LLVM bug 39720

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

LLVM bug 39720

Oleg Smolsky via cfe-dev
Hi,

I am writing here in relation to:

https://bugs.llvm.org/show_bug.cgi?id=39720

because of earlier advice given in regards to it:

On Thu, 15 Nov 2018 15:24:30 +0000 Tim Northover wrote:

> cfe-dev is a much bigger list, and relevant given
> that there's a pretty good chance it's an LLVM bug

On Tue, 20 Nov 2018 11:18:58 -0500 Vedant Kumar wrote:

> This bug would be a bit easier to triage with a
> backtrace attached, ideally with some contextual
> disassembly around the crashing instruction.

I am not a developer and I don't know how to do that,
so if any particular info is needed please someone
ELI5 how to provide it. This is a really critical bug
in the explained scenario as it results in a
non-executable binary.

--
George
_______________________________________________
cfe-dev mailing list
[hidden email]
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev
Reply | Threaded
Open this post in threaded view
|

Re: LLVM bug 39720

Oleg Smolsky via cfe-dev
On 22 Nov 2018, at 12:53, George via cfe-dev <[hidden email]> wrote:

>
> I am writing here in relation to:
>
> https://bugs.llvm.org/show_bug.cgi?id=39720
>
> because of earlier advice given in regards to it:
>
> On Thu, 15 Nov 2018 15:24:30 +0000 Tim Northover wrote:
>
>> cfe-dev is a much bigger list, and relevant given
>> that there's a pretty good chance it's an LLVM bug
>
> On Tue, 20 Nov 2018 11:18:58 -0500 Vedant Kumar wrote:
>
>> This bug would be a bit easier to triage with a
>> backtrace attached, ideally with some contextual
>> disassembly around the crashing instruction.
>
> I am not a developer and I don't know how to do that,
> so if any particular info is needed please someone
> ELI5 how to provide it. This is a really critical bug
> in the explained scenario as it results in a
> non-executable binary.
Maybe post that binary and a corresponding core dump somewhere, so
somebody else can load those in a debugger?  Or let your favorite crash
reporting tool (catchsegv or any of the Linux built-in crash reporters)
generate a full report, and post it in the PR?

In any case, you could try the following.  Make sure the gdb package is
installed, then run:

  gdb /opt/ungoogled-chromium/chrome

When gdb is finished loading the executable, type "run", and wait until
the program crashes.  Then execute the following gdb commands:

  bt full
  info registers
  disassemble

and post the full output of those commands in the PR.

-Dimitry


_______________________________________________
cfe-dev mailing list
[hidden email]
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev

signature.asc (230 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: LLVM bug 39720

Oleg Smolsky via cfe-dev
On Thu, 22 Nov 2018 13:09:03 +0100 Dimitry Andric
wrote:

> Maybe post that binary and a corresponding core dump
> somewhere, so somebody else can load those in a
> debugger?

# du -hs /opt/ungoogled-chromium/
288M    /opt/ungoogled-chromium/
# du -hs /opt/ungoogled-chromium/chrome
199M    /opt/ungoogled-chromium/chrome

I don't know where to post something of such size. But
using the shared build scripts it can be reproduced
quite easily.

> Or let your favorite crash reporting tool (catchsegv
> or any of the Linux built-in crash reporters)
> generate a full report, and post it in the PR?

I have no idea what you are talking about (not a
developer as mentioned). I checked but couldn't find
such tool:

# zypper se catchsegv
Loading repository data...
Reading installed packages...
No matching items found.

> In any case, you could try the following.  Make sure
> the gdb package is installed, then run:
>
>   gdb /opt/ungoogled-chromium/chrome
>
> When gdb is finished loading the executable, type
> "run", and wait until the program crashes.  Then
> execute the following gdb commands:
>
>   bt full
>   info registers
>   disassemble
>
> and post the full output of those commands in the PR.

Done:

https://bugs.llvm.org/show_bug.cgi?id=39720#c2

--
George
_______________________________________________
cfe-dev mailing list
[hidden email]
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev