minix/minix/kernel/debug.c

565 lines
11 KiB
C
Raw Normal View History

/* This file implements kernel debugging functionality that is not included
* in the standard kernel. Available functionality includes timing of lock
* functions and sanity checking of the scheduling queues.
2005-04-21 16:53:53 +02:00
*/
#include "kernel/kernel.h"
2010-06-24 15:31:40 +02:00
#include <minix/callnr.h>
#include <minix/sysutil.h>
2010-11-18 16:12:01 +01:00
#include <minix/u64.h>
#include <limits.h>
#include <string.h>
#include <assert.h>
#define MAX_LOOP (NR_PROCS + NR_TASKS)
2012-03-25 20:25:53 +02:00
int runqueues_ok_cpu(unsigned cpu)
{
int q, l = 0;
register struct proc *xp;
struct proc **rdy_head, **rdy_tail;
rdy_head = get_cpu_var(cpu, run_q_head);
rdy_tail = get_cpu_var(cpu, run_q_tail);
for (xp = BEG_PROC_ADDR; xp < END_PROC_ADDR; ++xp) {
xp->p_found = 0;
if (l++ > MAX_LOOP) panic("check error");
}
for (q=l=0; q < NR_SCHED_QUEUES; q++) {
2005-09-11 18:44:06 +02:00
if (rdy_head[q] && !rdy_tail[q]) {
printf("head but no tail in %d\n", q);
return 0;
}
2005-09-11 18:44:06 +02:00
if (!rdy_head[q] && rdy_tail[q]) {
printf("tail but no head in %d\n", q);
return 0;
}
if (rdy_tail[q] && rdy_tail[q]->p_nextready) {
printf("tail and tail->next not null in %d\n", q);
return 0;
}
for(xp = rdy_head[q]; xp; xp = xp->p_nextready) {
2010-03-27 15:31:00 +01:00
const vir_bytes vxp = (vir_bytes) xp;
vir_bytes dxp;
if(vxp < (vir_bytes) BEG_PROC_ADDR || vxp >= (vir_bytes) END_PROC_ADDR) {
printf("xp out of range\n");
return 0;
}
dxp = vxp - (vir_bytes) BEG_PROC_ADDR;
if(dxp % sizeof(struct proc)) {
printf("xp not a real pointer");
return 0;
}
if(!proc_ptr_ok(xp)) {
printf("xp bogus pointer");
return 0;
}
if (RTS_ISSET(xp, RTS_SLOT_FREE)) {
printf("scheduling error: dead proc q %d %d\n",
q, xp->p_endpoint);
return 0;
}
if (!proc_is_runnable(xp)) {
printf("scheduling error: unready on runq %d proc %d\n",
q, xp->p_nr);
return 0;
}
2005-09-11 18:44:06 +02:00
if (xp->p_priority != q) {
printf("scheduling error: wrong priority q %d proc %d ep %d name %s\n",
q, xp->p_nr, xp->p_endpoint, xp->p_name);
return 0;
}
2005-09-11 18:44:06 +02:00
if (xp->p_found) {
printf("scheduling error: double sched q %d proc %d\n",
q, xp->p_nr);
return 0;
}
xp->p_found = 1;
if (!xp->p_nextready && rdy_tail[q] != xp) {
printf("sched err: last element not tail q %d proc %d\n",
q, xp->p_nr);
return 0;
}
if (l++ > MAX_LOOP) {
printf("loop in schedule queue?");
return 0;
}
}
}
for (xp = BEG_PROC_ADDR; xp < END_PROC_ADDR; ++xp) {
if(!proc_ptr_ok(xp)) {
printf("xp bogus pointer in proc table\n");
return 0;
}
Primary goal for these changes is: - no longer have kernel have its own page table that is loaded on every kernel entry (trap, interrupt, exception). the primary purpose is to reduce the number of required reloads. Result: - kernel can only access memory of process that was running when kernel was entered - kernel must be mapped into every process page table, so traps to kernel keep working Problem: - kernel must often access memory of arbitrary processes (e.g. send arbitrary processes messages); this can't happen directly any more; usually because that process' page table isn't loaded at all, sometimes because that memory isn't mapped in at all, sometimes because it isn't mapped in read-write. So: - kernel must be able to map in memory of any process, in its own address space. Implementation: - VM and kernel share a range of memory in which addresses of all page tables of all processes are available. This has two purposes: . Kernel has to know what data to copy in order to map in a range . Kernel has to know where to write the data in order to map it in That last point is because kernel has to write in the currently loaded page table. - Processes and kernel are separated through segments; kernel segments haven't changed. - The kernel keeps the process whose page table is currently loaded in 'ptproc.' - If it wants to map in a range of memory, it writes the value of the page directory entry for that range into the page directory entry in the currently loaded map. There is a slot reserved for such purposes. The kernel can then access this memory directly. - In order to do this, its segment has been increased (and the segments of processes start where it ends). - In the pagefault handler, detect if the kernel is doing 'trappable' memory access (i.e. a pagefault isn't a fatal error) and if so, - set the saved instruction pointer to phys_copy_fault, breaking out of phys_copy - set the saved eax register to the address of the page fault, both for sanity checking and for checking in which of the two ranges that phys_copy was called with the fault occured - Some boot-time processes do not have their own page table, and are mapped in with the kernel, and separated with segments. The kernel detects this using HASPT. If such a process has to be scheduled, any page table will work and no page table switch is done. Major changes in kernel are - When accessing user processes memory, kernel no longer explicitly checks before it does so if that memory is OK. It simply makes the mapping (if necessary), tries to do the operation, and traps the pagefault if that memory isn't present; if that happens, the copy function returns EFAULT. So all of the CHECKRANGE_OR_SUSPEND macros are gone. - Kernel no longer has to copy/read and parse page tables. - A message copying optimisation: when messages are copied, and the recipient isn't mapped in, they are copied into a buffer in the kernel. This is done in QueueMess. The next time the recipient is scheduled, this message is copied into its memory. This happens in schedcheck(). This eliminates the mapping/copying step for messages, and makes it easier to deliver messages. This eliminates soft_notify. - Kernel no longer creates a page table at all, so the vm_setbuf and pagetable writing in memory.c is gone. Minor changes in kernel are - ipc_stats thrown out, wasn't used - misc flags all renamed to MF_* - NOREC_* macros to enter and leave functions that should not be called recursively; just sanity checks really - code to fully decode segment selectors and descriptors to print on exceptions - lots of vmassert()s added, only executed if DEBUG_VMASSERT is 1
2009-09-21 16:31:52 +02:00
if (isemptyp(xp))
continue;
if(proc_is_runnable(xp) && !xp->p_found) {
printf("sched error: ready proc %d not on queue\n", xp->p_nr);
return 0;
}
}
/* All is ok. */
return 1;
}
Primary goal for these changes is: - no longer have kernel have its own page table that is loaded on every kernel entry (trap, interrupt, exception). the primary purpose is to reduce the number of required reloads. Result: - kernel can only access memory of process that was running when kernel was entered - kernel must be mapped into every process page table, so traps to kernel keep working Problem: - kernel must often access memory of arbitrary processes (e.g. send arbitrary processes messages); this can't happen directly any more; usually because that process' page table isn't loaded at all, sometimes because that memory isn't mapped in at all, sometimes because it isn't mapped in read-write. So: - kernel must be able to map in memory of any process, in its own address space. Implementation: - VM and kernel share a range of memory in which addresses of all page tables of all processes are available. This has two purposes: . Kernel has to know what data to copy in order to map in a range . Kernel has to know where to write the data in order to map it in That last point is because kernel has to write in the currently loaded page table. - Processes and kernel are separated through segments; kernel segments haven't changed. - The kernel keeps the process whose page table is currently loaded in 'ptproc.' - If it wants to map in a range of memory, it writes the value of the page directory entry for that range into the page directory entry in the currently loaded map. There is a slot reserved for such purposes. The kernel can then access this memory directly. - In order to do this, its segment has been increased (and the segments of processes start where it ends). - In the pagefault handler, detect if the kernel is doing 'trappable' memory access (i.e. a pagefault isn't a fatal error) and if so, - set the saved instruction pointer to phys_copy_fault, breaking out of phys_copy - set the saved eax register to the address of the page fault, both for sanity checking and for checking in which of the two ranges that phys_copy was called with the fault occured - Some boot-time processes do not have their own page table, and are mapped in with the kernel, and separated with segments. The kernel detects this using HASPT. If such a process has to be scheduled, any page table will work and no page table switch is done. Major changes in kernel are - When accessing user processes memory, kernel no longer explicitly checks before it does so if that memory is OK. It simply makes the mapping (if necessary), tries to do the operation, and traps the pagefault if that memory isn't present; if that happens, the copy function returns EFAULT. So all of the CHECKRANGE_OR_SUSPEND macros are gone. - Kernel no longer has to copy/read and parse page tables. - A message copying optimisation: when messages are copied, and the recipient isn't mapped in, they are copied into a buffer in the kernel. This is done in QueueMess. The next time the recipient is scheduled, this message is copied into its memory. This happens in schedcheck(). This eliminates the mapping/copying step for messages, and makes it easier to deliver messages. This eliminates soft_notify. - Kernel no longer creates a page table at all, so the vm_setbuf and pagetable writing in memory.c is gone. Minor changes in kernel are - ipc_stats thrown out, wasn't used - misc flags all renamed to MF_* - NOREC_* macros to enter and leave functions that should not be called recursively; just sanity checks really - code to fully decode segment selectors and descriptors to print on exceptions - lots of vmassert()s added, only executed if DEBUG_VMASSERT is 1
2009-09-21 16:31:52 +02:00
#ifdef CONFIG_SMP
2012-03-25 20:25:53 +02:00
static int runqueues_ok_all(void)
{
unsigned c;
for (c = 0 ; c < ncpus; c++) {
if (!runqueues_ok_cpu(c))
return 0;
}
return 1;
}
2012-03-25 20:25:53 +02:00
int runqueues_ok(void)
{
return runqueues_ok_all();
}
#else
2012-03-25 20:25:53 +02:00
int runqueues_ok(void)
{
return runqueues_ok_cpu(0);
}
#endif
2012-03-25 20:25:53 +02:00
char *
rtsflagstr(const u32_t flags)
Primary goal for these changes is: - no longer have kernel have its own page table that is loaded on every kernel entry (trap, interrupt, exception). the primary purpose is to reduce the number of required reloads. Result: - kernel can only access memory of process that was running when kernel was entered - kernel must be mapped into every process page table, so traps to kernel keep working Problem: - kernel must often access memory of arbitrary processes (e.g. send arbitrary processes messages); this can't happen directly any more; usually because that process' page table isn't loaded at all, sometimes because that memory isn't mapped in at all, sometimes because it isn't mapped in read-write. So: - kernel must be able to map in memory of any process, in its own address space. Implementation: - VM and kernel share a range of memory in which addresses of all page tables of all processes are available. This has two purposes: . Kernel has to know what data to copy in order to map in a range . Kernel has to know where to write the data in order to map it in That last point is because kernel has to write in the currently loaded page table. - Processes and kernel are separated through segments; kernel segments haven't changed. - The kernel keeps the process whose page table is currently loaded in 'ptproc.' - If it wants to map in a range of memory, it writes the value of the page directory entry for that range into the page directory entry in the currently loaded map. There is a slot reserved for such purposes. The kernel can then access this memory directly. - In order to do this, its segment has been increased (and the segments of processes start where it ends). - In the pagefault handler, detect if the kernel is doing 'trappable' memory access (i.e. a pagefault isn't a fatal error) and if so, - set the saved instruction pointer to phys_copy_fault, breaking out of phys_copy - set the saved eax register to the address of the page fault, both for sanity checking and for checking in which of the two ranges that phys_copy was called with the fault occured - Some boot-time processes do not have their own page table, and are mapped in with the kernel, and separated with segments. The kernel detects this using HASPT. If such a process has to be scheduled, any page table will work and no page table switch is done. Major changes in kernel are - When accessing user processes memory, kernel no longer explicitly checks before it does so if that memory is OK. It simply makes the mapping (if necessary), tries to do the operation, and traps the pagefault if that memory isn't present; if that happens, the copy function returns EFAULT. So all of the CHECKRANGE_OR_SUSPEND macros are gone. - Kernel no longer has to copy/read and parse page tables. - A message copying optimisation: when messages are copied, and the recipient isn't mapped in, they are copied into a buffer in the kernel. This is done in QueueMess. The next time the recipient is scheduled, this message is copied into its memory. This happens in schedcheck(). This eliminates the mapping/copying step for messages, and makes it easier to deliver messages. This eliminates soft_notify. - Kernel no longer creates a page table at all, so the vm_setbuf and pagetable writing in memory.c is gone. Minor changes in kernel are - ipc_stats thrown out, wasn't used - misc flags all renamed to MF_* - NOREC_* macros to enter and leave functions that should not be called recursively; just sanity checks really - code to fully decode segment selectors and descriptors to print on exceptions - lots of vmassert()s added, only executed if DEBUG_VMASSERT is 1
2009-09-21 16:31:52 +02:00
{
static char str[100];
str[0] = '\0';
2012-08-15 13:12:11 +02:00
#define FLAG(n) if(flags & n) { strlcat(str, #n " ", sizeof(str)); }
Primary goal for these changes is: - no longer have kernel have its own page table that is loaded on every kernel entry (trap, interrupt, exception). the primary purpose is to reduce the number of required reloads. Result: - kernel can only access memory of process that was running when kernel was entered - kernel must be mapped into every process page table, so traps to kernel keep working Problem: - kernel must often access memory of arbitrary processes (e.g. send arbitrary processes messages); this can't happen directly any more; usually because that process' page table isn't loaded at all, sometimes because that memory isn't mapped in at all, sometimes because it isn't mapped in read-write. So: - kernel must be able to map in memory of any process, in its own address space. Implementation: - VM and kernel share a range of memory in which addresses of all page tables of all processes are available. This has two purposes: . Kernel has to know what data to copy in order to map in a range . Kernel has to know where to write the data in order to map it in That last point is because kernel has to write in the currently loaded page table. - Processes and kernel are separated through segments; kernel segments haven't changed. - The kernel keeps the process whose page table is currently loaded in 'ptproc.' - If it wants to map in a range of memory, it writes the value of the page directory entry for that range into the page directory entry in the currently loaded map. There is a slot reserved for such purposes. The kernel can then access this memory directly. - In order to do this, its segment has been increased (and the segments of processes start where it ends). - In the pagefault handler, detect if the kernel is doing 'trappable' memory access (i.e. a pagefault isn't a fatal error) and if so, - set the saved instruction pointer to phys_copy_fault, breaking out of phys_copy - set the saved eax register to the address of the page fault, both for sanity checking and for checking in which of the two ranges that phys_copy was called with the fault occured - Some boot-time processes do not have their own page table, and are mapped in with the kernel, and separated with segments. The kernel detects this using HASPT. If such a process has to be scheduled, any page table will work and no page table switch is done. Major changes in kernel are - When accessing user processes memory, kernel no longer explicitly checks before it does so if that memory is OK. It simply makes the mapping (if necessary), tries to do the operation, and traps the pagefault if that memory isn't present; if that happens, the copy function returns EFAULT. So all of the CHECKRANGE_OR_SUSPEND macros are gone. - Kernel no longer has to copy/read and parse page tables. - A message copying optimisation: when messages are copied, and the recipient isn't mapped in, they are copied into a buffer in the kernel. This is done in QueueMess. The next time the recipient is scheduled, this message is copied into its memory. This happens in schedcheck(). This eliminates the mapping/copying step for messages, and makes it easier to deliver messages. This eliminates soft_notify. - Kernel no longer creates a page table at all, so the vm_setbuf and pagetable writing in memory.c is gone. Minor changes in kernel are - ipc_stats thrown out, wasn't used - misc flags all renamed to MF_* - NOREC_* macros to enter and leave functions that should not be called recursively; just sanity checks really - code to fully decode segment selectors and descriptors to print on exceptions - lots of vmassert()s added, only executed if DEBUG_VMASSERT is 1
2009-09-21 16:31:52 +02:00
FLAG(RTS_SLOT_FREE);
FLAG(RTS_PROC_STOP);
FLAG(RTS_SENDING);
FLAG(RTS_RECEIVING);
FLAG(RTS_SIGNALED);
FLAG(RTS_SIG_PENDING);
FLAG(RTS_P_STOP);
FLAG(RTS_NO_PRIV);
FLAG(RTS_NO_ENDPOINT);
FLAG(RTS_VMINHIBIT);
FLAG(RTS_PAGEFAULT);
FLAG(RTS_VMREQUEST);
FLAG(RTS_VMREQTARGET);
FLAG(RTS_PREEMPTED);
FLAG(RTS_NO_QUANTUM);
Primary goal for these changes is: - no longer have kernel have its own page table that is loaded on every kernel entry (trap, interrupt, exception). the primary purpose is to reduce the number of required reloads. Result: - kernel can only access memory of process that was running when kernel was entered - kernel must be mapped into every process page table, so traps to kernel keep working Problem: - kernel must often access memory of arbitrary processes (e.g. send arbitrary processes messages); this can't happen directly any more; usually because that process' page table isn't loaded at all, sometimes because that memory isn't mapped in at all, sometimes because it isn't mapped in read-write. So: - kernel must be able to map in memory of any process, in its own address space. Implementation: - VM and kernel share a range of memory in which addresses of all page tables of all processes are available. This has two purposes: . Kernel has to know what data to copy in order to map in a range . Kernel has to know where to write the data in order to map it in That last point is because kernel has to write in the currently loaded page table. - Processes and kernel are separated through segments; kernel segments haven't changed. - The kernel keeps the process whose page table is currently loaded in 'ptproc.' - If it wants to map in a range of memory, it writes the value of the page directory entry for that range into the page directory entry in the currently loaded map. There is a slot reserved for such purposes. The kernel can then access this memory directly. - In order to do this, its segment has been increased (and the segments of processes start where it ends). - In the pagefault handler, detect if the kernel is doing 'trappable' memory access (i.e. a pagefault isn't a fatal error) and if so, - set the saved instruction pointer to phys_copy_fault, breaking out of phys_copy - set the saved eax register to the address of the page fault, both for sanity checking and for checking in which of the two ranges that phys_copy was called with the fault occured - Some boot-time processes do not have their own page table, and are mapped in with the kernel, and separated with segments. The kernel detects this using HASPT. If such a process has to be scheduled, any page table will work and no page table switch is done. Major changes in kernel are - When accessing user processes memory, kernel no longer explicitly checks before it does so if that memory is OK. It simply makes the mapping (if necessary), tries to do the operation, and traps the pagefault if that memory isn't present; if that happens, the copy function returns EFAULT. So all of the CHECKRANGE_OR_SUSPEND macros are gone. - Kernel no longer has to copy/read and parse page tables. - A message copying optimisation: when messages are copied, and the recipient isn't mapped in, they are copied into a buffer in the kernel. This is done in QueueMess. The next time the recipient is scheduled, this message is copied into its memory. This happens in schedcheck(). This eliminates the mapping/copying step for messages, and makes it easier to deliver messages. This eliminates soft_notify. - Kernel no longer creates a page table at all, so the vm_setbuf and pagetable writing in memory.c is gone. Minor changes in kernel are - ipc_stats thrown out, wasn't used - misc flags all renamed to MF_* - NOREC_* macros to enter and leave functions that should not be called recursively; just sanity checks really - code to fully decode segment selectors and descriptors to print on exceptions - lots of vmassert()s added, only executed if DEBUG_VMASSERT is 1
2009-09-21 16:31:52 +02:00
return str;
}
2012-03-25 20:25:53 +02:00
char *
miscflagstr(const u32_t flags)
Primary goal for these changes is: - no longer have kernel have its own page table that is loaded on every kernel entry (trap, interrupt, exception). the primary purpose is to reduce the number of required reloads. Result: - kernel can only access memory of process that was running when kernel was entered - kernel must be mapped into every process page table, so traps to kernel keep working Problem: - kernel must often access memory of arbitrary processes (e.g. send arbitrary processes messages); this can't happen directly any more; usually because that process' page table isn't loaded at all, sometimes because that memory isn't mapped in at all, sometimes because it isn't mapped in read-write. So: - kernel must be able to map in memory of any process, in its own address space. Implementation: - VM and kernel share a range of memory in which addresses of all page tables of all processes are available. This has two purposes: . Kernel has to know what data to copy in order to map in a range . Kernel has to know where to write the data in order to map it in That last point is because kernel has to write in the currently loaded page table. - Processes and kernel are separated through segments; kernel segments haven't changed. - The kernel keeps the process whose page table is currently loaded in 'ptproc.' - If it wants to map in a range of memory, it writes the value of the page directory entry for that range into the page directory entry in the currently loaded map. There is a slot reserved for such purposes. The kernel can then access this memory directly. - In order to do this, its segment has been increased (and the segments of processes start where it ends). - In the pagefault handler, detect if the kernel is doing 'trappable' memory access (i.e. a pagefault isn't a fatal error) and if so, - set the saved instruction pointer to phys_copy_fault, breaking out of phys_copy - set the saved eax register to the address of the page fault, both for sanity checking and for checking in which of the two ranges that phys_copy was called with the fault occured - Some boot-time processes do not have their own page table, and are mapped in with the kernel, and separated with segments. The kernel detects this using HASPT. If such a process has to be scheduled, any page table will work and no page table switch is done. Major changes in kernel are - When accessing user processes memory, kernel no longer explicitly checks before it does so if that memory is OK. It simply makes the mapping (if necessary), tries to do the operation, and traps the pagefault if that memory isn't present; if that happens, the copy function returns EFAULT. So all of the CHECKRANGE_OR_SUSPEND macros are gone. - Kernel no longer has to copy/read and parse page tables. - A message copying optimisation: when messages are copied, and the recipient isn't mapped in, they are copied into a buffer in the kernel. This is done in QueueMess. The next time the recipient is scheduled, this message is copied into its memory. This happens in schedcheck(). This eliminates the mapping/copying step for messages, and makes it easier to deliver messages. This eliminates soft_notify. - Kernel no longer creates a page table at all, so the vm_setbuf and pagetable writing in memory.c is gone. Minor changes in kernel are - ipc_stats thrown out, wasn't used - misc flags all renamed to MF_* - NOREC_* macros to enter and leave functions that should not be called recursively; just sanity checks really - code to fully decode segment selectors and descriptors to print on exceptions - lots of vmassert()s added, only executed if DEBUG_VMASSERT is 1
2009-09-21 16:31:52 +02:00
{
static char str[100];
str[0] = '\0';
FLAG(MF_REPLY_PEND);
FLAG(MF_DELIVERMSG);
FLAG(MF_KCALL_RESUME);
Primary goal for these changes is: - no longer have kernel have its own page table that is loaded on every kernel entry (trap, interrupt, exception). the primary purpose is to reduce the number of required reloads. Result: - kernel can only access memory of process that was running when kernel was entered - kernel must be mapped into every process page table, so traps to kernel keep working Problem: - kernel must often access memory of arbitrary processes (e.g. send arbitrary processes messages); this can't happen directly any more; usually because that process' page table isn't loaded at all, sometimes because that memory isn't mapped in at all, sometimes because it isn't mapped in read-write. So: - kernel must be able to map in memory of any process, in its own address space. Implementation: - VM and kernel share a range of memory in which addresses of all page tables of all processes are available. This has two purposes: . Kernel has to know what data to copy in order to map in a range . Kernel has to know where to write the data in order to map it in That last point is because kernel has to write in the currently loaded page table. - Processes and kernel are separated through segments; kernel segments haven't changed. - The kernel keeps the process whose page table is currently loaded in 'ptproc.' - If it wants to map in a range of memory, it writes the value of the page directory entry for that range into the page directory entry in the currently loaded map. There is a slot reserved for such purposes. The kernel can then access this memory directly. - In order to do this, its segment has been increased (and the segments of processes start where it ends). - In the pagefault handler, detect if the kernel is doing 'trappable' memory access (i.e. a pagefault isn't a fatal error) and if so, - set the saved instruction pointer to phys_copy_fault, breaking out of phys_copy - set the saved eax register to the address of the page fault, both for sanity checking and for checking in which of the two ranges that phys_copy was called with the fault occured - Some boot-time processes do not have their own page table, and are mapped in with the kernel, and separated with segments. The kernel detects this using HASPT. If such a process has to be scheduled, any page table will work and no page table switch is done. Major changes in kernel are - When accessing user processes memory, kernel no longer explicitly checks before it does so if that memory is OK. It simply makes the mapping (if necessary), tries to do the operation, and traps the pagefault if that memory isn't present; if that happens, the copy function returns EFAULT. So all of the CHECKRANGE_OR_SUSPEND macros are gone. - Kernel no longer has to copy/read and parse page tables. - A message copying optimisation: when messages are copied, and the recipient isn't mapped in, they are copied into a buffer in the kernel. This is done in QueueMess. The next time the recipient is scheduled, this message is copied into its memory. This happens in schedcheck(). This eliminates the mapping/copying step for messages, and makes it easier to deliver messages. This eliminates soft_notify. - Kernel no longer creates a page table at all, so the vm_setbuf and pagetable writing in memory.c is gone. Minor changes in kernel are - ipc_stats thrown out, wasn't used - misc flags all renamed to MF_* - NOREC_* macros to enter and leave functions that should not be called recursively; just sanity checks really - code to fully decode segment selectors and descriptors to print on exceptions - lots of vmassert()s added, only executed if DEBUG_VMASSERT is 1
2009-09-21 16:31:52 +02:00
return str;
}
2012-03-25 20:25:53 +02:00
char *
Userspace scheduling - cotributed by Bjorn Swift - In this first phase, scheduling is moved from the kernel to the PM server. The next steps are to a) moving scheduling to its own server and b) include useful information in the "out of quantum" message, so that the scheduler can make use of this information. - The kernel process table now keeps record of who is responsible for scheduling each process (p_scheduler). When this pointer is NULL, the process will be scheduled by the kernel. If such a process runs out of quantum, the kernel will simply renew its quantum an requeue it. - When PM loads, it will take over scheduling of all running processes, except system processes, using sys_schedctl(). Essentially, this only results in taking over init. As children inherit a scheduler from their parent, user space programs forked by init will inherit PM (for now) as their scheduler. - Once a process has been assigned a scheduler, and runs out of quantum, its RTS_NO_QUANTUM flag will be set and the process dequeued. The kernel will send a message to the scheduler, on the process' behalf, informing the scheduler that it has run out of quantum. The scheduler can take what ever action it pleases, based on its policy, and then reschedule the process using the sys_schedule() system call. - Balance queues does not work as before. While the old in-kernel function used to renew the quantum of processes in the highest priority run queue, the user-space implementation only acts on processes that have been bumped down to a lower priority queue. This approach reacts slower to changes than the old one, but saves us sending a sys_schedule message for each process every time we balance the queues. Currently, when processes are moved up a priority queue, their quantum is also renewed, but this can be fiddled with. - do_nice has been removed from kernel. PM answers to get- and setpriority calls, updates it's own nice variable as well as the max_run_queue. This will be refactored once scheduling is moved to a separate server. We will probably have PM update it's local nice value and then send a message to whoever is scheduling the process. - changes to fix an issue in do_fork() where processes could run out of quantum but bypassing the code path that handles it correctly. The future plan is to remove the policy from do_fork() and implement it in userspace too.
2010-03-29 13:07:20 +02:00
schedulerstr(struct proc *scheduler)
{
if (scheduler != NULL)
{
return scheduler->p_name;
}
return "KERNEL";
}
2012-03-25 20:25:53 +02:00
static void
2010-07-07 00:05:21 +02:00
print_proc_name(struct proc *pp)
{
char *name = pp->p_name;
endpoint_t ep = pp->p_endpoint;
if(name) {
printf("%s(%d)", name, ep);
}
else {
printf("%d", ep);
}
}
2012-03-25 20:25:53 +02:00
static void
2010-07-07 00:05:21 +02:00
print_endpoint(endpoint_t ep)
{
int proc_nr;
struct proc *pp = NULL;
switch(ep) {
case ANY:
printf("ANY");
break;
case SELF:
printf("SELF");
break;
case NONE:
printf("NONE");
break;
default:
if(!isokendpt(ep, &proc_nr)) {
printf("??? %d\n", ep);
}
else {
pp = proc_addr(proc_nr);
if(isemptyp(pp)) {
printf("??? empty slot %d\n", proc_nr);
}
else {
print_proc_name(pp);
}
}
break;
}
}
2012-03-25 20:25:53 +02:00
static void
2010-07-07 00:05:21 +02:00
print_sigmgr(struct proc *pp)
{
endpoint_t sig_mgr, bak_sig_mgr;
sig_mgr = priv(pp) ? priv(pp)->s_sig_mgr : NONE;
bak_sig_mgr = priv(pp) ? priv(pp)->s_bak_sig_mgr : NONE;
if(sig_mgr == NONE) { printf("no sigmgr"); return; }
2010-07-07 00:05:21 +02:00
printf("sigmgr ");
print_endpoint(sig_mgr);
if(bak_sig_mgr != NONE) {
printf(" / ");
print_endpoint(bak_sig_mgr);
}
}
2012-03-25 20:25:53 +02:00
void print_proc(struct proc *pp)
{
endpoint_t dep;
printf("%d: %s %d prio %d time %d/%d cycles 0x%x%08x cpu %2d "
2012-10-08 03:38:03 +02:00
"pdbr 0x%lx rts %s misc %s sched %s ",
proc_nr(pp), pp->p_name, pp->p_endpoint,
pp->p_priority, pp->p_user_time,
pp->p_sys_time, ex64hi(pp->p_cycles),
ex64lo(pp->p_cycles), pp->p_cpu,
2012-10-08 03:38:03 +02:00
#if defined(__i386__)
pp->p_seg.p_cr3,
2012-10-08 03:38:03 +02:00
#elif defined(__arm__)
pp->p_seg.p_ttbr,
#endif
rtsflagstr(pp->p_rts_flags), miscflagstr(pp->p_misc_flags),
schedulerstr(pp->p_scheduler));
2010-07-07 00:05:21 +02:00
print_sigmgr(pp);
dep = P_BLOCKEDON(pp);
if(dep != NONE) {
printf(" blocked on: ");
2010-07-07 00:05:21 +02:00
print_endpoint(dep);
}
2010-07-07 00:05:21 +02:00
printf("\n");
}
2012-03-25 20:25:53 +02:00
static void print_proc_depends(struct proc *pp, const int level)
{
struct proc *depproc = NULL;
endpoint_t dep;
#define COL { int i; for(i = 0; i < level; i++) printf("> "); }
if(level >= NR_PROCS) {
printf("loop??\n");
return;
}
COL
print_proc(pp);
COL
proc_stacktrace(pp);
dep = P_BLOCKEDON(pp);
if(dep != NONE && dep != ANY) {
int procno;
if(isokendpt(dep, &procno)) {
depproc = proc_addr(procno);
if(isemptyp(depproc))
depproc = NULL;
}
if (depproc)
print_proc_depends(depproc, level+1);
}
}
2012-03-25 20:25:53 +02:00
void print_proc_recursive(struct proc *pp)
{
print_proc_depends(pp, 0);
}
#if DEBUG_DUMPIPC || DEBUG_DUMPIPCF
static const char *mtypename(int mtype, int *possible_callname)
2010-06-24 15:31:40 +02:00
{
char *callname = NULL, *errname = NULL;
/* use generated file to recognize message types
*
* we try to match both error numbers and call numbers, as the
* reader can probably decide from context what's going on.
*
* whenever it might be a call number we tell the caller so the
* call message fields can be decoded if known.
*/
switch(mtype) {
#define IDENT(x) case x: callname = #x; *possible_callname = 1; break;
#include "kernel/extracted-mtype.h"
#undef IDENT
}
switch(mtype) {
#define IDENT(x) case x: errname = #x; break;
#include "kernel/extracted-errno.h"
#undef IDENT
2010-06-24 15:31:40 +02:00
}
/* no match */
if(!errname && !callname)
return NULL;
/* 2 matches */
if(errname && callname) {
static char typename[100];
strcpy(typename, errname);
strcat(typename, " / ");
strcat(typename, callname);
return typename;
}
if(errname) return errname;
assert(callname);
return callname;
2010-06-24 15:31:40 +02:00
}
2012-03-25 20:25:53 +02:00
static void printproc(struct proc *rp)
{
if (rp)
printf(" %s(%d)", rp->p_name, rp - proc);
else
