Re: [Qemu-devel] [PATCH 0/7] target-arm: Translate based on TB flags, not CPUState

[David Turner]

For what is worth, this was fixed in the Android emulator with the following patch:

http://android.git.kernel.org/?p=platform/external/qemu.git;a=commit;h=01e9608cb62901d13b330f851a260a2082e81a06

And sorry, we don't have any time to adapt this to the upstream QEMU sources right now.

Hope this helps

On Fri, Jan 7, 2011 at 5:12 PM, Peter Maydell <address@hidden> wrote:

On 7 January 2011 16:01, Aurelien Jarno <address@hidden> wrote:
> In overall I think it's the correct approach to fix the issue, this is
> a really good cleanup. I have tested this patch series, and it clearly
> improve armv7 support. However I am surprised it doesn't fix the issue
> mentioned in https://bugs.launchpad.net/qemu/+bug/581335 , which seems
> to be the same issue. Executing the testcase still returns 1 instead of
> 0 on QEMU.

I think that bug is a different problem, which is that when we take
an unexpected exception we don't correctly reset the condexec bits
when restarting. I think that the right way to fix that is for gen_pc_load()
to restore not just the PC but also the condexec bits (in the same way
that on MIPS we restore env->hflags).

I haven't quite managed to work that up into a patch yet, though
(figuring out exactly where we do and don't need to output code
to update the condexec bits is a bit tricky). I have a long plane flight
on Sunday though so might have another hack at it then :-)

-- PMM