[Qemu-devel] Re: [PATCH] Guest debugging support for KVM

[Jan Kiszka <jan.kiszka@web.de>]

[-- Attachment #1: Type: text/plain, Size: 2322 bytes --]

Anthony Liguori wrote:
> Jan Kiszka wrote:
>> [ Also available via git://git.kiszka.org/qemu.git queue/gdb ]
>>
>> This is a backport of the guest debugging support for the KVM
>> accelerator that is now part of the KVM tree. It implements the reworked
>> KVM kernel API for guest debugging (KVM_CAP_SET_GUEST_DEBUG) which is
>> not yet part of any mainline kernel but will probably be 2.6.30 stuff.
>> So far supported is x86, but PPC is expected to catch up soon.
>>
>> Core features are:
>>  - unlimited soft-breakpoints via code patching
>>  - hardware-assisted x86 breakpoints and watchpoints
>>
>> Open issues: Writing soft breakpoints to BIOS code. cpu_memory_rw_debug
>> does not work here. Suggestion to fix this cleanly are welcome.
>>
>> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
>> ---
>>
>>  configure         |    4 +
>>  exec.c            |   10 ++-
>>  gdbstub.c         |   29 +++++++--
>>  gdbstub.h         |    7 ++
>>  kvm-all.c         |  172
>> +++++++++++++++++++++++++++++++++++++++++++++++++++++
>>  kvm.h             |   41 +++++++++++++
>>  target-i386/kvm.c |  172
>> +++++++++++++++++++++++++++++++++++++++++++++++++++++
>>  7 files changed, 423 insertions(+), 12 deletions(-)
>>
>> diff --git a/configure b/configure
>> index a8ea55a..afbe1e0 100755
>> --- a/configure
>> +++ b/configure
>> @@ -479,11 +479,11 @@ case "$cpu" in
>>             ARCH_CFLAGS="-march=z900"
>>             ;;
>>      i386)
>> -           ARCH_CFLAGS="-m32"
>> +           ARCH_CFLAGS="-m32 -DCONFIG_X86"
>>             ARCH_LDFLAGS="-m32"
>>             ;;
>>      x86_64)
>> -           ARCH_CFLAGS="-m64"
>> +           ARCH_CFLAGS="-m64 -DCONFIG_X86"
>>             ARCH_LDFLAGS="-m64"
>>             ;;
>>  esac
>>   
> 
> This seems unnecessary, no?

Nope. I suggested and Avi merged a patch to define
KVM_CAP_SET_GUEST_DEBUG only for those archs (ie. currently x86) that
really support it. That avoids having to define empty stubs on the other
KVM archs.

So we need to define the proper arch switch to make the CAP appear in
the kernel header. Also note that there are much more conditional CAP
defines queued up or already merged into mainline.

> 
> The rest looks good.
> 
> Regards,
> 
> Anthony Liguori
> 
> 
> 

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 257 bytes --]