2008-01-30 20:32:49 +08:00
|
|
|
/*
|
2009-02-04 01:02:36 +08:00
|
|
|
|
|
|
|
|
x86 function call convention, 64-bit:
|
|
|
|
|
-------------------------------------
|
|
|
|
|
arguments | callee-saved | extra caller-saved | return
|
|
|
|
|
[callee-clobbered] | | [callee-clobbered] |
|
|
|
|
|
---------------------------------------------------------------------------
|
|
|
|
|
rdi rsi rdx rcx r8-9 | rbx rbp [*] r12-15 | r10-11 | rax, rdx [**]
|
|
|
|
|
|
|
|
|
|
( rsp is obviously invariant across normal function calls. (gcc can 'merge'
|
|
|
|
|
functions when it sees tail-call optimization possibilities) rflags is
|
|
|
|
|
clobbered. Leftover arguments are passed over the stack frame.)
|
|
|
|
|
|
|
|
|
|
[*] In the frame-pointers case rbp is fixed to the stack frame.
|
|
|
|
|
|
|
|
|
|
[**] for struct return values wider than 64 bits the return convention is a
|
|
|
|
|
bit more complex: up to 128 bits width we return small structures
|
|
|
|
|
straight in rax, rdx. For structures larger than that (3 words or
|
|
|
|
|
larger) the caller puts a pointer to an on-stack return struct
|
|
|
|
|
[allocated in the caller's stack frame] into the first argument - i.e.
|
|
|
|
|
into rdi. All other arguments shift up by one in this case.
|
|
|
|
|
Fortunately this case is rare in the kernel.
|
|
|
|
|
|
|
|
|
|
For 32-bit we have the following conventions - kernel is built with
|
|
|
|
|
-mregparm=3 and -freg-struct-return:
|
|
|
|
|
|
|
|
|
|
x86 function calling convention, 32-bit:
|
|
|
|
|
----------------------------------------
|
|
|
|
|
arguments | callee-saved | extra caller-saved | return
|
|
|
|
|
[callee-clobbered] | | [callee-clobbered] |
|
|
|
|
|
-------------------------------------------------------------------------
|
|
|
|
|
eax edx ecx | ebx edi esi ebp [*] | <none> | eax, edx [**]
|
|
|
|
|
|
|
|
|
|
( here too esp is obviously invariant across normal function calls. eflags
|
|
|
|
|
is clobbered. Leftover arguments are passed over the stack frame. )
|
|
|
|
|
|
|
|
|
|
[*] In the frame-pointers case ebp is fixed to the stack frame.
|
|
|
|
|
|
|
|
|
|
[**] We build with -freg-struct-return, which on 32-bit means similar
|
|
|
|
|
semantics as on 64-bit: edx can be used for a second return value
|
|
|
|
|
(i.e. covering integer and structure sizes up to 64 bits) - after that
|
|
|
|
|
it gets more complex and more expensive: 3-word or larger struct returns
|
|
|
|
|
get done in the caller's frame and the pointer to the return struct goes
|
|
|
|
|
into regparm0, i.e. eax - the other arguments shift up and the
|
|
|
|
|
function's register parameters degenerate to regparm=2 in essence.
|
|
|
|
|
|
|
|
|
|
*/
|
|
|
|
|
|
2012-10-03 01:01:25 +08:00
|
|
|
#include <asm/dwarf2.h>
|
2009-02-04 01:02:36 +08:00
|
|
|
|
2013-08-14 20:51:00 +08:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
|
|
2009-02-04 01:02:36 +08:00
|
|
|
/*
|
2012-09-26 16:28:22 +08:00
|
|
|
* 64-bit system call stack frame layout defines and helpers,
|
|
|
|
|
* for assembly code:
|
2008-01-30 20:32:49 +08:00
|
|
|
*/
|
2005-04-17 06:20:36 +08:00
|
|
|
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
/* The layout forms the "struct pt_regs" on the stack: */
|
|
|
|
|
/*
|
|
|
|
|
* C ABI says these regs are callee-preserved. They aren't saved on kernel entry
|
|
|
|
|
* unless syscall needs a complete, fully filled "struct pt_regs".
|
|
|
|
|
*/
|
|
|
|
|
#define R15 0*8
|
|
|
|
|
#define R14 1*8
|
|
|
|
|
#define R13 2*8
|
|
|
|
|
#define R12 3*8
|
|
|
|
|
#define RBP 4*8
|
|
|
|
|
#define RBX 5*8
|
|
|
|
|
/* These regs are callee-clobbered. Always saved on kernel entry. */
|
|
|
|
|
#define R11 6*8
|
|
|
|
|
#define R10 7*8
|
|
|
|
|
#define R9 8*8
|
|
|
|
|
#define R8 9*8
|
|
|
|
|
#define RAX 10*8
|
|
|
|
|
#define RCX 11*8
|
|
|
|
|
#define RDX 12*8
|
|
|
|
|
#define RSI 13*8
|
|
|
|
|
#define RDI 14*8
|
|
|
|
|
/*
|
|
|
|
|
* On syscall entry, this is syscall#. On CPU exception, this is error code.
|
|
|
|
|
* On hw interrupt, it's IRQ number:
|
|
|
|
|
*/
|
|
|
|
|
#define ORIG_RAX 15*8
|
|
|
|
|
/* Return frame for iretq */
|
|
|
|
|
#define RIP 16*8
|
|
|
|
|
#define CS 17*8
|
|
|
|
|
#define EFLAGS 18*8
|
|
|
|
|
#define RSP 19*8
|
|
|
|
|
#define SS 20*8
|
|
|
|
|
|
2015-02-27 06:40:36 +08:00
|
|
|
#define SIZEOF_PTREGS 21*8
|
|
|
|
|
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.macro ALLOC_PT_GPREGS_ON_STACK addskip=0
|
|
|
|
|
subq $15*8+\addskip, %rsp
|
|
|
|
|
CFI_ADJUST_CFA_OFFSET 15*8+\addskip
|
|
|
|
|
.endm
|
2011-06-01 04:21:51 +08:00
|
|
|
|
2015-03-10 02:39:21 +08:00
|
|
|
.macro SAVE_C_REGS_HELPER offset=0 rax=1 rcx=1 r8910=1 r11=1
|
|
|
|
|
.if \r11
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
movq_cfi r11, 6*8+\offset
|
2015-03-10 02:39:21 +08:00
|
|
|
.endif
|
|
|
|
|
.if \r8910
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
movq_cfi r10, 7*8+\offset
|
|
|
|
|
movq_cfi r9, 8*8+\offset
|
|
|
|
|
movq_cfi r8, 9*8+\offset
|
2014-09-06 06:13:55 +08:00
|
|
|
.endif
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.if \rax
|
|
|
|
|
movq_cfi rax, 10*8+\offset
|
|
|
|
|
.endif
|
|
|
|
|
.if \rcx
|
|
|
|
|
movq_cfi rcx, 11*8+\offset
|
2005-04-17 06:20:36 +08:00
|
|
|
.endif
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
movq_cfi rdx, 12*8+\offset
|
|
|
|
|
movq_cfi rsi, 13*8+\offset
|
|
|
|
|
movq_cfi rdi, 14*8+\offset
|
|
|
|
|
.endm
|
|
|
|
|
.macro SAVE_C_REGS offset=0
|
2015-03-10 02:39:21 +08:00
|
|
|
SAVE_C_REGS_HELPER \offset, 1, 1, 1, 1
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.endm
|
|
|
|
|
.macro SAVE_C_REGS_EXCEPT_RAX_RCX offset=0
|
2015-03-10 02:39:21 +08:00
|
|
|
SAVE_C_REGS_HELPER \offset, 0, 0, 1, 1
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.endm
|
|
|
|
|
.macro SAVE_C_REGS_EXCEPT_R891011
|
2015-03-10 02:39:21 +08:00
|
|
|
SAVE_C_REGS_HELPER 0, 1, 1, 0, 0
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.endm
|
|
|
|
|
.macro SAVE_C_REGS_EXCEPT_RCX_R891011
|
2015-03-10 02:39:21 +08:00
|
|
|
SAVE_C_REGS_HELPER 0, 1, 0, 0, 0
|
|
|
|
|
.endm
|
|
|
|
|
.macro SAVE_C_REGS_EXCEPT_RAX_RCX_R11
|
|
|
|
|
SAVE_C_REGS_HELPER 0, 0, 0, 1, 0
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.endm
|
2011-06-01 04:21:51 +08:00
|
|
|
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.macro SAVE_EXTRA_REGS offset=0
|
|
|
|
|
movq_cfi r15, 0*8+\offset
|
|
|
|
|
movq_cfi r14, 1*8+\offset
|
|
|
|
|
movq_cfi r13, 2*8+\offset
|
|
|
|
|
movq_cfi r12, 3*8+\offset
|
|
|
|
|
movq_cfi rbp, 4*8+\offset
|
|
|
|
|
movq_cfi rbx, 5*8+\offset
|
|
|
|
|
.endm
|
|
|
|
|
.macro SAVE_EXTRA_REGS_RBP offset=0
|
|
|
|
|
movq_cfi rbp, 4*8+\offset
|
