On Tue, Nov 21, 2017 at 08:01:16PM +0100, Marek Olšák wrote:
> Where is patch 1?
The patches didn't end up in the same thread. Really sorry about that, still
getting used to sending patches via mailing list.
The subject is
> [Mesa-dev] [PATCH 1/4] st/mesa: Propagate framebuffer validation failure
Callers of dri_drawable->allocate_textures are unaware of failures
occurring in the provided allocation function, happily chugging along
with potentially invalid state. Update the interface and
{dri2,drisw}_allocate_textures to return a boolean indicating success.
---
src/gallium/state_trackers/dr
DRI routines in X drivers can return null buffers if they're unable to
allocate them as requested (for example, [1]) but the DRI 2 protocol
doesn't expose a mechanism to communicate this explicitly. Instead, we
assume that zeroed out buffer info structs are invalid and go from
there.
This commit a
st_framebuffer_validate discards the result of the validate interface
method, preventing functions up the call stack from being notified of
any failure. Update st_framebuffer_validate to resolve this, and amend
st_api_make_current to propagate this failure up the stack.
---
src/mesa/state_tracker/
dri_make_current discards the return value of st_api_make_current,
preventing callers up the stack being made aware of any failures.
---
src/gallium/state_trackers/dri/dri_context.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/gallium/state_trackers/dri/dri_context.c
generate_array_index fails to check whether the target of a subroutine
call exists in the AST, potentially passing around null ir_rvalue
pointers eventuating in abort/segfault.
Fixes: fd01840c0bd3b22d058a65a17ad30e3b45813b60
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100438
---
src/co