printf(" kernel");
}
2012-03-25 20:25:53 +02:00
static void printparam(const char *name, const void *data, size_t size)
2010-06-24 15:31:40 +02:00
{
printf(" %s=", name);
switch (size) {
case sizeof(char): printf("%d", *(char *) data); break;
case sizeof(short): printf("%d", *(short *) data); break;
case sizeof(int): printf("%d", *(int *) data); break;
default: printf("(%u bytes)", size); break;
}
}
#ifdef DEBUG_DUMPIPC_NAMES
static int namematch(char **names, int nnames, char *name)
{
int i;
for(i = 0; i < nnames; i++)
if(!strcmp(names[i], name))
return 1;
return 0;
}
#endif
void printmsg(message *msg, struct proc *src, struct proc *dst,
char operation, int printparams)
2010-06-24 15:31:40 +02:00
{
const char *name;
int mtype = msg->m_type, mightbecall = 0;
#ifdef DEBUG_DUMPIPC_NAMES
{
char *names[] = DEBUG_DUMPIPC_NAMES;
int nnames = sizeof(names)/sizeof(names[0]);
/* skip printing messages for messages neither to
* or from DEBUG_DUMPIPC_EP if it is defined; either
* can be NULL to indicate kernel
*/
if(!(src && namematch(names, nnames, src->p_name)) &&
!(dst && namematch(names, nnames, dst->p_name))) {
return;
}
}
#endif
2010-06-24 15:31:40 +02:00
/* source, destination and message type */
printf("%c", operation);
printproc(src);
printproc(dst);
name = mtypename(mtype, &mightbecall);
2010-06-24 15:31:40 +02:00
if (name) {
printf(" %s(%d/0x%x)", name, mtype, mtype);
2010-06-24 15:31:40 +02:00
} else {
printf(" %d/0x%x", mtype, mtype);
2010-06-24 15:31:40 +02:00
}
if (mightbecall && printparams) {
2010-06-24 15:31:40 +02:00
#define IDENT(x, y) if (mtype == x) printparam(#y, &msg->y, sizeof(msg->y));
#include "kernel/extracted-mfield.h"
2010-06-24 15:31:40 +02:00
#undef IDENT
}
printf("\n");
}
#endif
#if DEBUG_IPCSTATS
#define IPCPROCS (NR_PROCS+1) /* number of slots we need */
#define KERNELIPC NR_PROCS /* slot number to use for kernel calls */
2012-03-25 20:25:53 +02:00
static int messages[IPCPROCS][IPCPROCS];
#define PRINTSLOTS 20
2012-03-25 20:25:53 +02:00
static struct {
int src, dst, messages;
} winners[PRINTSLOTS];
2012-03-25 20:25:53 +02:00
static int total, goodslots;
2012-03-25 20:25:53 +02:00
static void printstats(int ticks)
{
int i;
for(i = 0; i < goodslots; i++) {
#define name(s) (s == KERNELIPC ? "kernel" : proc_addr(s)->p_name)
#define persec(n) (system_hz*(n)/ticks)
char *n1 = name(winners[i].src),
*n2 = name(winners[i].dst);
printf("%2d. %8s -> %8s %9d/s\n",
i, n1, n2, persec(winners[i].messages));
}
printf("total %d/s\n", persec(total));
}
2012-03-25 20:25:53 +02:00
static void sortstats(void)
{
/* Print top message senders/receivers. */
int src_slot, dst_slot;
total = goodslots = 0;
for(src_slot = 0; src_slot < IPCPROCS; src_slot++) {
for(dst_slot = 0; dst_slot < IPCPROCS; dst_slot++) {
int w = PRINTSLOTS, rem,
n = messages[src_slot][dst_slot];
total += n;
while(w > 0 && n > winners[w-1].messages)
w--;
if(w >= PRINTSLOTS) continue;
/* This combination has beaten the current winners
* and should be inserted at position 'w.'
*/
rem = PRINTSLOTS-w-1;
assert(rem >= 0);
assert(rem < PRINTSLOTS);
if(rem > 0) {
assert(w+1 <= PRINTSLOTS-1);
assert(w >= 0);
memmove(&winners[w+1], &winners[w],
rem*sizeof(winners[0]));
}
winners[w].src = src_slot;
winners[w].dst = dst_slot;
winners[w].messages = n;
if(goodslots < PRINTSLOTS) goodslots++;
}
}
}
2010-06-24 15:31:40 +02:00
#define proc2slot(p, s) { \
if(p) { s = p->p_nr; } \
else { s = KERNELIPC; } \
assert(s >= 0 && s < IPCPROCS); \
}
2012-03-25 20:25:53 +02:00
static void statmsg(message *msg, struct proc *srcp, struct proc *dstp)
2010-06-24 15:31:40 +02:00
{
int src, dst, now, secs, dt;
static int lastprint;
/* Stat message. */
assert(src);
proc2slot(srcp, src);
proc2slot(dstp, dst);
messages[src][dst]++;
/* Print something? */
now = get_monotonic();
dt = now - lastprint;
secs = dt/system_hz;
if(secs >= 30) {
memset(winners, 0, sizeof(winners));
sortstats();
printstats(dt);
memset(messages, 0, sizeof(messages));
lastprint = now;
}
}
#endif
#if DEBUG_IPC_HOOK
2012-03-25 20:25:53 +02:00
void hook_ipc_msgkcall(message *msg, struct proc *proc)
{
#if DEBUG_DUMPIPC
printmsg(msg, proc, NULL, 'k', 1);
#endif
2010-06-24 15:31:40 +02:00
}
2012-03-25 20:25:53 +02:00
void hook_ipc_msgkresult(message *msg, struct proc *proc)
2010-06-24 15:31:40 +02:00
{
#if DEBUG_DUMPIPC
printmsg(msg, NULL, proc, 'k', 0);
#endif
#if DEBUG_IPCSTATS
statmsg(msg, proc, NULL);
#endif
2010-06-24 15:31:40 +02:00
}
2012-03-25 20:25:53 +02:00
void hook_ipc_msgrecv(message *msg, struct proc *src, struct proc *dst)
2010-06-24 15:31:40 +02:00
{
#if DEBUG_DUMPIPC
printmsg(msg, src, dst, 'r', 0);
#endif
#if DEBUG_IPCSTATS
statmsg(msg, src, dst);
#endif
2010-06-24 15:31:40 +02:00
}
2012-03-25 20:25:53 +02:00
void hook_ipc_msgsend(message *msg, struct proc *src, struct proc *dst)
2010-06-24 15:31:40 +02:00
{
#if DEBUG_DUMPIPC
printmsg(msg, src, dst, 's', 1);
#endif
}
2012-03-25 20:25:53 +02:00
void hook_ipc_clear(struct proc *p)
{
#if DEBUG_IPCSTATS
int slot, i;
assert(p);
proc2slot(p, slot);
for(i = 0; i < IPCPROCS; i++)
messages[slot][i] = messages[i][slot] = 0;
#endif
2010-06-24 15:31:40 +02:00
}
#endif