Re: [Qemu-devel] [PULL 10/25] aarch64-linux-user: Add support for SVE signal frame records

[Peter Maydell]

On 9 March 2018 at 17:26, Peter Maydell <address@hidden> wrote:
> From: Richard Henderson <address@hidden>
>
> Depending on the currently selected size of the SVE vector registers,
> we can either store the data within the "standard" allocation, or we
> may beedn to allocate additional space with an EXTRA record.
>
> Signed-off-by: Richard Henderson <address@hidden>
> Message-id: address@hidden
> Reviewed-by: Peter Maydell <address@hidden>
> Signed-off-by: Peter Maydell <address@hidden>
> ---
>  linux-user/signal.c | 210 
> +++++++++++++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 192 insertions(+), 18 deletions(-)

I did a 'git bisect' looking for when we introduced the segv
bug described in https://bugs.launchpad.net/qemu/+bug/1761535,
and git bisect thinks this commit is it.

At least for me, with a xenial aarch64 chroot, with this
patch if I chroot in and run 'ls' then we get a spurious
segfault (I think in the guest bash):

e104462:xenial:chroot$ sudo chroot xenial-aarch64
address@hidden:/# ls
Deadlock.class    bin   home                  hotspot_pid17113.log
mnt   root   srv  var
Deadlock.java     boot  hotspot_pid14759.log  hotspot_pid19407.log
opt   run    sys
HelloWorld.class  dev   hotspot_pid16542.log  lib
proc  sbin   tmp
HelloWorld.java   etc   hotspot_pid16895.log  media
risu  sdbg9  usr
qemu: uncaught target signal 11 (Segmentation fault) - core dumped
Segmentation fault (core dumped)

Reverting f914baef8e39f7 on master fixes the segfault.

Richard (Henderson): I'll have a closer look at this on Monday if you don't
get to it first...

thanks
-- PMM