2005-04-17 06:20:36 +08:00
|
|
|
.endm
|
|
|
|
|
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.macro RESTORE_EXTRA_REGS offset=0
|
|
|
|
|
movq_cfi_restore 0*8+\offset, r15
|
|
|
|
|
movq_cfi_restore 1*8+\offset, r14
|
|
|
|
|
movq_cfi_restore 2*8+\offset, r13
|
|
|
|
|
movq_cfi_restore 3*8+\offset, r12
|
|
|
|
|
movq_cfi_restore 4*8+\offset, rbp
|
|
|
|
|
movq_cfi_restore 5*8+\offset, rbx
|
|
|
|
|
.endm
|
2008-01-30 20:32:49 +08:00
|
|
|
|
2015-04-05 02:55:19 +08:00
|
|
|
.macro ZERO_EXTRA_REGS
|
|
|
|
|
xorl %r15d, %r15d
|
|
|
|
|
xorl %r14d, %r14d
|
|
|
|
|
xorl %r13d, %r13d
|
|
|
|
|
xorl %r12d, %r12d
|
|
|
|
|
xorl %ebp, %ebp
|
|
|
|
|
xorl %ebx, %ebx
|
|
|
|
|
.endm
|
|
|
|
|
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.macro RESTORE_C_REGS_HELPER rstor_rax=1, rstor_rcx=1, rstor_r11=1, rstor_r8910=1, rstor_rdx=1
|
2011-06-01 04:21:53 +08:00
|
|
|
.if \rstor_r11
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
movq_cfi_restore 6*8, r11
|
2005-04-17 06:20:36 +08:00
|
|
|
.endif
|
2011-06-01 04:21:53 +08:00
|
|
|
.if \rstor_r8910
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
movq_cfi_restore 7*8, r10
|
|
|
|
|
movq_cfi_restore 8*8, r9
|
|
|
|
|
movq_cfi_restore 9*8, r8
|
2005-04-17 06:20:36 +08:00
|
|
|
.endif
|
2011-06-01 04:21:53 +08:00
|
|
|
.if \rstor_rax
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
movq_cfi_restore 10*8, rax
|
2005-04-17 06:20:36 +08:00
|
|
|
.endif
|
2011-06-01 04:21:53 +08:00
|
|
|
.if \rstor_rcx
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
movq_cfi_restore 11*8, rcx
|
2005-04-17 06:20:36 +08:00
|
|
|
.endif
|
2011-06-01 04:21:53 +08:00
|
|
|
.if \rstor_rdx
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
movq_cfi_restore 12*8, rdx
|
2005-04-17 06:20:36 +08:00
|
|
|
.endif
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
movq_cfi_restore 13*8, rsi
|
|
|
|
|
movq_cfi_restore 14*8, rdi
|
2008-01-30 20:32:49 +08:00
|
|
|
.endm
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.macro RESTORE_C_REGS
|
|
|
|
|
RESTORE_C_REGS_HELPER 1,1,1,1,1
|
2005-04-17 06:20:36 +08:00
|
|
|
.endm
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.macro RESTORE_C_REGS_EXCEPT_RAX
|
|
|
|
|
RESTORE_C_REGS_HELPER 0,1,1,1,1
|
2008-01-30 20:32:49 +08:00
|
|
|
.endm
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.macro RESTORE_C_REGS_EXCEPT_RCX
|
|
|
|
|
RESTORE_C_REGS_HELPER 1,0,1,1,1
|
2005-04-17 06:20:36 +08:00
|
|
|
.endm
|
2015-02-27 06:40:38 +08:00
|
|
|
.macro RESTORE_C_REGS_EXCEPT_R11
|
|
|
|
|
RESTORE_C_REGS_HELPER 1,1,0,1,1
|
|
|
|
|
.endm
|
2015-03-10 02:39:21 +08:00
|
|
|
.macro RESTORE_C_REGS_EXCEPT_RCX_R11
|
|
|
|
|
RESTORE_C_REGS_HELPER 1,0,0,1,1
|
|
|
|
|
.endm
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.macro RESTORE_RSI_RDI
|
|
|
|
|
RESTORE_C_REGS_HELPER 0,0,0,0,0
|
|
|
|
|
.endm
|
|
|
|
|
.macro RESTORE_RSI_RDI_RDX
|
|
|
|
|
RESTORE_C_REGS_HELPER 0,0,0,0,1
|
2005-04-17 06:20:36 +08:00
|
|
|
.endm
|
2008-01-30 20:32:49 +08:00
|
|
|
|
x86/asm/entry/64: Always allocate a complete "struct pt_regs" on the kernel stack
The 64-bit entry code was using six stack slots less by not
saving/restoring registers which are callee-preserved according
to the C ABI, and was not allocating space for them.
Only when syscalls needed a complete "struct pt_regs" was
the complete area allocated and filled in.
As an additional twist, on interrupt entry a "slightly less
truncated pt_regs" trick is used, to make nested interrupt
stacks easier to unwind.
This proved to be a source of significant obfuscation and subtle
bugs. For example, 'stub_fork' had to pop the return address,
extend the struct, save registers, and push return address back.
Ugly. 'ia32_ptregs_common' pops return address and "returns" via
jmp insn, throwing a wrench into CPU return stack cache.
This patch changes the code to always allocate a complete
"struct pt_regs" on the kernel stack. The saving of registers
is still done lazily.
"Partial pt_regs" trick on interrupt stack is retained.
Macros which manipulate "struct pt_regs" on stack are reworked:
- ALLOC_PT_GPREGS_ON_STACK allocates the structure.
- SAVE_C_REGS saves to it those registers which are clobbered
by C code.
- SAVE_EXTRA_REGS saves to it all other registers.
- Corresponding RESTORE_* and REMOVE_PT_GPREGS_FROM_STACK macros
reverse it.
'ia32_ptregs_common', 'stub_fork' and friends lost their ugly dance
with the return pointer.
LOAD_ARGS32 in ia32entry.S now uses symbolic stack offsets
instead of magic numbers.
'error_entry' and 'save_paranoid' now use SAVE_C_REGS +
SAVE_EXTRA_REGS instead of having it open-coded yet again.
Patch was run-tested: 64-bit executables, 32-bit executables,
strace works.
Timing tests did not show measurable difference in 32-bit
and 64-bit syscalls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Drewry <wad@chromium.org>
Link: http://lkml.kernel.org/r/1423778052-21038-2-git-send-email-dvlasenk@redhat.com
Link: http://lkml.kernel.org/r/b89763d354aa23e670b9bdf3a40ae320320a7c2e.1424989793.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-02-27 06:40:27 +08:00
|
|
|
.macro REMOVE_PT_GPREGS_FROM_STACK addskip=0
|
|
|
|
|
addq $15*8+\addskip, %rsp
|
|
|
|
|
CFI_ADJUST_CFA_OFFSET -(15*8+\addskip)
|
2005-04-17 06:20:36 +08:00
|
|
|
.endm
|
|
|
|
|
|
|
|
|
|
.macro icebp
|
|
|
|
|
.byte 0xf1
|
|
|
|
|
.endm
|
2013-08-14 20:51:00 +08:00
|
|
|
|
|
|
|
|
#else /* CONFIG_X86_64 */
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* For 32bit only simplified versions of SAVE_ALL/RESTORE_ALL. These
|
|
|
|
|
* are different from the entry_32.S versions in not changing the segment
|
|
|
|
|
* registers. So only suitable for in kernel use, not when transitioning
|
|
|
|
|
* from or to user space. The resulting stack frame is not a standard
|
|
|
|
|
* pt_regs frame. The main use case is calling C code from assembler
|
|
|
|
|
* when all the registers need to be preserved.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
.macro SAVE_ALL
|
2015-02-27 06:40:25 +08:00
|
|
|
pushl_cfi_reg eax
|
|
|
|
|
pushl_cfi_reg ebp
|
|
|
|
|
pushl_cfi_reg edi
|
|
|
|
|
pushl_cfi_reg esi
|
|
|
|
|
pushl_cfi_reg edx
|
|
|
|
|
pushl_cfi_reg ecx
|
|
|
|
|
pushl_cfi_reg ebx
|
2013-08-14 20:51:00 +08:00
|
|
|
.endm
|
|
|
|
|
|
|
|
|
|
.macro RESTORE_ALL
|
2015-02-27 06:40:25 +08:00
|
|
|
popl_cfi_reg ebx
|
|
|
|
|
popl_cfi_reg ecx
|
|
|
|
|
popl_cfi_reg edx
|
|
|
|
|
popl_cfi_reg esi
|
|
|
|
|
popl_cfi_reg edi
|
|
|
|
|
popl_cfi_reg ebp
|
|
|
|
|
popl_cfi_reg eax
|
2013-08-14 20:51:00 +08:00
|
|
|
.endm
|
|
|
|
|
|
|
|
|
|
#endif /* CONFIG_X86_64 */
|
|
|
|
|
|
