https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #17 from Mark Wielaard ---
Sorry, we are waiting for me to cleanup the DRM_IOCTL related ioctls.
I did the SYNC_IOC (sync_file) ioctls cleanly in the attachment to comment #15.
Just need to go over the DRM_IOCTL ioctrls to make sure we handl
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #16 from Michael Catanzaro ---
Are we waiting for code review now?
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #15 from Mark Wielaard ---
Created attachment 173179
--> https://bugs.kde.org/attachment.cgi?id=173179&action=edit
Implement stable variant of sync_file ioctls
We implemented an old staging android variant of the sync_file
ioctls. But the
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #14 from Michael Catanzaro ---
Good news: with your latest patch, there are no more fd errors other than fd
leaks. If I determine that any of these look like valgrind issues, then I'll
report a follow-up bug.
--
You are receiving this mail
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #13 from Michael Catanzaro ---
I am only just now realizing that both of the ioctls that I mentioned as
follow-ups, SYNC_IOC_MERGE and DRM_IOCTL_PRIME_HANDLE_TO_FD, are also the ones
that I mentioned in comment #1 and which you had already a
https://bugs.kde.org/show_bug.cgi?id=492422
Mark Wielaard changed:
What|Removed |Added
Attachment #173151|0 |1
is obsolete|
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #11 from Michael Catanzaro ---
(In reply to Michael Catanzaro from comment #10)
> I suspect we're just missing one last ioctl that creates fds. I will
> investigate.
I believe it's coming from DRM_IOCTL_PRIME_HANDLE_TO_FD. (It's coming from
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #10 from Michael Catanzaro ---
(In reply to Mark Wielaard from comment #8)
> Could you test this updated patch to see if it handles file descriptor
> tracking for all three ioctls now?
Yes. There is a massive improvement in the results I se
https://bugs.kde.org/show_bug.cgi?id=492422
Paul Floyd changed:
What|Removed |Added
CC||pjfl...@wanadoo.fr
--- Comment #9 from Paul Floyd
https://bugs.kde.org/show_bug.cgi?id=492422
Mark Wielaard changed:
What|Removed |Added
Attachment #173148|0 |1
is obsolete|
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #7 from Mark Wielaard ---
(In reply to Mark Wielaard from comment #6)
> Urgh, somehow the VKI_SYNC_IOC_MERGE macro in valgrind is broken/wrong.
O no. We got the constant and data structure definitions for sync_file from
some ancient android
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #6 from Mark Wielaard ---
Urgh, somehow the VKI_SYNC_IOC_MERGE macro in valgrind is broken/wrong.
If you add this patch on top things should work.
Scratching my head now what we did wrong and why we need to hardcode the hex
value.
diff --g
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #5 from Michael Catanzaro ---
(In reply to Mark Wielaard from comment #3)
> Could you test with this?
It fixes the issue I reported. Thank you!
But there is another one that looks almost identical:
==130394== File descriptor 28: file desc
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #4 from Michael Catanzaro ---
(In reply to Mark Wielaard from comment #2)
> I'll take a look at those other ioctls. Is there a description of how the
> DRM_IOCTLs handle fds in general? Are there any that close/destroy an fd? Or
> are you al
https://bugs.kde.org/show_bug.cgi?id=492422
Mark Wielaard changed:
What|Removed |Added
Attachment #173142|0 |1
is obsolete|
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #2 from Mark Wielaard ---
Created attachment 173142
--> https://bugs.kde.org/attachment.cgi?id=173142&action=edit
Quick and dirty record_fd_open for DRM_IOCTL_SYNCOBJ_HANDLE_TO_FD
Thanks for the report. I was unaware that there were ioctl
https://bugs.kde.org/show_bug.cgi?id=492422
Mark Wielaard changed:
What|Removed |Added
CC||ahajk...@redhat.com,
|
https://bugs.kde.org/show_bug.cgi?id=492422
--- Comment #1 from Michael Catanzaro ---
Unfortunately Benjamin thinks we'll need a lot more than one new ioctl for
`valgrind --track-fds=yes` to be useful for GTK applications, including at
least SYNC_IOC_MERGE and DRM_IOCTL_PRIME_HANDLE_TO_FD.
I al
https://bugs.kde.org/show_bug.cgi?id=492422
muzikn...@gmail.com changed:
What|Removed |Added
CC||muzikn...@gmail.com
--
You are receiving
19 matches
Mail list logo
