jenna
/
kernel_samsung_sm7125
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
119756 Commits
5 Branches
0 Tags
1.2 GiB
Tag: Branch: Tree: 7105de84f1
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '7105de84f1'
${ noResults }
kernel_samsung_sm7125/arch/sparc/kernel/rtrap_32.S

323 lines
6.6 KiB
Raw Normal View History Unescape

sparc: remove CVS keywords This patch removes the CVS keywords that weren't updated for a long time from comments. Signed-off-by: Adrian Bunk <bunk@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
/*
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
20 years ago
* rtrap.S: Return from Sparc trap low-level code.
*
* Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
*/
#include <asm/page.h>
#include <asm/ptrace.h>
#include <asm/psr.h>
#include <asm/asi.h>
#include <asm/smp.h>
#include <asm/contregs.h>
#include <asm/winmacro.h>
#include <asm/asmmacro.h>
#include <asm/thread_info.h>
#define t_psr l0
#define t_pc l1
#define t_npc l2
#define t_wim l3
#define twin_tmp1 l4
#define glob_tmp g4
#define curptr g6
/* 7 WINDOW SPARC PATCH INSTRUCTIONS */
.globl rtrap_7win_patch1, rtrap_7win_patch2, rtrap_7win_patch3
.globl rtrap_7win_patch4, rtrap_7win_patch5
rtrap_7win_patch1: srl %t_wim, 0x6, %glob_tmp
rtrap_7win_patch2: and %glob_tmp, 0x7f, %glob_tmp
rtrap_7win_patch3: srl %g1, 7, %g2
rtrap_7win_patch4: srl %g2, 6, %g2
rtrap_7win_patch5: and %g1, 0x7f, %g1
/* END OF PATCH INSTRUCTIONS */
/* We need to check for a few things which are:
* 1) The need to call schedule() because this
* processes quantum is up.
* 2) Pending signals for this process, if any
* exist we need to call do_signal() to do
* the needy.
*
* Else we just check if the rett would land us
* in an invalid window, if so we need to grab
* it off the user/kernel stack first.
*/
.globl ret_trap_entry, rtrap_patch1, rtrap_patch2
.globl rtrap_patch3, rtrap_patch4, rtrap_patch5
.globl ret_trap_lockless_ipi
ret_trap_entry:
ret_trap_lockless_ipi:
andcc %t_psr, PSR_PS, %g0
sparc: Fix debugger syscall restart interactions. So, forever, we've had this ptrace_signal_deliver implementation which tries to handle all of the nasties that can occur when the debugger looks at a process about to take a signal. It's meant to address all of these issues inside of the kernel so that the debugger need not be mindful of such things. Problem is, this doesn't work. The idea was that we should do the syscall restart business first, so that the debugger captures that state. Otherwise, if the debugger for example saves the child's state, makes the child execute something else, then restores the saved state, we won't handle the syscall restart properly because we lose the "we're in a syscall" state. The code here worked for most cases, but if the debugger actually passes the signal through to the child unaltered, it's possible that we would do a syscall restart when we shouldn't have. In particular this breaks the case of debugging a process under a gdb which is being debugged by yet another gdb. gdb uses sigsuspend to wait for SIGCHLD of the inferior, but if gdb itself is being debugged by a top-level gdb we get a ptrace_stop(). The top-level gdb does a PTRACE_CONT with SIGCHLD to let the inferior gdb see the signal. But ptrace_signal_deliver() assumed the debugger would cancel out the signal and therefore did a syscall restart, because the return error was ERESTARTNOHAND. Fix this by simply making ptrace_signal_deliver() a nop, and providing a way for the debugger to control system call restarting properly: 1) Report a "in syscall" software bit in regs->{tstate,psr}. It is set early on in trap entry to a system call and is fully visible to the debugger via ptrace() and regsets. 2) Test this bit right before doing a syscall restart. We have to do a final recheck right after get_signal_to_deliver() in case the debugger cleared the bit during ptrace_stop(). 3) Clear the bit in trap return so we don't accidently try to set that bit in the real register. As a result we also get a ptrace_{is,clear}_syscall() for sparc32 just like sparc64 has. M68K has this same exact bug, and is now the only other user of the ptrace_signal_deliver hook. It needs to be fixed in the same exact way as sparc. Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
sethi %hi(PSR_SYSCALL), %g1
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
20 years ago
be 1f
sparc: Fix debugger syscall restart interactions. So, forever, we've had this ptrace_signal_deliver implementation which tries to handle all of the nasties that can occur when the debugger looks at a process about to take a signal. It's meant to address all of these issues inside of the kernel so that the debugger need not be mindful of such things. Problem is, this doesn't work. The idea was that we should do the syscall restart business first, so that the debugger captures that state. Otherwise, if the debugger for example saves the child's state, makes the child execute something else, then restores the saved state, we won't handle the syscall restart properly because we lose the "we're in a syscall" state. The code here worked for most cases, but if the debugger actually passes the signal through to the child unaltered, it's possible that we would do a syscall restart when we shouldn't have. In particular this breaks the case of debugging a process under a gdb which is being debugged by yet another gdb. gdb uses sigsuspend to wait for SIGCHLD of the inferior, but if gdb itself is being debugged by a top-level gdb we get a ptrace_stop(). The top-level gdb does a PTRACE_CONT with SIGCHLD to let the inferior gdb see the signal. But ptrace_signal_deliver() assumed the debugger would cancel out the signal and therefore did a syscall restart, because the return error was ERESTARTNOHAND. Fix this by simply making ptrace_signal_deliver() a nop, and providing a way for the debugger to control system call restarting properly: 1) Report a "in syscall" software bit in regs->{tstate,psr}. It is set early on in trap entry to a system call and is fully visible to the debugger via ptrace() and regsets. 2) Test this bit right before doing a syscall restart. We have to do a final recheck right after get_signal_to_deliver() in case the debugger cleared the bit during ptrace_stop(). 3) Clear the bit in trap return so we don't accidently try to set that bit in the real register. As a result we also get a ptrace_{is,clear}_syscall() for sparc32 just like sparc64 has. M68K has this same exact bug, and is now the only other user of the ptrace_signal_deliver hook. It needs to be fixed in the same exact way as sparc. Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
andn %t_psr, %g1, %t_psr
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
20 years ago
wr %t_psr, 0x0, %psr
b ret_trap_kernel
nop
1:
ld [%curptr + TI_FLAGS], %g2
andcc %g2, (_TIF_NEED_RESCHED), %g0
be signal_p
nop
call schedule
nop
ld [%curptr + TI_FLAGS], %g2
signal_p:
sparc: Create and use TIF_NOTIFY_RESUME. Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
andcc %g2, _TIF_DO_NOTIFY_RESUME_MASK, %g0
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
20 years ago
bz,a ret_trap_continue
ld [%sp + STACKFRAME_SZ + PT_PSR], %t_psr
sparc: Create and use TIF_NOTIFY_RESUME. Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
mov %g2, %o2
[SPARC]: Add support for *at(), ppoll, and pselect syscalls. This also includes by necessity _TIF_RESTORE_SIGMASK support, which actually resulted in a lot of cleanups. The sparc signal handling code is quite a mess and I should clean it up some day. Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago
mov %l5, %o1
sparc: Create and use TIF_NOTIFY_RESUME. Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
call do_notify_resume
[SPARC]: Add support for *at(), ppoll, and pselect syscalls. This also includes by necessity _TIF_RESTORE_SIGMASK support, which actually resulted in a lot of cleanups. The sparc signal handling code is quite a mess and I should clean it up some day. Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago
add %sp, STACKFRAME_SZ, %o0 ! pt_regs ptr
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
20 years ago
/* Fall through. */
ld [%sp + STACKFRAME_SZ + PT_PSR], %t_psr
clr %l6
ret_trap_continue:
sparc: Fix debugger syscall restart interactions. So, forever, we've had this ptrace_signal_deliver implementation which tries to handle all of the nasties that can occur when the debugger looks at a process about to take a signal. It's meant to address all of these issues inside of the kernel so that the debugger need not be mindful of such things. Problem is, this doesn't work. The idea was that we should do the syscall restart business first, so that the debugger captures that state. Otherwise, if the debugger for example saves the child's state, makes the child execute something else, then restores the saved state, we won't handle the syscall restart properly because we lose the "we're in a syscall" state. The code here worked for most cases, but if the debugger actually passes the signal through to the child unaltered, it's possible that we would do a syscall restart when we shouldn't have. In particular this breaks the case of debugging a process under a gdb which is being debugged by yet another gdb. gdb uses sigsuspend to wait for SIGCHLD of the inferior, but if gdb itself is being debugged by a top-level gdb we get a ptrace_stop(). The top-level gdb does a PTRACE_CONT with SIGCHLD to let the inferior gdb see the signal. But ptrace_signal_deliver() assumed the debugger would cancel out the signal and therefore did a syscall restart, because the return error was ERESTARTNOHAND. Fix this by simply making ptrace_signal_deliver() a nop, and providing a way for the debugger to control system call restarting properly: 1) Report a "in syscall" software bit in regs->{tstate,psr}. It is set early on in trap entry to a system call and is fully visible to the debugger via ptrace() and regsets. 2) Test this bit right before doing a syscall restart. We have to do a final recheck right after get_signal_to_deliver() in case the debugger cleared the bit during ptrace_stop(). 3) Clear the bit in trap return so we don't accidently try to set that bit in the real register. As a result we also get a ptrace_{is,clear}_syscall() for sparc32 just like sparc64 has. M68K has this same exact bug, and is now the only other user of the ptrace_signal_deliver hook. It needs to be fixed in the same exact way as sparc. Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
sethi %hi(PSR_SYSCALL), %g1
andn %t_psr, %g1, %t_psr
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
20 years ago
wr %t_psr, 0x0, %psr
WRITE_PAUSE
ld [%curptr + TI_W_SAVED], %twin_tmp1
orcc %g0, %twin_tmp1, %g0
be ret_trap_nobufwins
nop
wr %t_psr, PSR_ET, %psr
WRITE_PAUSE
mov 1, %o1
call try_to_clear_window_buffer
add %sp, STACKFRAME_SZ, %o0
b signal_p
ld [%curptr + TI_FLAGS], %g2
ret_trap_nobufwins:
/* Load up the user's out registers so we can pull
* a window from the stack, if necessary.
*/
LOAD_PT_INS(sp)
/* If there are already live user windows in the
* set we can return from trap safely.
*/
ld [%curptr + TI_UWINMASK], %twin_tmp1
orcc %g0, %twin_tmp1, %g0
bne ret_trap_userwins_ok
nop
/* Calculate new %wim, we have to pull a register
* window from the users stack.
*/
ret_trap_pull_one_window:
rd %wim, %t_wim
sll %t_wim, 0x1, %twin_tmp1
rtrap_patch1: srl %t_wim, 0x7, %glob_tmp
or %glob_tmp, %twin_tmp1, %glob_tmp
rtrap_patch2: and %glob_tmp, 0xff, %glob_tmp
wr %glob_tmp, 0x0, %wim
/* Here comes the architecture specific
* branch to the user stack checking routine
* for return from traps.
*/
.globl rtrap_mmu_patchme
rtrap_mmu_patchme: b sun4c_rett_stackchk
