Re: [PATCH 00/21] objtool: vmlinux.o and CLANG LTO support

[Josh Poimboeuf]

On Thu, Jan 14, 2021 at 04:41:28PM -0800, Sami Tolvanen wrote:
> Hi Josh,
> 
> On Thu, Jan 14, 2021 at 11:40 AM Josh Poimboeuf <jpoimboe@xxxxxxxxxx> wrote:
> >
> > Add support for proper vmlinux.o validation, which will be needed for
> > Sami's upcoming x86 LTO set.  (And vmlinux validation is the future for
> > objtool anyway, for other reasons.)
> >
> > This isn't 100% done -- most notably, crypto still needs to be supported
> > -- but I think this gets us most of the way there.
> >
> > This can also be found at
> >
> >   git://git.kernel.org/pub/scm/linux/kernel/git/jpoimboe/linux.git objtool-vmlinux
> >
> > And for more testing it can be combined with Sami's x86 LTO patches:
> >
> >   https://github.com/samitolvanen/linux clang-lto
> 
> Thank you for sending these! I applied this series on top of the
> clang-lto tree and built allyesconfig with LTO_CLANG enabled and the
> following crypto options disabled:
> 
> CRYPTO_AES_NI_INTEL
> CRYPTO_CAMELLIA_AESNI_AVX2_X86_64
> CRYPTO_SHA1_SSSE3
> CRYPTO_SHA256_SSSE3
> CRYPTO_SHA512_SSSE3
> CRYPTO_CRC32C_INTEL
> 
> I can confirm that all the warnings I previously saw are now fixed,
> but I'm seeing a few new ones:
> 
> vmlinux.o: warning: objtool: balance_leaf_when_delete()+0x17d4: stack
> state mismatch: cfa1=7+192 cfa2=7+176
> vmlinux.o: warning: objtool: internal_move_pointers_items()+0x9f7:
> stack state mismatch: cfa1=7+160 cfa2=7+176
> vmlinux.o: warning: objtool: strncpy_from_user()+0x181: call to
> do_strncpy_from_user() with UACCESS enabled
> vmlinux.o: warning: objtool: strnlen_user()+0x12b: call to
> do_strnlen_user() with UACCESS enabled
> vmlinux.o: warning: objtool: i915_gem_execbuffer2_ioctl()+0x390: call
> to __ubsan_handle_negate_overflow() with UACCESS enabled
> vmlinux.o: warning: objtool: .text.snd_trident_free_voice: unexpected
> end of section
> 
> I haven't had a chance to take a closer look yet, but some of these
> are probably related to
> https://github.com/ClangBuiltLinux/linux/issues/1192. However, I can
> reproduce these also with ToT Clang, not just with Clang 11.

Thanks, I'm able to reproduce these.  Will take a look.

-- 
Josh