Add MMX detection when getting registers in Linux

For CPUs that don't support the MMX instruction set, such pre-Pentium III or
industrial x86 embedded PCs, the minidump fails when it tries to retrieve MMX
specific registers.

This patch adds MMX detection for that call.

Tested on Ubuntu 12.04 with i686, and on a custom Linux distro on a Vortex86DX
microcontroller.

[Ted Mielczarek]

This LGTM, I'll land it as soon as I find out if you have a CLA on file.

[Lei Zhang (chromium)]

lgtm

You need to sign the Google CLA before we can accept this code:
https://developers.google.com/open-source/cla/individual

https://breakpad.appspot.com/455002/diff/1/src/client/linux/minidump_writer/l...
File src/client/linux/minidump_writer/linux_ptrace_dumper.cc (right):

https://breakpad.appspot.com/455002/diff/1/src/client/linux/minidump_writer/l...
src/client/linux/minidump_writer/linux_ptrace_dumper.cc:49: #include <cpuid.h>
nit: alphabetical order please

https://breakpad.appspot.com/455002/diff/1/src/client/linux/minidump_writer/l...
src/client/linux/minidump_writer/linux_ptrace_dumper.cc:198: if (edx & bit_MMX)
{
nit: extra trailing tab.

https://breakpad.appspot.com/455002/diff/1/src/client/linux/minidump_writer/l...
src/client/linux/minidump_writer/linux_ptrace_dumper.cc:203: memset(
&info->fpxregs, 0, sizeof(info->fpxregs) );
nit: no spaces on either side of parenthesis.

[aras.vaichas]

Google CLA has been signed off.

Aras


On 26 November 2013 23:49, <thestig@chromium.org> wrote:

> lgtm
>
> You need to sign the Google CLA before we can accept this code:
> https://developers.google.com/open-source/cla/individual
>
>
> https://breakpad.appspot.com/455002/diff/1/src/client/
> linux/minidump_writer/linux_ptrace_dumper.cc
> File src/client/linux/minidump_writer/linux_ptrace_dumper.cc (right):
>
> https://breakpad.appspot.com/455002/diff/1/src/client/
> linux/minidump_writer/linux_ptrace_dumper.cc#newcode49
> src/client/linux/minidump_writer/linux_ptrace_dumper.cc:49: #include
> <cpuid.h>
> nit: alphabetical order please
>
> https://breakpad.appspot.com/455002/diff/1/src/client/
> linux/minidump_writer/linux_ptrace_dumper.cc#newcode198
> src/client/linux/minidump_writer/linux_ptrace_dumper.cc:198: if (edx &
> bit_MMX) {
> nit: extra trailing tab.
>
> https://breakpad.appspot.com/455002/diff/1/src/client/
> linux/minidump_writer/linux_ptrace_dumper.cc#newcode203
> src/client/linux/minidump_writer/linux_ptrace_dumper.cc:203: memset(
> &info->fpxregs, 0, sizeof(info->fpxregs) );
> nit: no spaces on either side of parenthesis.
>
> https://breakpad.appspot.com/455002/
>

[Ted Mielczarek]

On 2013/12/09 09:02:30, aras.vaichas wrote:
> Google CLA has been signed off.

thestig: can you verify this and land? (Or poke me and I'll land it.)

[aras.vaichas]

I don't know what you mean by "land".

Sorry, but I have no way to verify it. This was a patch I made as a
contractor at a previous company for a specific piece of hardware. I don't
have the target hardware, the development environment or even the original
patch anymore.

The patch was taken from working code, so I know it worked on my platform.
I tested it on my development PC (which had MMX registers) and the target
CPU (no MMX registers).

Aras


On 9 December 2013 20:05, <ted.mielczarek@gmail.com> wrote:

> On 2013/12/09 09:02:30, aras.vaichas wrote:
>
>> Google CLA has been signed off.
>>
>
> thestig: can you verify this and land? (Or poke me and I'll land it.)
>
> https://breakpad.appspot.com/455002/
>

[Lei Zhang (chromium)]

On 2013/12/09 20:05:22, Ted Mielczarek wrote:
> On 2013/12/09 09:02:30, aras.vaichas wrote:
> > Google CLA has been signed off.
> 
> thestig: can you verify this and land? (Or poke me and I'll land it.)

CLA signed. Will land in a minute.

[Lei Zhang (chromium)]

On 2013/12/09 20:13:51, aras.vaichas wrote:
> I don't know what you mean by "land".

Land = commit.

> Sorry, but I have no way to verify it. This was a patch I made as a
> contractor at a previous company for a specific piece of hardware. I don't
> have the target hardware, the development environment or even the original
> patch anymore.
> 
> The patch was taken from working code, so I know it worked on my platform.
> I tested it on my development PC (which had MMX registers) and the target
> CPU (no MMX registers).

I believe someone recently mentioned on the bug that they had similar hardware
and this patch fixes the problem for them.

[Lei Zhang (chromium)]

Feel free to Edit -> Close this issue since I have committed it for you.