Hi,
There are 3 patches in this series.
1) Trivial clean up in entry_64.S.
2) Add comment about registers on exit in entry_64.S.
3) Add documentation about registers on entry and exit.
(2) and (3) are based on the discussion we had at:
https://lore.kernel.org/lkml/[email protected]/
This series is based on commit:
24556728c305886b8bb05bf2ac7e20cf7db3e314 ("Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm")
Thanks!
Signed-off-by: Ammar Faizi <[email protected]>
---
Ammar Faizi (3):
x86/entry/64: Clean up spaces after instruction
x86/entry/64: Add info about registers on exit
Documentation: x86-64: Document registers on entry and exit
Documentation/x86/entry_64.rst | 47 ++++++++++++++++++++++++++++++++++
arch/x86/entry/entry_64.S | 27 ++++++++++++++-----
2 files changed, 67 insertions(+), 7 deletions(-)
base-commit: 24556728c305886b8bb05bf2ac7e20cf7db3e314
--
Ammar Faizi
There was a controversial discussion about the wording in the System
V ABI document regarding what registers the kernel is allowed to
clobber when the userspace executes syscall.
The resolution of the discussion was reviewing the clobber list in
the glibc source. For a historical reason in the glibc source, the
kernel must restore all registers before returning to the userspace
(except for rax, rcx and r11).
On Wed, 13 Oct 2021 at 16:24:28 +0000, Michael Matz <[email protected]> wrote:
> It might also be interesting to know that while the wording of the psABI
> was indeed intended to imply that all argument registers are potentially
> clobbered (like with normal calls) glibc's inline assembler to call
> syscalls relies on most registers to actually be preserved:
>
> # define REGISTERS_CLOBBERED_BY_SYSCALL "cc", "r11", "cx"
> ...
> #define internal_syscall6(number, arg1, arg2, arg3, arg4, arg5, arg6) \
> ({ \
> unsigned long int resultvar; \
> TYPEFY (arg6, __arg6) = ARGIFY (arg6); \
> TYPEFY (arg5, __arg5) = ARGIFY (arg5); \
> TYPEFY (arg4, __arg4) = ARGIFY (arg4); \
> TYPEFY (arg3, __arg3) = ARGIFY (arg3); \
> TYPEFY (arg2, __arg2) = ARGIFY (arg2); \
> TYPEFY (arg1, __arg1) = ARGIFY (arg1); \
> register TYPEFY (arg6, _a6) asm ("r9") = __arg6; \
> register TYPEFY (arg5, _a5) asm ("r8") = __arg5; \
> register TYPEFY (arg4, _a4) asm ("r10") = __arg4; \
> register TYPEFY (arg3, _a3) asm ("rdx") = __arg3; \
> register TYPEFY (arg2, _a2) asm ("rsi") = __arg2; \
> register TYPEFY (arg1, _a1) asm ("rdi") = __arg1; \
> asm volatile ( \
> "syscall\n\t" \
> : "=a" (resultvar) \
> : "0" (number), "r" (_a1), "r" (_a2), "r" (_a3), "r" (_a4), \
> "r" (_a5), "r" (_a6) \
> : "memory", REGISTERS_CLOBBERED_BY_SYSCALL); \
> (long int) resultvar; \
> })
>
>
> Note in particular the missing clobbers or outputs of any of the argument
> regs.
>
> So, even though the psABI (might have) meant something else, as glibc is
> doing the above we in fact have a de-facto standard that the kernel can't
> clobber any of the argument regs. The wording and the linux x86-64
> syscall implementation (and use in glibc) all come from the same time in
> 2001, so there never was a time when the kernel was not saving/restoring
> the arg registers, so it can't stop now.
>
> In effect this means the psABI should be clarified to explicitely say the
> the arg registers aren't clobbered, i.e. that the mentioned list of
> clobbered regs isn't inclusive but exclusive. I will do that.
>
> When I was discussing this with Boris earlier I hadn't yet looked at glibc
> use but only gave my interpretation from memory and reading. Obviously
> reality trumps anything like that :-)
Link: https://lore.kernel.org/lkml/[email protected]/
Link: https://gitlab.com/x86-psABIs/x86-64-ABI/-/merge_requests/25
This documents "registers on entry" and "registers on exit".
