[Mesa-dev] [Mesa-stable] [PATCH] st/nine: Immediately upload user provided textures
Dylan Baker
dylan at pnwbakers.com
Wed Mar 6 20:03:39 UTC 2019
Quoting Axel Davy (2019-03-06 11:27:09)
> On 06/03/2019 20:13, Dylan Baker wrote:
> > Quoting Axel Davy (2019-01-22 12:08:05)
> >> Fixes regression caused by
> >> 42d672fa6a766363e5703f119607f7c7975918aa
> >> st/nine: Bind src not dst in nine_context_box_upload
> >>
> >> Before that patch, for user provided textures,
> >> when the texture was destroyed, the safety
> >> check for pending uploads, which according to
> >> the code "Following condition cannot happen currently",
> >> was flushing the queue and thus triggering the upload.
> >>
> >> After the patch, the texture destruction was delayed after
> >> the upload. However the user frees the texture buffer,
> >> as it thinks the texture released.
> >>
> >> Instead of reverting the faulty patch,
> >> this patch instead flushes the csmt queue right away
> >> after queuing the upload for this type of textures.
> >> This is more future-proof, as we may want to bind the
> >> surface for other reasons in the future.
> >>
> >> Signed-off-by: Axel Davy <davyaxel0 at gmail.com>
> >> Cc: 18.3 <mesa-stable at lists.freedesktop.org>
> >> ---
> >> The regression affects Mesa 18.3.2. At least HL2 lost coast
> >> is affected, has artifacts and crashes at the menu.
> >>
> >> src/gallium/state_trackers/nine/surface9.c | 13 +++++++++++++
> >> 1 file changed, 13 insertions(+)
> >>
> >> diff --git a/src/gallium/state_trackers/nine/surface9.c b/src/gallium/state_trackers/nine/surface9.c
> >> index f94f7c62583..7f4ecf559e8 100644
> >> --- a/src/gallium/state_trackers/nine/surface9.c
> >> +++ b/src/gallium/state_trackers/nine/surface9.c
> >> @@ -668,6 +668,19 @@ NineSurface9_CopyMemToDefault( struct NineSurface9 *This,
> >> From->data, From->stride,
> >> 0, /* depth = 1 */
> >> &src_box);
> >> + if (From->texture == D3DRTYPE_TEXTURE) {
> >> + struct NineTexture9 *tex =
> >> + NineTexture9(From->base.base.container);
> >> + /* D3DPOOL_SYSTEMMEM with buffer content passed
> >> + * from the user: execute the upload right now.
> >> + * It is possible it is enough to delay upload
> >> + * until the surface refcount is 0, but the
> >> + * bind refcount may not be 0, and thus the dtor
> >> + * is not executed (and doesn't trigger the
> >> + * pending_uploads_counter check). */
> >> + if (!tex->managed_buffer)
> >> + nine_csmt_process(This->base.base.device);
> >> + }
> >>
> >> if (This->data_conversion)
> >> (void) util_format_translate(This->format_conversion,
> >> --
> >> 2.20.1
> >>
> > Should I pick this to 19.0 as well after it lands?
> >
> > Dylan
>
> Well, it looks like to be already in the 19.0 tree:
> https://cgit.freedesktop.org/mesa/mesa/commit/?h=19.0&id=d7433c22e6c9624ca5275a3cd35be79caed9fffc
>
>
> Yours,
>
>
> Axel Davy
>
I don't know why, but my email client shows this as being sent yesterday! Since
I wrote my client it's my fault too :(
Sorry for the noise,
Dylan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: signature
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20190306/8150dc7a/attachment.sig>
More information about the mesa-dev
mailing list