Re: [RFC PATCH v3 5/6] hurd: Make it possible to call memcpy very early

[Samuel Thibault]

Applied, thanks!

Sergey Bugaev, le sam. 29 avril 2023 23:18:21 +0300, a ecrit:
> Normally, in static builds, the first code that runs is _start, in e.g.
> sysdeps/x86_64/start.S, which quickly calls __libc_start_main, passing
> it the argv etc. Among the first things __libc_start_main does is
> initializing the tunables (based on env), then CPU features, and then
> calls _dl_relocate_static_pie (). Specifically, this runs ifunc
> resolvers to pick, based on the CPU features discovered earlier, the
> most suitable implementation of "string" functions such as memcpy.
> 
> Before that point, calling memcpy (or other ifunc-resolved functions)
> will not work.
> 
> In the Hurd port, things are more complex. In order to get argv/env for
> our process, glibc normally needs to do an RPC to the exec server,
> unless our args/env are already located on the stack (which is what
> happens to bootstrap processes spawned by GNU Mach). Fetching our
> argv/env from the exec server has to be done before the call to
> __libc_start_main, since we need to know what our argv/env are to pass
> them to __libc_start_main.
> 
> On the other hand, the implementation of the RPC (and other initial
> setup needed on the Hurd before __libc_start_main can be run) is not
> very trivial. In particular, it may (and on x86_64, will) use memcpy.
> But as described above, calling memcpy before __libc_start_main can not
> work, since the GOT entry for it is not yet initialized at that point.
> 
> Work around this by pre-filling the GOT entry with the baseline version
> of memcpy, __memcpy_sse2_unaligned. This makes it possible for early
> calls to memcpy to just work. The initial value of the GOT entry is
> unused on x86_64, and changing it won't interfere with the relocation
> being performed later: once _dl_relocate_static_pie () is called, the
> baseline version will get replaced with the most suitable one, and that
> is what subsequent calls of memcpy are going to call.
> 
> Checked on x86_64-gnu.
> 
> Signed-off-by: Sergey Bugaev <bugaevc@gmail.com>
> ---
> Changes since v1:
> - drop the stpncpy, since it's apparently not required during early
>   startup;
> - as a result of the above, there are no longer any changes to the
>   i386 version;
> - drop the PIC/non-PIC split, we can always use %rip-relative addressing
>   on x86_64;
> - as mentioned somewhere in the v1 thread, I have, since posting the v1,
>   actually gone and checked that the relocations do work and the proper,
>   more effecient memcpy version does get installed into the GOT slot and
>   invoked whenever anything calls memcpy;
> - convinced myself that this is not a terrible hack but rather an OK
>   solution;
> - worked out how this would be done on an architecture that (like i386,
>   unlike x86_64) does need the original value in the GOT to perform the
>   relocation, but (unlike i386, like x86_64) still uses an ifunc-selected
>   memcpy in static builds: namely, we'd simply put the original ifunc
>   address back into the GOT slot a few lines below, after the call to
>   _hurd_stack_setup.
> 
>  sysdeps/mach/hurd/x86_64/static-start.S | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/sysdeps/mach/hurd/x86_64/static-start.S 
> b/sysdeps/mach/hurd/x86_64/static-start.S
> index 982d3d52..cc8e2410 100644
> --- a/sysdeps/mach/hurd/x86_64/static-start.S
> +++ b/sysdeps/mach/hurd/x86_64/static-start.S
> @@ -19,6 +19,9 @@
>       .text
>       .globl _start
>  _start:
> +
> +     leaq __memcpy_sse2_unaligned(%rip), %rax
> +     movq %rax, memcpy@GOTPCREL(%rip)
>       call _hurd_stack_setup
>       xorq %rdx, %rdx
>       jmp _start1
> -- 
> 2.40.1
> 

-- 
Samuel
---
Pour une évaluation indépendante, transparente et rigoureuse !
Je soutiens la Commission d'Évaluation de l'Inria.