Bug 76738 - [snb regression] pll readout blows up on p1 == 0
Summary: [snb regression] pll readout blows up on p1 == 0
Status: CLOSED WONTFIX
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: XOrg git
Hardware: Other All
: highest normal
Assignee: Daniel Vetter
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-03-28 14:17 UTC by Peter Senna Tschudin
Modified: 2017-07-24 22:55 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
dmesg with drm.debug=0xe (47.31 KB, text/plain)
2014-03-28 14:17 UTC, Peter Senna Tschudin
no flags Details
Dmesg with drm.debug=0xe after applying patch from Daniel Vetter (71.85 KB, application/gzip)
2014-03-28 14:19 UTC, Peter Senna Tschudin
no flags Details
output of intel_reg_dumper with nomodeset (14.79 KB, text/plain)
2014-03-29 13:01 UTC, Peter Senna Tschudin
no flags Details

Description Peter Senna Tschudin 2014-03-28 14:17:33 UTC
Created attachment 96533 [details]
dmesg with drm.debug=0xe

When Fedora updated the Kernel package from 3.12 to 3.13 my notebook stopped booting (Kernel freezes) when a 2560 x 1440 high res monitor is attached. I have tried using 3.13.6 from kernel.org and the problem persists. The problem can be partially solved by passing nomodeset to Kernel which will make the Kernel to boot, but the screens never comes to life. When using 3.14-rc7 it boots again, the screen works, but with some warning messages.

The problem persists on linux-3.14-rc8
Comment 1 Peter Senna Tschudin 2014-03-28 14:19:08 UTC
Created attachment 96534 [details]
Dmesg with drm.debug=0xe after applying patch from Daniel Vetter

Dmesg with drm.debug=0xe and with the patch:

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 7be5984431bb..3c0f8f095a6b 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -450,6 +450,8 @@ static void i9xx_clock(int refclk, intel_clock_t *clock)
 {
        clock->m = i9xx_dpll_compute_m(clock);
        clock->p = clock->p1 * clock->p2;
+       printk("p1 = %u, p2 = %u, n = %u\n", clock->p1, clock->p2,
+              clock->n);
        if (WARN_ON(clock->n + 2 == 0 || clock->p == 0))
                return;
        clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n + 2);
Comment 2 Peter Senna Tschudin 2014-03-28 14:21:06 UTC
The lkml thread: http://www.kernelhub.org/?p=2&msg=435493
Comment 3 Daniel Vetter 2014-03-28 19:34:25 UTC
Can you please boot with nomodeset (and no X if possible) and grab the output of the intel_reg_dumper tool form intel-gpu-tools? Apparently your BIOS sets up the screen in a way we deem impossible.
Comment 4 Peter Senna Tschudin 2014-03-29 13:01:41 UTC
Created attachment 96586 [details]
output of intel_reg_dumper with nomodeset

Setting nomodest created the same behavior of 3.13 in which the Kernel boots but the screens never come to life.
Comment 5 Peter Senna Tschudin 2014-03-29 13:09:39 UTC
Details of my setup:
 - I use a dock station from Toshiba which has one HDMI and one VGA output.
 - The monitor connected to HDMI output has a resolution that is not supported by the GPU. To use the monitor with full resolution I reduce the refresh rate to 38Hz with:

$ xrandr --newmode "2560x1440_38.00" 189.75 2560 2712 2976 3392 1440 1443 1448 1474 -hsync +vsync
$ xrandr --addmode HDMI3 "2560x1440_38.00"
$ xrandr --output HDMI3 --mode "2560x1440_38.00"

Why I think it is working with 3.14:
 - During the boot the VGA monitor is being detected as the primary while on 3.13 the one at HDMI port was being detected as primary. I never saw a text console on the high res monitor, it never worked.
Comment 6 Daniel Vetter 2014-03-29 14:29:19 UTC
Just to check: When you boot with nomodeset, does any screen work?

According to the reg dump HDMI is enabled, VGA is disabled, but the pll is complete bs.

Also, does the BIOS manage to light anything up properly? At what point do you start to see the black screen (linux or boot-loader/grub or ...)?

Can you try to remove any special options from grub like gfxmode or similar and let grub run in legacy vga mode?
Comment 7 Peter Senna Tschudin 2014-03-29 14:42:44 UTC
>Just to check: When you boot with nomodeset, does any screen work?
No, none of the external screens work. I had to login and run intel_reg_dumper without seeing anything. Note that while the notebook is on the dock, the lid is closed. The dock has a power button...

>Also, does the BIOS manage to light anything up properly? At what point do you
>start to see the black screen (linux or boot-loader/grub or ...)?
No nothing before the Kernel powers up any of the screens. Nor the bios, nor grub. Without nomodeset on 3.14, I can see text messages of services starting on VGA.

>Can you try to remove any special options from grub like gfxmode or similar and
>let grub run in legacy vga mode?
Current configuration:
linux   /boot/vmlinuz-3.14.0-rc8+ root=UUID=135a331c-88e4-4131-8630-93fc16045d95 ro rd.md=0 rd.lvm=0 rd.dm=0 vconsole.keymap=us  rd.luks=0 vconsole.font=latarcyrheb-sun16 quiet

What should I try?
Comment 8 Imre Deak 2014-09-17 13:45:31 UTC
(In reply to comment #7) 
> >Can you try to remove any special options from grub like gfxmode or similar and
> >let grub run in legacy vga mode?
> Current configuration:
> linux   /boot/vmlinuz-3.14.0-rc8+
> root=UUID=135a331c-88e4-4131-8630-93fc16045d95 ro rd.md=0 rd.lvm=0 rd.dm=0
> vconsole.keymap=us  rd.luks=0 vconsole.font=latarcyrheb-sun16 quiet
> 
> What should I try?

Just to clarify, does the HDMI output work at all with 3.14-rc7? There is at least an HDMI modeset in the log that seems to succeed.

As Daniel pointed out, the BIOS or boot loader seems to configure the PLL with an invalid config, but then (at least 3.14-rc7 and for VGA) the driver modeset succeeds. Did you try already to upgrade your BIOS?

To make sure it's not a bootloader issue by forcing GRUB to use text mode: GRUB has additional global options besides those you listed above, you should check places like /etc/default/grub or /boot/grub/grub.cfg. You'd need to set something like GRUB_GFXPAYLOAD_LINUX=text for GRUB2.0, please check the GRUB docs to be sure. Afterwards we'd need another reg dump after you booted with nomodeset.

As I understood 3.12 worked with HDMI attached but 3.13 hanged. Could you do a kernel bisect between those two versions?

Thanks.
Comment 9 Peter Senna Tschudin 2014-09-17 13:59:47 UTC
On Wed, Sep 17, 2014 at 3:45 PM,  <bugzilla-daemon@freedesktop.org> wrote:
> Comment # 8 on bug 76738 from Imre Deak
>
> (In reply to comment #7)
>> >Can you try to remove any special options from grub like gfxmode or
>> > similar and
>> >let grub run in legacy vga mode?
>> Current configuration:
>> linux   /boot/vmlinuz-3.14.0-rc8+
>> root=UUID=135a331c-88e4-4131-8630-93fc16045d95 ro rd.md=0 rd.lvm=0 rd.dm=0
>> vconsole.keymap=us  rd.luks=0 vconsole.font=latarcyrheb-sun16 quiet
>>
>> What should I try?
>
> Just to clarify, does the HDMI output work at all with 3.14-rc7? There is at
> least an HDMI modeset in the log that seems to succeed.

I'm sorry but I have a different hardware setup now, with a newer
version of the GPU in which the problem doesn't happen any more.

>
> As Daniel pointed out, the BIOS or boot loader seems to configure the PLL
> with
> an invalid config, but then (at least 3.14-rc7 and for VGA) the driver
> modeset
> succeeds. Did you try already to upgrade your BIOS?
Toshiba has discontinued the model, but I was using the latest bios available.

>
> To make sure it's not a bootloader issue by forcing GRUB to use text mode:
> GRUB
> has additional global options besides those you listed above, you should
> check
> places like /etc/default/grub or /boot/grub/grub.cfg. You'd need to set
> something like GRUB_GFXPAYLOAD_LINUX=text for GRUB2.0, please check the GRUB
> docs to be sure. Afterwards we'd need another reg dump after you booted with
> nomodeset.
I don't have the hw any more, sorry.

>
> As I understood 3.12 worked with HDMI attached but 3.13 hanged. Could you do
> a
> kernel bisect between those two versions?
I don't have the hw any more, sorry.

>
> Thanks.
>
> ________________________________
> You are receiving this mail because:
>
> You reported the bug.
Comment 10 Imre Deak 2014-09-17 14:08:42 UTC
(In reply to comment #9)
> On Wed, Sep 17, 2014 at 3:45 PM,  <bugzilla-daemon@freedesktop.org> wrote:
> > Comment # 8 on bug 76738 from Imre Deak
> > As I understood 3.12 worked with HDMI attached but 3.13 hanged. Could you do
> > a kernel bisect between those two versions?
> I don't have the hw any more, sorry.
> > Thanks.

Ok, thanks for the feedback and sorry that we couldn't be of more help here. Closing the bug for now.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.