https://bugzilla.kernel.org/show_bug.cgi?id=77471
Bug ID: 77471 Summary: VESA VGA to radeondri handoff results in boot hang Product: Drivers Version: 2.5 Kernel Version: 3.15-rc8 Hardware: x86-64 OS: Linux Tree: Mainline Status: NEW Severity: normal Priority: P1 Component: Video(DRI - non Intel) Assignee: drivers_video-dri@kernel-bugs.osdl.org Reporter: joshua.r.marshall.1991@gmail.com Regression: No
Created attachment 138461 --> https://bugzilla.kernel.org/attachment.cgi?id=138461&action=edit The kernel config I'm using
Suspecting a kernel memory leak involving radeon stuffs, I pulled down the kernel source, used "make localyesconfig" to automatically configure the kernel from my distro's configuration, did the typical procedure to make the kernel boot as per archlinux instructions and past experiance. During boot of the new kernel, a line saying the handoff from VESA VGA was handing off to radeondri for the framebuffer switchoff prints, CPU goes to 100% (from what I can tell -- fan speeds up and CPU kicks off heat) and does not progress. Even given a half hour, this doesn't change.
I'm not sure how to proceed other than filing this report.
https://bugzilla.kernel.org/show_bug.cgi?id=77471
--- Comment #1 from joshua.r.marshall.1991@gmail.com --- My distro kernel version is at 3.14.5
https://bugzilla.kernel.org/show_bug.cgi?id=77471
Alex Deucher alexdeucher@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |alexdeucher@gmail.com
--- Comment #2 from Alex Deucher alexdeucher@gmail.com --- Can you attach the dmesg output from a sucessful load the driver? What GPU is this?
https://bugzilla.kernel.org/show_bug.cgi?id=77471
--- Comment #3 from joshua.r.marshall.1991@gmail.com --- This is the R7 GPU on the A10-7850K
https://bugzilla.kernel.org/show_bug.cgi?id=77471
--- Comment #4 from joshua.r.marshall.1991@gmail.com --- Created attachment 138971 --> https://bugzilla.kernel.org/attachment.cgi?id=138971&action=edit New kernel config that does boot
I used localmodconfig for then, then added debug info -- my guess is something related to having the module built in vs. loaded. I ran into similar issues with my intel graphics a while back.
https://bugzilla.kernel.org/show_bug.cgi?id=77471
--- Comment #5 from joshua.r.marshall.1991@gmail.com --- Created attachment 138981 --> https://bugzilla.kernel.org/attachment.cgi?id=138981&action=edit dmesg output
I found where my kernel is leaking memory (USB audio related) but this hand off should still not have been an issue.
https://bugzilla.kernel.org/show_bug.cgi?id=77471
--- Comment #6 from Alex Deucher alexdeucher@gmail.com --- Did you build the necessary firmware into your kernel image when you built the driver into the kernel? I don't think this has anything to do with the handoff.
https://bugzilla.kernel.org/show_bug.cgi?id=77471
--- Comment #7 from joshua.r.marshall.1991@gmail.com --- I'm not aware of anything involving firmware for building the modules separately or part of the main kernel. I follow the same build procedure for both.
https://bugzilla.kernel.org/show_bug.cgi?id=77471
--- Comment #8 from Alex Deucher alexdeucher@gmail.com --- You need to build the firmware into the kernel image if you build the driver into the kernel otherwise, the driver will not be able to fetch the firmware. What you may be seeing is the kernel firmware loading waiting for firmware to become available, but it never does since it's not actually in the kernel binary.
https://bugzilla.kernel.org/show_bug.cgi?id=77471
joshua.r.marshall.1991@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |INVALID Status|NEW |RESOLVED
--- Comment #9 from joshua.r.marshall.1991@gmail.com --- That looks to be it.
dri-devel@lists.freedesktop.org
-
bugzilla-daemon＠bugzilla.kernel.org