Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Already on GitHub? Sign in to your account

Xvba #28

Merged
merged 1 commit into from Mar 11, 2012

Conversation

Projects
None yet
3 participants

fritsch commented Mar 10, 2012

XVBA::Present does not secure render enough. If NOT (outPic AND NOTisDuplicate)
the outPic is set, but the render stays potentially NuLL. Which brings CopyYV12 to segfault
as render is not checked for !NULL

Only set outPic if m_presentPicture->render is valid?

@fritsch fritsch xvba: fix newphreak segfault with Shared Surfaces disabled.
XVBA::Present does not secure render enough. If NOT (outPic AND NOTisDuplicate)
the outPic is set, but the render stays potentially NuLL. Which brings CopyYV12 to segfault
as render is not checked for !NULL

Only set outPic if render present?
af70f2e

It should print out here. But will see if the surface is not present, too.

This does indeed prevent it from segfaulting.

fritsch commented Mar 10, 2012

From the log you showed (not online any more), this happened really often. Finding the cause we must.
This is btw. also the cause for this greenish stuff.

Owner

FernetMenta commented Mar 11, 2012

pulled, thanks

@FernetMenta FernetMenta added a commit that referenced this pull request Mar 11, 2012

@FernetMenta FernetMenta Merge pull request #28 from fritsch/xvba
Xvba
46bda28

@FernetMenta FernetMenta merged commit 46bda28 into FernetMenta:xvba Mar 11, 2012

fritsch commented Mar 11, 2012

Any Idea,

Why we get a picture without a render surface? This fixes the segfault
only, the greenish frames are the result of "old picture still
present", cause no render, aka XVBASurface available?

Am 11. März 2012 07:00 schrieb Rainer Hochecker
reply@reply.github.com:

pulled, thanks


Reply to this email directly or view it on GitHub:
#28 (comment)

                   Key-ID:     0x1A995A9B

                   keyserver: pgp.mit.edu

Fingerprint: 4606 DA19 EC2E 9A0B 0157  C81B DA07 CF63 1A99 5A9B

Owner

FernetMenta commented Mar 11, 2012

This will go away with the new design.

fritsch commented Mar 11, 2012

Yeah. I am sure, it will. But do not forget: This current version is
tested by over 2000 different computers. So making a hotfix, which
btw. does affect Shared Surfaces, also - as Present is called and the
picture wrongly bundled. If you have some hints, i will look after it
and fix it in the old version.

Am 11. März 2012 08:17 schrieb Rainer Hochecker
reply@reply.github.com:

This will go away with the new design.


Reply to this email directly or view it on GitHub:
#28 (comment)

                   Key-ID:     0x1A995A9B

                   keyserver: pgp.mit.edu

Fingerprint: 4606 DA19 EC2E 9A0B 0157  C81B DA07 CF63 1A99 5A9B

Owner

FernetMenta commented Mar 11, 2012

This issue is related to the biggest flaw in the current design, the flow of video surfaces and the way they are processed. As long as those kind of fixes (dbba4fc) are in the code, this will never work ok.

The hotfix will be the redesign which will make things much more stable. I don't think this will take longer than 2 weeks.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment