Re: x86 - clang / objtool status

From: Nick Desaulniers <ndesaulniers@google.com>
To: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>,
	LKML <linux-kernel@vger.kernel.org>,
	Nathan Chancellor <natechancellor@gmail.com>,
	clang-built-linux <clang-built-linux@googlegroups.com>,
	"maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT)"
	<x86@kernel.org>, Arnd Bergmann <arnd@arndb.de>,
	Sedat Dilek <sedat.dilek@gmail.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: x86 - clang / objtool status
Date: Wed, 24 Jul 2019 10:22:53 -0700	[thread overview]
Message-ID: <CAKwvOdk7d2UupG66Zt0zN_NB0xQOaCWBvMMWY_N_-uejw69-BQ@mail.gmail.com> (raw)
In-Reply-To: <20190724023946.yxsz5im22fz4zxrn@treble>

On Tue, Jul 23, 2019 at 7:43 PM Josh Poimboeuf <jpoimboe@redhat.com> wrote:
>
> On Thu, Jul 18, 2019 at 10:40:09PM +0200, Thomas Gleixner wrote:
>
> >   drivers/gpu/drm/i915/gem/i915_gem_execbuffer.o: warning: objtool: .altinstr_replacement+0x86: redundant UACCESS disable
>
> Looking at this one, I think I agree with objtool.
>
> PeterZ, Linus, I know y'all discussed this code a few months ago.
>
> __copy_from_user() already does a CLAC in its error path.  So isn't the
> user_access_end() redundant for the __copy_from_user() error path?
>
> Untested fix:
>
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
> index 5fae0e50aad0..41dab9ea33cd 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
> @@ -1628,6 +1628,7 @@ static int check_relocations(const struct drm_i915_gem_exec_object2 *entry)
>
>  static int eb_copy_relocations(const struct i915_execbuffer *eb)
>  {
> +       struct drm_i915_gem_relocation_entry *relocs;
>         const unsigned int count = eb->buffer_count;
>         unsigned int i;
>         int err;
> @@ -1635,7 +1636,6 @@ static int eb_copy_relocations(const struct i915_execbuffer *eb)
>         for (i = 0; i < count; i++) {
>                 const unsigned int nreloc = eb->exec[i].relocation_count;
>                 struct drm_i915_gem_relocation_entry __user *urelocs;
> -               struct drm_i915_gem_relocation_entry *relocs;
>                 unsigned long size;
>                 unsigned long copied;
>
> @@ -1663,14 +1663,8 @@ static int eb_copy_relocations(const struct i915_execbuffer *eb)
>
>                         if (__copy_from_user((char *)relocs + copied,
>                                              (char __user *)urelocs + copied,
> -                                            len)) {
> -end_user:
> -                               user_access_end();
> -end:
> -                               kvfree(relocs);
> -                               err = -EFAULT;
> -                               goto err;
> -                       }
> +                                            len))
> +                               goto end;
>
>                         copied += len;
>                 } while (copied < size);
> @@ -1699,10 +1693,14 @@ static int eb_copy_relocations(const struct i915_execbuffer *eb)
>
>         return 0;
>
> +end_user:
> +       user_access_end();
> +end:
> +       kvfree(relocs);
> +       err = -EFAULT;
>  err:
>         while (i--) {
> -               struct drm_i915_gem_relocation_entry *relocs =
> -                       u64_to_ptr(typeof(*relocs), eb->exec[i].relocs_ptr);
> +               relocs = u64_to_ptr(typeof(*relocs), eb->exec[i].relocs_ptr);
>                 if (eb->exec[i].relocation_count)
>                         kvfree(relocs);
>         }

Thanks for this patch.
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Reported-by: Sedat Dilek <sedat.dilek@gmail.com>
Link: https://github.com/ClangBuiltLinux/linux/issues/617
Fixes: 2889caa92321 ("drm/i915: Eliminate lots of iterations over the
execobjects array")

-- 
Thanks,
~Nick Desaulniers