From: Peter Zijlstra <peterz@infradead.org>
To: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>,
LKML <linux-kernel@vger.kernel.org>,
Nick Desaulniers <ndesaulniers@google.com>,
Nathan Chancellor <natechancellor@gmail.com>,
clang-built-linux <clang-built-linux@googlegroups.com>,
x86@kernel.org, Arnd Bergmann <arnd@arndb.de>,
Sedat Dilek <sedat.dilek@gmail.com>,
Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: x86 - clang / objtool status
Date: Wed, 24 Jul 2019 16:10:40 +0200 [thread overview]
Message-ID: <20190724141040.GA31425@hirez.programming.kicks-ass.net> (raw)
In-Reply-To: <20190724133516.GB31381@hirez.programming.kicks-ass.net>
On Wed, Jul 24, 2019 at 03:35:16PM +0200, Peter Zijlstra wrote:
> On Wed, Jul 24, 2019 at 07:55:25AM -0500, Josh Poimboeuf wrote:
> > b) why doesn't objtool detect the case I found?
>
> With GCC you mean? Yes, that is really really weird.
>
> Let me go stare at objdump output for this file (which doesn't build
> with:
>
> make O=defconfig-build/ drivers/gpu/drm/i915/gem/i915_gem_execbuffer.o
> )
0000 0000000000000240 <eb_copy_relocations.isra.34>:
0000 240: 41 57 push %r15
0002 242: 41 56 push %r14
0004 244: 41 55 push %r13
0006 246: 41 54 push %r12
0008 248: 55 push %rbp
0009 249: 53 push %rbx
000a 24a: 48 83 ec 20 sub $0x20,%rsp
000e 24e: 85 f6 test %esi,%esi
0010 250: 74 39 je 28b <eb_copy_relocations.isra.34+0x4b>
0012 252: 89 74 24 14 mov %esi,0x14(%rsp)
0016 256: 45 31 f6 xor %r14d,%r14d
0019 259: 48 c7 04 24 00 00 00 movq $0x0,(%rsp)
0020 260: 00
0021 261: 48 89 7c 24 08 mov %rdi,0x8(%rsp)
0026 266: 48 8b 44 24 08 mov 0x8(%rsp),%rax
002b 26b: 48 8b 34 24 mov (%rsp),%rsi
002f 26f: 48 03 30 add (%rax),%rsi
0032 272: 44 8b 46 04 mov 0x4(%rsi),%r8d
0036 276: 45 85 c0 test %r8d,%r8d
0039 279: 75 23 jne 29e <eb_copy_relocations.isra.34+0x5e>
003b 27b: 41 83 c6 01 add $0x1,%r14d
003f 27f: 48 83 04 24 38 addq $0x38,(%rsp)
0044 284: 44 3b 74 24 14 cmp 0x14(%rsp),%r14d
0049 289: 75 db jne 266 <eb_copy_relocations.isra.34+0x26>
004b 28b: 31 db xor %ebx,%ebx
004d 28d: 48 83 c4 20 add $0x20,%rsp
0051 291: 89 d8 mov %ebx,%eax
0053 293: 5b pop %rbx
0054 294: 5d pop %rbp
0055 295: 41 5c pop %r12
0057 297: 41 5d pop %r13
0059 299: 41 5e pop %r14
005b 29b: 41 5f pop %r15
005d 29d: c3 retq
(<- from +39)
005e 29e: 48 83 c6 08 add $0x8,%rsi
0062 2a2: 44 89 c7 mov %r8d,%edi
0065 2a5: e8 26 ff ff ff callq 1d0 <check_relocations.isra.32>
006a 2aa: 85 c0 test %eax,%eax
006c 2ac: 0f 85 35 01 00 00 jne 3e7 <eb_copy_relocations.isra.34+0x1a7>
0072 2b2: 48 8b 44 24 08 mov 0x8(%rsp),%rax
0077 2b7: 48 8b 0c 24 mov (%rsp),%rcx
007b 2bb: ba ff ff ff ff mov $0xffffffff,%edx
0080 2c0: be c0 0c 00 00 mov $0xcc0,%esi
0085 2c5: 48 8b 00 mov (%rax),%rax
0088 2c8: 4c 8b 6c 08 08 mov 0x8(%rax,%rcx,1),%r13
008d 2cd: 44 89 c0 mov %r8d,%eax
0090 2d0: 49 89 c4 mov %rax,%r12
0093 2d3: 48 89 44 24 18 mov %rax,0x18(%rsp)
0098 2d8: 49 c1 e4 05 shl $0x5,%r12
009c 2dc: 4c 89 e7 mov %r12,%rdi
009f 2df: e8 00 00 00 00 callq 2e4 <eb_copy_relocations.isra.34+0xa4>
00a0 2e0: R_X86_64_PLT32 kvmalloc_node-0x4
00a4 2e4: 49 89 c7 mov %rax,%r15
00a7 2e7: 48 85 c0 test %rax,%rax
00aa 2ea: 0f 84 e8 00 00 00 je 3d8 <eb_copy_relocations.isra.34+0x198>
00b0 2f0: 31 ed xor %ebp,%ebp
00b2 2f2: eb 08 jmp 2fc <eb_copy_relocations.isra.34+0xbc>
(<- from +e0)
00b4 2f4: 48 01 dd add %rbx,%rbp
00b7 2f7: 49 39 ec cmp %rbp,%r12
00ba 2fa: 76 73 jbe 36f <eb_copy_relocations.isra.34+0x12f>
(<- from +b2)
00bc 2fc: 4c 89 e3 mov %r12,%rbx
00bf 2ff: b8 00 00 00 80 mov $0x80000000,%eax
00c4 304: 49 8d 3c 2f lea (%r15,%rbp,1),%rdi
00c8 308: 48 29 eb sub %rbp,%rbx
00cb 30b: 49 8d 74 2d 00 lea 0x0(%r13,%rbp,1),%rsi
00d0 310: 48 39 c3 cmp %rax,%rbx
00d3 313: 48 0f 47 d8 cmova %rax,%rbx
00d7 317: 89 da mov %ebx,%edx
00d9 319: e8 00 00 00 00 callq 31e <eb_copy_relocations.isra.34+0xde>
00da 31a: R_X86_64_PLT32 copy_user_generic_unrolled-0x4
00de 31e: 85 c0 test %eax,%eax
00e0 320: 74 d2 je 2f4 <eb_copy_relocations.isra.34+0xb4>
00e2 322: 4c 89 f8 mov %r15,%rax
00e5 325: 4c 8b 7c 24 08 mov 0x8(%rsp),%r15
00ea 32a: 90 nop
00eb 32b: 90 nop
00ec 32c: 90 nop
^^^ CLAC
And that most certainly should trigger...
Let me gdb that objtool thing.
next prev parent reply other threads:[~2019-07-24 14:11 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-18 20:40 x86 - clang / objtool status Thomas Gleixner
2019-07-18 20:58 ` Nathan Chancellor
2019-07-19 6:39 ` Thomas Gleixner
2019-07-19 7:00 ` Arnd Bergmann
2019-07-19 7:03 ` Nathan Chancellor
2019-07-24 16:57 ` Nick Desaulniers
2019-07-18 22:42 ` Nick Desaulniers
2019-07-19 6:44 ` Thomas Gleixner
2019-07-19 11:37 ` Sedat Dilek
2019-07-19 13:48 ` Sedat Dilek
2019-07-22 15:40 ` Sedat Dilek
2019-07-25 6:17 ` Sedat Dilek
2019-07-24 2:43 ` Josh Poimboeuf
2019-07-24 7:47 ` Peter Zijlstra
2019-07-24 12:37 ` Josh Poimboeuf
2019-07-24 12:55 ` Josh Poimboeuf
2019-07-24 13:35 ` Peter Zijlstra
2019-07-24 14:05 ` Josh Poimboeuf
2019-07-24 14:10 ` Peter Zijlstra [this message]
2019-07-24 16:48 ` [PATCH] objtool: Improve UACCESS coverage Peter Zijlstra
2019-07-24 16:54 ` Nathan Chancellor
2019-07-24 16:55 ` Nick Desaulniers
2019-07-24 18:30 ` Sedat Dilek
2019-07-24 18:32 ` Sedat Dilek
2019-07-24 16:52 ` x86 - clang / objtool status Peter Zijlstra
2019-07-24 17:22 ` Nick Desaulniers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190724141040.GA31425@hirez.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=arnd@arndb.de \
--cc=clang-built-linux@googlegroups.com \
--cc=jpoimboe@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=natechancellor@gmail.com \
--cc=ndesaulniers@google.com \
--cc=sedat.dilek@gmail.com \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox