git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Johannes Sixt <j.sixt@viscovery.net>
To: Michael Wookey <michaelwookey@gmail.com>,
	Junio C Hamano <gitster@pobox.com>
Cc: eduard stefan <eduard.stefan@gmail.com>,
	Tay Ray Chuan <rctay89@gmail.com>,
	Git Mailing List <git@vger.kernel.org>,
	msysgit@googlegroups.com
Subject: Re: Git 1.6.5-rc git clone unhandled exception using http protocol
Date: Tue, 13 Oct 2009 12:53:28 +0200	[thread overview]
Message-ID: <4AD45C28.4080501@viscovery.net> (raw)
In-Reply-To: <d2e97e800910130310wa9731a6j9b9bdd25047ade85@mail.gmail.com>


Michael Wookey schrieb:
> Using the above repository, I see the same crash with msysGit at git
> revision 1.6.5. Using windbg as the post-mortem debugger, the
> following information is captured:
> 
>   (a14.e8c): Access violation - code c0000005 (!!! second chance !!!)
>   eax=00000000 ebx=00000000 ecx=ffffffff edx=0046cc00 esi=0046f98f edi=00000000
>   eip=00420354 esp=0022fd80 ebp=0022fda8 iopl=0         nv up ei pl zr na pe nc
>   cs=001b  ss=0023  ds=0023  es=0023  fs=003b  gs=0000             efl=00010246
> 
> ...and the faulting instruction is:
> 
>   git_remote_curl+0x20354:
>   00420354 f2ae            repne scas byte ptr es:[edi]
> 
> so, a NULL dereference. The initial disassembly of the function is this:
> 
>   0:000> u 0042033C
>   git_remote_curl+0x2033c:
>   0042033c 55              push    ebp
>   0042033d 89e5            mov     ebp,esp
>   0042033f 57              push    edi
>   00420340 56              push    esi
>   00420341 53              push    ebx
>   00420342 83ec1c          sub     esp,1Ch
>   00420345 8b5d08          mov     ebx,dword ptr [ebp+8]
>   00420348 8b750c          mov     esi,dword ptr [ebp+0Ch]
>   0042034b 31c0            xor     eax,eax
>   0042034d b9ffffffff      mov     ecx,0FFFFFFFFh
>   00420352 89df            mov     edi,ebx
>   00420354 f2ae            repne scas byte ptr es:[edi]
>   00420356 f7d1            not     ecx
>   00420358 8d51ff          lea     edx,[ecx-1]
>   0042035b b9ffffffff      mov     ecx,0FFFFFFFFh
>   00420360 89f7            mov     edi,esi
>   00420362 f2ae            repne scas byte ptr es:[edi]
>   00420364 f7d1            not     ecx
>   00420366 49              dec     ecx
>   00420367 7466            je      git_remote_curl+0x203cf (004203cf)
>   00420369 85d2            test    edx,edx
>   0042036b 0f84b1000000    je      git_remote_curl+0x20422 (00420422)
>   00420371 89f7            mov     edi,esi
>   00420373 89de            mov     esi,ebx
>   ...
> 
> So its the first parameter that is NULL. The second parameter is:
> 
>   0:000> da poi(ebp+c)
>   0046f98f  "libexec/git-core"
> 
> I don't know how to build msysGit so that symbols are generated so
> I've attempted to reconstruct the source code; which ends up looking
> something like the following:
> 
>   int some_unknown_func(char *arg1, char *arg2)
>   {
>       len1 = strlen(arg1) - 1;  // <- crash here
>       len2 = strlen(arg2);
>       len3 = len2 - 1;
> 
>       if (len2 != 1) {
>           if (!len1)
>               return 0;
>           for (;;) {
>               x = arg1[len1 - 1];
> 
>               if (x != '/' && x != '\\') {
>                   --len1;
>                   --len3;
>                   if (arg1[len1] != arg2[len3])
>                       return 0;
>               } else {
>                   ...
>               }
>           }
>       }
>   }
> 
> Perhaps those more familiar with git's sources might recognise code
> that looks similar to the above sequence.

Wow, this is great work, thank you very much! The function is
strip_path_suffix(). And here is a patch that fixes the crash.

--- >8 ---
From: Johannes Sixt <j6t@kdbg.org>
Subject: [PATCH] remote-curl: add missing initialization of argv0_path

All programs, in particular also the stand-alone programs (non-builtins)
must call git_extract_argv0_path(argv[0]) in order to help builds that
derive the installation prefix at runtime, such as the MinGW build.
Without this call, the program segfaults (or raises an assertion
failure).

Signed-off-by: Johannes Sixt <j6t@kdbg.org>
---
 remote-curl.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/remote-curl.c b/remote-curl.c
index ad6a163..d8d276a 100644
--- a/remote-curl.c
+++ b/remote-curl.c
@@ -82,6 +82,7 @@ int main(int argc, const char **argv)
 	const char *url;
 	struct walker *walker = NULL;

+	git_extract_argv0_path(argv[0]);
 	setup_git_directory();
 	if (argc < 2) {
 		fprintf(stderr, "Remote needed\n");
-- 
1.6.5.1024.g31034.dirty

  reply	other threads:[~2009-10-13 10:53 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-10-09 17:27 Git 1.6.5-rc git clone unhandled exception using http protocol eduard stefan
2009-10-10 10:55 ` Tay Ray Chuan
     [not found]   ` <4AD09F5E.9090304@gmail.com>
     [not found]     ` <be6fef0d0910100811l325d3df1jdf8d3d9dd51e3385@mail.gmail.com>
2009-10-10 15:55       ` eduard stefan
2009-10-10 16:07         ` Tay Ray Chuan
2009-10-10 16:52           ` eduard stefan
2009-10-10 22:54             ` Tay Ray Chuan
2009-10-11 21:12               ` eduard stefan
2009-10-13  3:36                 ` Git 1.6.5 " eduard stefan
2009-10-13 10:10             ` Git 1.6.5-rc " Michael Wookey
2009-10-13 10:53               ` Johannes Sixt [this message]
2009-10-13 11:43                 ` Michael Wookey
2009-10-13 21:06                   ` Junio C Hamano

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=4AD45C28.4080501@viscovery.net \
    --to=j.sixt@viscovery.net \
    --cc=eduard.stefan@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=michaelwookey@gmail.com \
    --cc=msysgit@googlegroups.com \
    --cc=rctay89@gmail.com \
    /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;
as well as URLs for NNTP newsgroup(s).