Cc: Andy Lutomirski <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Borislav Petkov <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: "H. Peter Anvin" <[email protected]>
Cc: Michael Matz <[email protected]>
Cc: "H.J. Lu" <[email protected]>
Cc: Jonathan Corbet <[email protected]>
Cc: Willy Tarreau <[email protected]>
Cc: x86-ml <[email protected]>
Cc: lkml <[email protected]>
Cc: GNU/Weeb Mailing List <[email protected]>
Signed-off-by: Ammar Faizi <[email protected]>
---
Documentation/x86/entry_64.rst | 47 ++++++++++++++++++++++++++++++++++
1 file changed, 47 insertions(+)
diff --git a/Documentation/x86/entry_64.rst b/Documentation/x86/entry_64.rst
index e433e08f7018..3f2007e2a938 100644
--- a/Documentation/x86/entry_64.rst
+++ b/Documentation/x86/entry_64.rst
@@ -108,3 +108,50 @@ We try to only use IST entries and the paranoid entry code for vectors
that absolutely need the more expensive check for the GS base - and we
generate all 'normal' entry points with the regular (faster) paranoid=0
variant.
+
+
+Registers on entry:
+-------------------
+
+ - rax system call number
+ - rcx return address
+ - r11 saved rflags (note: r11 is callee-clobbered register in C ABI)
+ - rdi arg0
+ - rsi arg1
+ - rdx arg2
+ - r10 arg3 (needs to be moved to rcx to conform to C ABI)
+ - r8 arg4
+ - r9 arg5
+
+
+Registers on exit:
+------------------
+
+This explanation is adapted from a controversial discussion about the
+wording in the System V ABI document regarding what registers the
+kernel is allowed to clobber when the userspace executes syscall.
+
+The resolution of the discussion was reviewing the clobber list in the
+glibc source. For a historical reason in the glibc source, the kernel
+must restore all registers before returning to the userspace (except
+for rax, rcx and r11).
+
+For the detailed story, see the email from Michael Matz:
+
+https://lore.kernel.org/lkml/[email protected]/
+
+The kernel saves/restores all registers, so the userspace using it can
+assume the value of all registers are still intact, except for the rax,
+rcx, and r11.
+
+When the kernel returns to the userspace, only 3 registers are
+clobbered:
+
+ - rax syscall return value
+ - rcx return address
+ - r11 rflags
+
+The kernel syscall entry actually also saves/restores rcx and r11, but
+at that point, they have already been overwritten by the syscall
+instruction itself with the userspace rip and rflags value, they will
+be used by the sysret instruction when returning to the userspace.
--
2.32.0
On 1/7/22 15:52, Ammar Faizi wrote:
> There was a controversial discussion about the wording in the System
> V ABI document regarding what registers the kernel is allowed to
> clobber when the userspace executes syscall.
>
> The resolution of the discussion was reviewing the clobber list in
> the glibc source. For a historical reason in the glibc source, the
> kernel must restore all registers before returning to the userspace
> (except for rax, rcx and r11).
>
> On Wed, 13 Oct 2021 at 16:24:28 +0000, Michael Matz <[email protected]> wrote:
>> It might also be interesting to know that while the wording of the psABI
>> was indeed intended to imply that all argument registers are potentially
>> clobbered (like with normal calls) glibc's inline assembler to call
>> syscalls relies on most registers to actually be preserved:
>>
>> # define REGISTERS_CLOBBERED_BY_SYSCALL "cc", "r11", "cx"
>> ...
>> #define internal_syscall6(number, arg1, arg2, arg3, arg4, arg5, arg6) \
>> ({ \
>> unsigned long int resultvar; \
>> TYPEFY (arg6, __arg6) = ARGIFY (arg6); \
>> TYPEFY (arg5, __arg5) = ARGIFY (arg5); \
>> TYPEFY (arg4, __arg4) = ARGIFY (arg4); \
>> TYPEFY (arg3, __arg3) = ARGIFY (arg3); \
>> TYPEFY (arg2, __arg2) = ARGIFY (arg2); \
>> TYPEFY (arg1, __arg1) = ARGIFY (arg1); \
>> register TYPEFY (arg6, _a6) asm ("r9") = __arg6; \
>> register TYPEFY (arg5, _a5) asm ("r8") = __arg5; \
>> register TYPEFY (arg4, _a4) asm ("r10") = __arg4; \
>> register TYPEFY (arg3, _a3) asm ("rdx") = __arg3; \
>> register TYPEFY (arg2, _a2) asm ("rsi") = __arg2; \
>> register TYPEFY (arg1, _a1) asm ("rdi") = __arg1; \
>> asm volatile ( \
>> "syscall\n\t" \
>> : "=a" (resultvar) \
>> : "0" (number), "r" (_a1), "r" (_a2), "r" (_a3), "r" (_a4), \
>> "r" (_a5), "r" (_a6) \
>> : "memory", REGISTERS_CLOBBERED_BY_SYSCALL); \
>> (long int) resultvar; \
>> })
>>
>>
>> Note in particular the missing clobbers or outputs of any of the argument
>> regs.
>>
>> So, even though the psABI (might have) meant something else, as glibc is
>> doing the above we in fact have a de-facto standard that the kernel can't
>> clobber any of the argument regs. The wording and the linux x86-64
>> syscall implementation (and use in glibc) all come from the same time in
>> 2001, so there never was a time when the kernel was not saving/restoring
>> the arg registers, so it can't stop now.
>>
>> In effect this means the psABI should be clarified to explicitely say the
>> the arg registers aren't clobbered, i.e. that the mentioned list of
>> clobbered regs isn't inclusive but exclusive. I will do that.
>>
>> When I was discussing this with Boris earlier I hadn't yet looked at glibc
>> use but only gave my interpretation from memory and reading. Obviously
>> reality trumps anything like that :-)
>
> Link: https://lore.kernel.org/lkml/[email protected]/
> Link: https://gitlab.com/x86-psABIs/x86-64-ABI/-/merge_requests/25
>
> This documents "registers on entry" and "registers on exit".
>
> Cc: Andy Lutomirski <[email protected]>
> Cc: Thomas Gleixner <[email protected]>
> Cc: Ingo Molnar <[email protected]>
> Cc: Borislav Petkov <[email protected]>
> Cc: Dave Hansen <[email protected]>
> Cc: "H. Peter Anvin" <[email protected]>
> Cc: Michael Matz <[email protected]>
> Cc: "H.J. Lu" <[email protected]>
> Cc: Jonathan Corbet <[email protected]>
> Cc: Willy Tarreau <[email protected]>
> Cc: x86-ml <[email protected]>
> Cc: lkml <[email protected]>
> Cc: GNU/Weeb Mailing List <[email protected]>
> Signed-off-by: Ammar Faizi <[email protected]>
> ---
> Documentation/x86/entry_64.rst | 47 ++++++++++++++++++++++++++++++++++
> 1 file changed, 47 insertions(+)
>
> diff --git a/Documentation/x86/entry_64.rst b/Documentation/x86/entry_64.rst
> index e433e08f7018..3f2007e2a938 100644
> --- a/Documentation/x86/entry_64.rst
> +++ b/Documentation/x86/entry_64.rst
> @@ -108,3 +108,50 @@ We try to only use IST entries and the paranoid entry code for vectors
> that absolutely need the more expensive check for the GS base - and we
> generate all 'normal' entry points with the regular (faster) paranoid=0
> variant.
> +
> +
> +Registers on entry:
> +-------------------
This is SYSCALL64 registers on entry, not general registers on entry.
Also, this has little to do with the entry logic, so it probably doesn't
belong in this file.
On 1/8/22 7:02 AM, Andy Lutomirski wrote:
> On 1/7/22 15:52, Ammar Faizi wrote:
>> There was a controversial discussion about the wording in the System
>> V ABI document regarding what registers the kernel is allowed to
>> clobber when the userspace executes syscall.
>>
>> The resolution of the discussion was reviewing the clobber list in
>> the glibc source. For a historical reason in the glibc source, the
>> kernel must restore all registers before returning to the userspace
>> (except for rax, rcx and r11).
>>
[...]
>> diff --git a/Documentation/x86/entry_64.rst b/Documentation/x86/entry_64.rst
>> index e433e08f7018..3f2007e2a938 100644
>> --- a/Documentation/x86/entry_64.rst
>> +++ b/Documentation/x86/entry_64.rst
>> @@ -108,3 +108,50 @@ We try to only use IST entries and the paranoid entry code for vectors
>> that absolutely need the more expensive check for the GS base - and we
>> generate all 'normal' entry points with the regular (faster) paranoid=0
>> variant.
>> +
>> +
>> +Registers on entry:
>> +-------------------
>
> This is SYSCALL64 registers on entry, not general registers on entry.
> Also, this has little to do with the entry logic, so it probably
> doesn't belong in this file.
Ah right, I should be more specific saying it's for syscall64 entry
as there are 6 entries mentioned in this document.
Should syscall64 entry topic be documented? If not I will drop it,
otherwise suggest me a place for it.
I think we can document it here, but it needs to be more specific
saying it's for syscall64 entry.
Jonathan?
--
Ammar Faizi
On Fri, Jan 07, 2022 at 04:02:27PM -0800, Andy Lutomirski wrote:
> This is SYSCALL64 registers on entry, not general registers on entry. Also,
> this has little to do with the entry logic, so it probably doesn't belong in
> this file.
Right, except that syscall is also a kernel entry point so it kinda
belongs in a documentation file called "entry". :)
Srsly, I'd really like to keep the note about which registers glibc
considers clobbered and which not, documented somewhere as that is
practically an ABI which is not (yet) in the psABI doc.
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette