[Nouveau] [Bug 72651] New: [NV11] Hang during suspend/resume on Geforce 2 MX200 & MX400

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Thu Dec 12 14:28:32 PST 2013


https://bugs.freedesktop.org/show_bug.cgi?id=72651

          Priority: medium
            Bug ID: 72651
          Assignee: nouveau at lists.freedesktop.org
           Summary: [NV11] Hang during suspend/resume on Geforce 2 MX200 &
                    MX400
        QA Contact: xorg-team at lists.x.org
          Severity: normal
    Classification: Unclassified
                OS: Linux (All)
          Reporter: welnicki at students.mimuw.edu.pl
          Hardware: x86 (IA32)
            Status: NEW
           Version: unspecified
         Component: Driver/nouveau
           Product: xorg

Created attachment 90680
  --> https://bugs.freedesktop.org/attachment.cgi?id=90680&action=edit
Serial console dump from MX200 card (boot + crash)

Hibernation and suspend/resume always hangs on a GeForce2 MX200/MX400 (NV11)
with nouveau loaded.

Steps to reproduce:
- boot a kernel, load nouveau
- echo "disk" > /sys/power/state
- machine hangs after suspending devices but before the image writing begins
- after several seconds, the monitor complains about "input signal out of
range"

Same thing happens with "echo freeze > /sys/power/state".
If nouveau is unloaded before hibernation, everything works fine.

This happens on two different systems with a GeForce2 MX200/400:
1) Athlon XP 1700+ on an Asrock K7VT2,
   GeForce2 MX400 64MB AGP, with VGA output
2) Pentium II 400 on a 440BX-based Gigabyte GA-BX2000+,
   GeForce2 MX200 32MB AGP, with VGA and TV-out (S-Video + Cinch)

I'm using CentOS 6 userspace, and various kernels:
linux-2.6.32-358.18.1.el6
linux-3.0.x
linux-3.10.x
linux-3.12.1
linux-3.12.4

Using a serial console I was able to narrow down the hang to run_tmds_table(),
called from nv04_dfp_restore().
Dmesg output and VBIOS of both cards is attached.

Note that neither of my cards has DVI/HDMI.
It seems that the flat panel connector is wrongly fabricated in
nouveau_bios.c:fabricate_dcb_encoder_table().

Commenting out the call to fabricate_dcb_output(dcb, DCB_OUTPUT_TMDS, ...)
makes hibernation work fine (patch attached).

The current condition of checking bios->tmds.output{0,1}_script_ptr seems
wrong. But since I don't have any NV11 with DVI output, I'm not sure what to
key creating the TMDS output on.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/nouveau/attachments/20131212/e7da88f6/attachment.html>


More information about the Nouveau mailing list