andcc %fp, 0x7, %g0
ret_trap_userwins_ok:
LOAD_PT_PRIV(sp, t_psr, t_pc, t_npc)
or %t_pc, %t_npc, %g2
andcc %g2, 0x3, %g0
sparc32: fix rtrap.S typo Fix compile problem in rtrap.S arch/sparc/kernel/built-in.o: In function `ret_trap_userwins_ok': arch/sparc/kernel/rtrap.S:(.text+0x1900): undefined reference to `PSR_SYCALL' Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
sethi %hi(PSR_SYSCALL), %g2
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
20 years ago
be 1f
sparc: Fix debugger syscall restart interactions. So, forever, we've had this ptrace_signal_deliver implementation which tries to handle all of the nasties that can occur when the debugger looks at a process about to take a signal. It's meant to address all of these issues inside of the kernel so that the debugger need not be mindful of such things. Problem is, this doesn't work. The idea was that we should do the syscall restart business first, so that the debugger captures that state. Otherwise, if the debugger for example saves the child's state, makes the child execute something else, then restores the saved state, we won't handle the syscall restart properly because we lose the "we're in a syscall" state. The code here worked for most cases, but if the debugger actually passes the signal through to the child unaltered, it's possible that we would do a syscall restart when we shouldn't have. In particular this breaks the case of debugging a process under a gdb which is being debugged by yet another gdb. gdb uses sigsuspend to wait for SIGCHLD of the inferior, but if gdb itself is being debugged by a top-level gdb we get a ptrace_stop(). The top-level gdb does a PTRACE_CONT with SIGCHLD to let the inferior gdb see the signal. But ptrace_signal_deliver() assumed the debugger would cancel out the signal and therefore did a syscall restart, because the return error was ERESTARTNOHAND. Fix this by simply making ptrace_signal_deliver() a nop, and providing a way for the debugger to control system call restarting properly: 1) Report a "in syscall" software bit in regs->{tstate,psr}. It is set early on in trap entry to a system call and is fully visible to the debugger via ptrace() and regsets. 2) Test this bit right before doing a syscall restart. We have to do a final recheck right after get_signal_to_deliver() in case the debugger cleared the bit during ptrace_stop(). 3) Clear the bit in trap return so we don't accidently try to set that bit in the real register. As a result we also get a ptrace_{is,clear}_syscall() for sparc32 just like sparc64 has. M68K has this same exact bug, and is now the only other user of the ptrace_signal_deliver hook. It needs to be fixed in the same exact way as sparc. Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
andn %t_psr, %g2, %t_psr
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
20 years ago
b ret_trap_unaligned_pc
add %sp, STACKFRAME_SZ, %o0
1:
LOAD_PT_YREG(sp, g1)
LOAD_PT_GLOBALS(sp)
wr %t_psr, 0x0, %psr
WRITE_PAUSE
jmp %t_pc
rett %t_npc
ret_trap_unaligned_pc:
ld [%sp + STACKFRAME_SZ + PT_PC], %o1
ld [%sp + STACKFRAME_SZ + PT_NPC], %o2
ld [%sp + STACKFRAME_SZ + PT_PSR], %o3
wr %t_wim, 0x0, %wim ! or else...
wr %t_psr, PSR_ET, %psr
WRITE_PAUSE
call do_memaccess_unaligned
nop
b signal_p
ld [%curptr + TI_FLAGS], %g2
ret_trap_kernel:
/* Will the rett land us in the invalid window? */
mov 2, %g1
sll %g1, %t_psr, %g1
rtrap_patch3: srl %g1, 8, %g2
or %g1, %g2, %g1
rd %wim, %g2
andcc %g2, %g1, %g0
be 1f ! Nope, just return from the trap
sll %g2, 0x1, %g1
/* We have to grab a window before returning. */
rtrap_patch4: srl %g2, 7, %g2
or %g1, %g2, %g1
rtrap_patch5: and %g1, 0xff, %g1
wr %g1, 0x0, %wim
/* Grrr, make sure we load from the right %sp... */
LOAD_PT_ALL(sp, t_psr, t_pc, t_npc, g1)
restore %g0, %g0, %g0
LOAD_WINDOW(sp)
b 2f
save %g0, %g0, %g0
/* Reload the entire frame in case this is from a
* kernel system call or whatever...
*/
1:
LOAD_PT_ALL(sp, t_psr, t_pc, t_npc, g1)
2:
sparc: Fix debugger syscall restart interactions. So, forever, we've had this ptrace_signal_deliver implementation which tries to handle all of the nasties that can occur when the debugger looks at a process about to take a signal. It's meant to address all of these issues inside of the kernel so that the debugger need not be mindful of such things. Problem is, this doesn't work. The idea was that we should do the syscall restart business first, so that the debugger captures that state. Otherwise, if the debugger for example saves the child's state, makes the child execute something else, then restores the saved state, we won't handle the syscall restart properly because we lose the "we're in a syscall" state. The code here worked for most cases, but if the debugger actually passes the signal through to the child unaltered, it's possible that we would do a syscall restart when we shouldn't have. In particular this breaks the case of debugging a process under a gdb which is being debugged by yet another gdb. gdb uses sigsuspend to wait for SIGCHLD of the inferior, but if gdb itself is being debugged by a top-level gdb we get a ptrace_stop(). The top-level gdb does a PTRACE_CONT with SIGCHLD to let the inferior gdb see the signal. But ptrace_signal_deliver() assumed the debugger would cancel out the signal and therefore did a syscall restart, because the return error was ERESTARTNOHAND. Fix this by simply making ptrace_signal_deliver() a nop, and providing a way for the debugger to control system call restarting properly: 1) Report a "in syscall" software bit in regs->{tstate,psr}. It is set early on in trap entry to a system call and is fully visible to the debugger via ptrace() and regsets. 2) Test this bit right before doing a syscall restart. We have to do a final recheck right after get_signal_to_deliver() in case the debugger cleared the bit during ptrace_stop(). 3) Clear the bit in trap return so we don't accidently try to set that bit in the real register. As a result we also get a ptrace_{is,clear}_syscall() for sparc32 just like sparc64 has. M68K has this same exact bug, and is now the only other user of the ptrace_signal_deliver hook. It needs to be fixed in the same exact way as sparc. Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago
sethi %hi(PSR_SYSCALL), %twin_tmp1
andn %t_psr, %twin_tmp1, %t_psr
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
20 years ago
wr %t_psr, 0x0, %psr
WRITE_PAUSE
jmp %t_pc
rett %t_npc
ret_trap_user_stack_is_bolixed:
wr %t_wim, 0x0, %wim
wr %t_psr, PSR_ET, %psr
WRITE_PAUSE
call window_ret_fault
add %sp, STACKFRAME_SZ, %o0
b signal_p
ld [%curptr + TI_FLAGS], %g2
sun4c_rett_stackchk:
be 1f
and %fp, 0xfff, %g1 ! delay slot
b ret_trap_user_stack_is_bolixed + 0x4
wr %t_wim, 0x0, %wim
/* See if we have to check the sanity of one page or two */
1:
add %g1, 0x38, %g1
sra %fp, 29, %g2
add %g2, 0x1, %g2
andncc %g2, 0x1, %g0
be 1f
andncc %g1, 0xff8, %g0
/* %sp is in vma hole, yuck */
b ret_trap_user_stack_is_bolixed + 0x4
wr %t_wim, 0x0, %wim
1:
be sun4c_rett_onepage /* Only one page to check */
lda [%fp] ASI_PTE, %g2
sun4c_rett_twopages:
add %fp, 0x38, %g1
sra %g1, 29, %g2
add %g2, 0x1, %g2
andncc %g2, 0x1, %g0
be 1f
lda [%g1] ASI_PTE, %g2
/* Second page is in vma hole */
b ret_trap_user_stack_is_bolixed + 0x4
wr %t_wim, 0x0, %wim
1:
srl %g2, 29, %g2
andcc %g2, 0x4, %g0
bne sun4c_rett_onepage
lda [%fp] ASI_PTE, %g2
/* Second page has bad perms */
b ret_trap_user_stack_is_bolixed + 0x4
wr %t_wim, 0x0, %wim
sun4c_rett_onepage:
srl %g2, 29, %g2
andcc %g2, 0x4, %g0
bne,a 1f
restore %g0, %g0, %g0
/* A page had bad page permissions, losing... */
b ret_trap_user_stack_is_bolixed + 0x4
wr %t_wim, 0x0, %wim
/* Whee, things are ok, load the window and continue. */
1:
LOAD_WINDOW(sp)
b ret_trap_userwins_ok
save %g0, %g0, %g0
.globl srmmu_rett_stackchk
srmmu_rett_stackchk:
bne ret_trap_user_stack_is_bolixed
sethi %hi(PAGE_OFFSET), %g1
cmp %g1, %fp
bleu ret_trap_user_stack_is_bolixed
mov AC_M_SFSR, %g1
lda [%g1] ASI_M_MMUREGS, %g0
lda [%g0] ASI_M_MMUREGS, %g1
or %g1, 0x2, %g1
sta %g1, [%g0] ASI_M_MMUREGS
restore %g0, %g0, %g0
LOAD_WINDOW(sp)
save %g0, %g0, %g0
andn %g1, 0x2, %g1
sta %g1, [%g0] ASI_M_MMUREGS
mov AC_M_SFAR, %g2
lda [%g2] ASI_M_MMUREGS, %g2
mov AC_M_SFSR, %g1
lda [%g1] ASI_M_MMUREGS, %g1
andcc %g1, 0x2, %g0
be ret_trap_userwins_ok
nop
b,a ret_trap_user_stack_is_bolixed