[Mesa-stable] [Mesa-dev] [PATCH v2 01/18] anv: Add the pci_id into the shader cache UUID

[Jason Ekstrand]

On Wed, Apr 26, 2017 at 1:20 AM, Juan A. Suarez Romero <jasuarez at igalia.com>
wrote:

> On Tue, 2017-04-25 at 23:42 -0700, Jason Ekstrand wrote:
> > On Mon, Apr 24, 2017 at 5:17 PM, Andres Gomez <agomez at igalia.com> wrote:
> > > Hi Jason,
> > >
> > > this patch landed in master with Cc to stable.
> > >
> > > However, the changes need commit 656e30b6860 and quite some other
> > > dependencies (util/build_id.h)
> > >
> > > Additionally, Tapani is mentioning problems in Android and the
> > > probability of the issue that is trying to fix seems unlikely (?).
> > >
> > > Should we drop this patch from the stable branch?
> >
> > We probably can but I'd mildly prefer not.  I'm not sure what issue
> Tapani is hitting but I suspect it's just android not setting the right
> --build-id flag.
> >
>
>
> Just to make it clear, we are talking about dropping it for 17.0.4, but
> keep for 17.1.
>
> Unfortunately, for 17.0.4 it would require several other commits that
> are not in the 17.0 branch.
>
> Another alternative would be submitting a specific patch for 17.0 that
> does the same, but do not rely on those commits.
>

I think dropping it is fine for now.  We can come up with something if we
actually ever see the bug in the wild

--Jason


> > > Br.
> > >
> > > On Thu, 2017-04-20 at 10:11 +0300, Tapani Pälli wrote:
> > > >
> > > > On 03/14/2017 04:26 AM, Jason Ekstrand wrote:
> > > > > This prevents a user from using a cache created on one hardware
> > > > > generation on a different one.  Of course, with Intel hardware,
> this
> > > > > requires moving their drive from one machine to another but it's
> still
> > > > > possible and we should prevent it.
> > > > >
> > > > > Reviewed-by: Chad Versace <chadversary at chromium.org>
> > > > > ---
> > > > >   src/intel/vulkan/anv_device.c | 22 +++++++++++++++++-----
> > > > >   1 file changed, 17 insertions(+), 5 deletions(-)
> > > > >
> > > > > diff --git a/src/intel/vulkan/anv_device.c
> b/src/intel/vulkan/anv_device.c
> > > > > index be2b2f3..5db0a6c 100644
> > > > > --- a/src/intel/vulkan/anv_device.c
> > > > > +++ b/src/intel/vulkan/anv_device.c
> > > > > @@ -32,6 +32,7 @@
> > > > >   #include "util/strtod.h"
> > > > >   #include "util/debug.h"
> > > > >   #include "util/build_id.h"
> > > > > +#include "util/mesa-sha1.h"
> > > > >   #include "util/vk_util.h"
> > > > >
> > > > >   #include "genxml/gen7_pack.h"
> > > > > @@ -53,17 +54,28 @@ compiler_perf_log(void *data, const char *fmt,
> ...)
> > > > >   }
> > > > >
> > > > >   static bool
> > > > > -anv_device_get_cache_uuid(void *uuid)
> > > > > +anv_device_get_cache_uuid(void *uuid, uint16_t pci_id)
> > > > >   {
> > > > >      const struct build_id_note *note =
> build_id_find_nhdr("libvulkan_intel.so");
> > > > >      if (!note)
> > > > >         return false;
> > > > >
> > > > > -   unsigned len = build_id_length(note);
> > > > > -   if (len < VK_UUID_SIZE)
> > > > > +   unsigned build_id_len = build_id_length(note);
> > > > > +   if (build_id_len < 20) /* It should be a SHA-1 */
> > > >
> > > > This line here causes issues for Android, for us the build_id_len
> seems
> > > > to be 16 (and worked before nicely as check was against VK_UUID_SIZE
> > > > which is 16). I just bisected this and haven't yet figured out why
> but
> > > > will check.
> > > >
> > > > // Tapani
> > > > _______________________________________________
> > > > mesa-dev mailing list
> > > > mesa-dev at lists.freedesktop.org
> > > > https://lists.freedesktop.org/mailman/listinfo/mesa-dev
> > > --
> > > Br,
> > >
> > > Andres
> > >
> >
> > _______________________________________________
> > mesa-stable mailing list
> > mesa-stable at lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/mesa-stable
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-stable/attachments/20170426/ffe3faae/attachment.html>