF framework now supports a new SEF Init request type from RS. 3 different callbacks are available (init_fresh, init_lu, init_restart) to specify initialization code when a service starts fresh, starts after a live update, or restarts. SYSTEM SERVICE CHANGES: - Initialization code for system services is now enclosed in a callback SEF will automatically call at init time. The return code of the callback will tell RS whether the initialization completed successfully. - Each init callback can access information passed by RS to initialize. As of now, each system service has access to the public entries of RS's system process table to gather all the information required to initialize. This design eliminates many existing or potential races at boot time and provides a uniform initialization interface to system services. The same interface will be reused for the upcoming publish/subscribe model to handle dynamic registration / deregistration of system services. VM CHANGES: - Uniform privilege management for all system services. Every service uses the same call mask format. For boot services, VM copies the call mask from init data. For dynamic services, VM still receives the call mask via rs_set_priv call that will be soon replaced by the upcoming publish/subscribe model. RS CHANGES: - The system process table has been reorganized and split into private entries and public entries. Only the latter ones are exposed to system services. - VM call masks are now entirely configured in rs/table.c - RS has now its own slot in the system process table. Only kernel tasks and user processes not included in the boot image are now left out from the system process table. - RS implements the initialization protocol for system services. - For services in the boot image, RS blocks till initialization is complete and panics when failure is reported back. Services are initialized in their order of appearance in the boot image priv table and RS blocks to implements synchronous initialization for every system service having the flag SF_SYNCH_BOOT set. - For services started dynamically, the initialization protocol is implemented as though it were the first ping for the service. In this case, if the system service fails to report back (or reports failure), RS brings the service down rather than trying to restart it.
2010-01-08 02:20:42 +01:00
/*===========================================================================*
* sef_cb_init_fresh *
*===========================================================================*/
PRIVATE int sef_cb_init_fresh(int type, sef_init_info_t *info)
2008-02-25 13:07:19 +01:00
{
Initialization protocol for system services. SYSLIB CHANGES: - SEF framework now supports a new SEF Init request type from RS. 3 different callbacks are available (init_fresh, init_lu, init_restart) to specify initialization code when a service starts fresh, starts after a live update, or restarts. SYSTEM SERVICE CHANGES: - Initialization code for system services is now enclosed in a callback SEF will automatically call at init time. The return code of the callback will tell RS whether the initialization completed successfully. - Each init callback can access information passed by RS to initialize. As of now, each system service has access to the public entries of RS's system process table to gather all the information required to initialize. This design eliminates many existing or potential races at boot time and provides a uniform initialization interface to system services. The same interface will be reused for the upcoming publish/subscribe model to handle dynamic registration / deregistration of system services. VM CHANGES: - Uniform privilege management for all system services. Every service uses the same call mask format. For boot services, VM copies the call mask from init data. For dynamic services, VM still receives the call mask via rs_set_priv call that will be soon replaced by the upcoming publish/subscribe model. RS CHANGES: - The system process table has been reorganized and split into private entries and public entries. Only the latter ones are exposed to system services. - VM call masks are now entirely configured in rs/table.c - RS has now its own slot in the system process table. Only kernel tasks and user processes not included in the boot image are now left out from the system process table. - RS implements the initialization protocol for system services. - For services in the boot image, RS blocks till initialization is complete and panics when failure is reported back. Services are initialized in their order of appearance in the boot image priv table and RS blocks to implements synchronous initialization for every system service having the flag SF_SYNCH_BOOT set. - For services started dynamically, the initialization protocol is implemented as though it were the first ping for the service. In this case, if the system service fails to report back (or reports failure), RS brings the service down rather than trying to restart it.
2010-01-08 02:20:42 +01:00
/* Initialize the amddev driver. */
2008-02-25 13:07:19 +01:00
int r, n_maps, n_domains, revision;
u16_t flags;
u32_t bits;
Initialization protocol for system services. SYSLIB CHANGES: - SEF framework now supports a new SEF Init request type from RS. 3 different callbacks are available (init_fresh, init_lu, init_restart) to specify initialization code when a service starts fresh, starts after a live update, or restarts. SYSTEM SERVICE CHANGES: - Initialization code for system services is now enclosed in a callback SEF will automatically call at init time. The return code of the callback will tell RS whether the initialization completed successfully. - Each init callback can access information passed by RS to initialize. As of now, each system service has access to the public entries of RS's system process table to gather all the information required to initialize. This design eliminates many existing or potential races at boot time and provides a uniform initialization interface to system services. The same interface will be reused for the upcoming publish/subscribe model to handle dynamic registration / deregistration of system services. VM CHANGES: - Uniform privilege management for all system services. Every service uses the same call mask format. For boot services, VM copies the call mask from init data. For dynamic services, VM still receives the call mask via rs_set_priv call that will be soon replaced by the upcoming publish/subscribe model. RS CHANGES: - The system process table has been reorganized and split into private entries and public entries. Only the latter ones are exposed to system services. - VM call masks are now entirely configured in rs/table.c - RS has now its own slot in the system process table. Only kernel tasks and user processes not included in the boot image are now left out from the system process table. - RS implements the initialization protocol for system services. - For services in the boot image, RS blocks till initialization is complete and panics when failure is reported back. Services are initialized in their order of appearance in the boot image priv table and RS blocks to implements synchronous initialization for every system service having the flag SF_SYNCH_BOOT set. - For services started dynamically, the initialization protocol is implemented as though it were the first ping for the service. In this case, if the system service fails to report back (or reports failure), RS brings the service down rather than trying to restart it.
2010-01-08 02:20:42 +01:00
printf("amddev: starting\n");
2008-02-25 13:07:19 +01:00
r= find_dev(&dev_devind, &dev_capptr);
if (!r)
return r;
flags= pci_attr_r16(dev_devind, dev_capptr+CAP_SD_INFO);
printf("amddev`init: flags = 0x%x\n", flags);
bits= read_reg(DEVF_CAP, 0);
n_maps= ((bits & DEVF_CAP_MAPS_MASK) >> DEVF_CAP_MAPS_SHIFT);
n_domains= ((bits & DEVF_CAP_DOMS_MASK) >> DEVF_CAP_DOMS_SHIFT);
revision= ((bits & DEVF_CAP_REV_MASK) >> DEVF_CAP_REV_SHIFT);
printf("amddev`init: DEVF_CAP = 0x%x (%d maps, %d domains, rev 0x%x)\n",
bits, n_maps, n_domains, revision);
printf("status = 0x%x, addr-lo = 0x%x, addr-hi = 0x%x\n",
read_reg(DEVF_ERR_STATUS, 0),
read_reg(DEVF_ERR_ADDR_LO, 0),
read_reg(DEVF_ERR_ADDR_HI, 0));
init_domain(0);
init_map(0);
#if 0
init_domain(1);
#endif
write_reg(DEVF_CR, 0, 0x10 | 0x8 | 0x4 | 1);
printf("after write: DEVF_CR: 0x%x\n", read_reg(DEVF_CR, 0));
Initialization protocol for system services. SYSLIB CHANGES: - SEF framework now supports a new SEF Init request type from RS. 3 different callbacks are available (init_fresh, init_lu, init_restart) to specify initialization code when a service starts fresh, starts after a live update, or restarts. SYSTEM SERVICE CHANGES: - Initialization code for system services is now enclosed in a callback SEF will automatically call at init time. The return code of the callback will tell RS whether the initialization completed successfully. - Each init callback can access information passed by RS to initialize. As of now, each system service has access to the public entries of RS's system process table to gather all the information required to initialize. This design eliminates many existing or potential races at boot time and provides a uniform initialization interface to system services. The same interface will be reused for the upcoming publish/subscribe model to handle dynamic registration / deregistration of system services. VM CHANGES: - Uniform privilege management for all system services. Every service uses the same call mask format. For boot services, VM copies the call mask from init data. For dynamic services, VM still receives the call mask via rs_set_priv call that will be soon replaced by the upcoming publish/subscribe model. RS CHANGES: - The system process table has been reorganized and split into private entries and public entries. Only the latter ones are exposed to system services. - VM call masks are now entirely configured in rs/table.c - RS has now its own slot in the system process table. Only kernel tasks and user processes not included in the boot image are now left out from the system process table. - RS implements the initialization protocol for system services. - For services in the boot image, RS blocks till initialization is complete and panics when failure is reported back. Services are initialized in their order of appearance in the boot image priv table and RS blocks to implements synchronous initialization for every system service having the flag SF_SYNCH_BOOT set. - For services started dynamically, the initialization protocol is implemented as though it were the first ping for the service. In this case, if the system service fails to report back (or reports failure), RS brings the service down rather than trying to restart it.
2010-01-08 02:20:42 +01:00
return(OK);
2008-02-25 13:07:19 +01:00
}
2010-01-11 15:22:29 +01:00
/* Returns 0 if no device found, or 1 if a device is found. */
2008-02-25 13:07:19 +01:00
static int find_dev(devindp, capaddrp)
int *devindp;
u8_t *capaddrp;
{
int r, devind, first;
u8_t capptr, type, next, subtype;
u16_t vid, did, status;
pci_init();
first= 1;
for(;;)
{
if (first)
{
first= 0;
r= pci_first_dev(&devind, &vid, &did);
if (!r)
{
printf("amddev`find_dev: no first dev\n");
2010-01-11 15:22:29 +01:00
return 0;
2008-02-25 13:07:19 +01:00
}
}
else
{
r= pci_next_dev(&devind, &vid, &did);
if (!r)
{
printf("amddev`find_dev: no next dev\n");
2010-01-11 15:22:29 +01:00
return 0;
2008-02-25 13:07:19 +01:00
}
}
printf("amddev`find_dev: got devind %d, vid 0x%x, did 0x%x\n",
devind, vid, did);
/* Check capabilities bit in the device status register */
status= pci_attr_r16(devind, PCI_SR);
if (!(status & PSR_CAPPTR))
continue;
capptr= (pci_attr_r8(devind, PCI_CAPPTR) & PCI_CP_MASK);
while (capptr != 0)
{
type = pci_attr_r8(devind, capptr+CAP_TYPE);
next= (pci_attr_r8(devind, capptr+CAP_NEXT) &
PCI_CP_MASK);
if (type == CAP_T_SECURE_DEV)
{
printf(
"amddev`find_dev: found secure device\n");
subtype= (pci_attr_r8(devind, capptr+
CAP_SD_INFO) & CAP_SD_SUBTYPE_MASK);
if (subtype == CAP_T_SD_DEV)
{
printf("amddev`find_dev: AMD DEV\n");
pci_reserve(devind);
*devindp= devind;
*capaddrp= capptr;
return 1;
}
}
capptr= next;
}
}
return 0;
}
static u32_t read_reg(int function, int index)
{
pci_attr_w32(dev_devind, dev_capptr + DEV_OP, ((function <<
DEV_OP_FUNC_SHIFT) | index));
return pci_attr_r32(dev_devind, dev_capptr + DEV_DATA);
}
static void write_reg(int function, int index, u32_t value)
{
pci_attr_w32(dev_devind, dev_capptr + DEV_OP, ((function <<
DEV_OP_FUNC_SHIFT) | index));
pci_attr_w32(dev_devind, dev_capptr + DEV_DATA, value);
}
static void init_domain(int index)
{
size_t size, memsize;
2008-02-25 13:07:19 +01:00
phys_bytes busaddr;
size= 0x100000 / 8;
table= alloc_contig(size, AC_ALIGN4K, &busaddr);
2008-02-25 13:07:19 +01:00
if (table == NULL)
panic("AMDDEV","malloc failed", NO_NUM);
if (index == 0)
{
memset(table, 0, size);
memsize= 0x37000 / 8;
printf("memsize = 0x%x / 8\n", memsize*8);
memset(table, 0xff, memsize);
}
else
{
memset(table, 0xff, size);
memset(table, 0x00, size);
}
printf("init_domain: busaddr = %p\n", busaddr);
write_reg(DEVF_BASE_HI, index, 0);
write_reg(DEVF_BASE_LO, index, busaddr | 3);
printf("after write: DEVF_BASE_LO: 0x%x\n",
read_reg(DEVF_BASE_LO, index));
}
static void init_map(unsigned int ix)
2008-02-25 13:07:19 +01:00
{
u32_t v, dom, busno, unit0, unit1;
dom= 1;
busno= 7;
unit1= 9;
unit0= 9;
v= (dom << 26) | (dom << 20) | (busno << 12) |
(0 << 11) | (unit1 << 6) |
(0 << 5) | (unit0 << 0);
write_reg(DEVF_MAP, ix, v);
2008-02-25 13:07:19 +01:00
printf("after write: DEVF_MAP: 0x%x\n", read_reg(DEVF_MAP, ix));
2008-02-25 13:07:19 +01:00
}
#if 0
2008-02-25 13:07:19 +01:00
static int do_add(message *m)
{
int r;
endpoint_t proc;
vir_bytes start;
size_t size;
phys_bytes busaddr;
proc= m->m_source;
start= m->m2_l1;
size= m->m2_l2;
#if 0
printf("amddev`do_add: got request for 0x%x@0x%x from %d\n",
size, start, proc);
#endif
if (start % I386_PAGE_SIZE)
2008-02-25 13:07:19 +01:00
{
printf("amddev`do_add: bad start 0x%x from proc %d\n",
start, proc);
return EINVAL;
}
if (size % I386_PAGE_SIZE)
2008-02-25 13:07:19 +01:00
{
printf("amddev`do_add: bad size 0x%x from proc %d\n",
size, proc);
return EINVAL;
}
r= sys_umap(proc, VM_D, (vir_bytes)start, size, &busaddr);
2008-02-25 13:07:19 +01:00
if (r != OK)
{
printf("amddev`do_add: umap failed for 0x%x@0x%x, proc %d\n",
size, start, proc);
return r;
}
add_range(busaddr, size);
}
#endif
2008-02-25 13:07:19 +01:00
static int do_add4pci(message *m)
{
int r, pci_bus, pci_dev, pci_func;
endpoint_t proc;
vir_bytes start;
size_t size;
phys_bytes busaddr;
proc= m->m_source;
start= m->m2_l1;
size= m->m2_l2;
pci_bus= m->m1_i1;
pci_dev= m->m1_i2;
pci_func= m->m1_i3;
printf(
"amddev`do_add4pci: got request for 0x%x@0x%x from %d for pci dev %u.%u.%u\n",
size, start, proc, pci_bus, pci_dev, pci_func);
if (start % I386_PAGE_SIZE)
2008-02-25 13:07:19 +01:00
{
printf("amddev`do_add4pci: bad start 0x%x from proc %d\n",
start, proc);
return EINVAL;
}
if (size % I386_PAGE_SIZE)
2008-02-25 13:07:19 +01:00
{
printf("amddev`do_add4pci: bad size 0x%x from proc %d\n",
size, proc);
return EINVAL;
}
printf("amddev`do_add4pci: should check with PCI\n");
r= sys_umap(proc, VM_D, (vir_bytes)start, size, &busaddr);
2008-02-25 13:07:19 +01:00
if (r != OK)
{
printf(
"amddev`do_add4pci: umap failed for 0x%x@0x%x, proc %d: %d\n",
size, start, proc, r);
return r;
}
r= adddma(proc, start, size);
2008-02-25 13:07:19 +01:00
if (r != 0)
{
r= -errno;
printf(
"amddev`do_add4pci: adddma failed for 0x%x@0x%x, proc %d: %d\n",
size, start, proc, r);
return r;
}
add_range(busaddr, size);
return OK;
}
static void add_range(u32_t busaddr, u32_t size)
{
u32_t o;
2008-02-25 13:07:19 +01:00
#if 0
printf("add_range: mapping 0x%x@0x%x\n", size, busaddr);
#endif
for (o= 0; o<size; o += I386_PAGE_SIZE)
2008-02-25 13:07:19 +01:00
{
u32_t bit= (busaddr+o)/I386_PAGE_SIZE;
table[bit/8] &= ~(1U << (bit % 8));
2008-02-25 13:07:19 +01:00
}
}
static void del_range(u32_t busaddr, u32_t size)
{
u32_t o, bit;
#if 0
printf("del_range: mapping 0x%x@0x%x\n", size, busaddr);
#endif
for (o= 0; o<size; o += I386_PAGE_SIZE)
2008-02-25 13:07:19 +01:00
{
bit= (busaddr+o)/I386_PAGE_SIZE;
2008-02-25 13:07:19 +01:00
table[bit/8] |= (1 << (bit % 8));
}
}
2010-01-11 15:22:29 +01:00
static void do_pm_notify(message *m)
2008-02-25 13:07:19 +01:00
{
int r;
endpoint_t proc_e;
phys_bytes base, size;
if (m->m_source != PM_PROC_NR)
{
printf("amddev`do_pm_notify: notify not from PM (from %d)\n",
m->m_source);
return;
}
for (;;)
{
r= getdma(&proc_e, &base, &size);
if (r == -1)
{
if (errno != -EAGAIN)
{
printf(
"amddev`do_pm_notify: getdma failed: %d\n",
errno);
}
break;
}
printf(
"amddev`do_pm_notify: deleting 0x%x@0x%x for proc %d\n",
size, base, proc_e);
del_range(base, size);
r= deldma(proc_e, base, size);
if (r == -1)
{
printf("amddev`do_pm_notify: deldma failed: %d\n",
errno);
break;
}
}
}
static void report_exceptions(void)
{
u32_t status;
status= read_reg(DEVF_ERR_STATUS, 0);
if (!(status & 0x80000000))
return;
printf("amddev: status = 0x%x, addr-lo = 0x%x, addr-hi = 0x%x\n",
status, read_reg(DEVF_ERR_ADDR_LO, 0),
read_reg(DEVF_ERR_ADDR_HI, 0));
write_reg(DEVF_ERR_STATUS, 0, 0);
}