Still time for v4.2? - c0165304e10f ("drm/i915: Only enable cursor if it can be enabled.")

Still time for v4.2? - c0165304e10f ("drm/i915: Only enable cursor if it can be enabled.")

[Bjørn Mork]

Hello, I see that I consistently get the warning below on v4.2-rc8.  I
believe you already fixed this a long time ago by commit c0165304e10f
("drm/i915: Only enable cursor if it can be enabled."), which I assume is
a stable candidate for v4.2.y.

But wouldn't it look better if you managed to squeeze this into v4.2?
The warning will probably cause a few unnecessary worries and reports
from users...

Sample WARNING from v4.2-rc8, minus the irrelevant list of modules:
------------[ cut here ]------------
WARNING: CPU: 1 PID: 2675 at drivers/gpu/drm/i915/intel_display.c:9669 intel_crtc_update_cursor+0x335/0x452 [i915]()
Missing switch case (0) in i9xx_update_cursor
CPU: 1 PID: 2675 Comm: Xorg Tainted: G        W       4.2.0-rc8 #339
Hardware name: LENOVO 2776LEG/2776LEG, BIOS 6EET55WW (3.15 ) 12/19/2011
 0000000000000009 ffff88022ca07b68 ffffffff81423887 0000000000000006
 ffff88022ca07bb8 ffff88022ca07ba8 ffffffff8104a9d6 0000000000070088
 ffffffffa036b141 ffff8802317c9000 000000000053e000 ffff880231790000
Call Trace:
 [<ffffffff81423887>] dump_stack+0x4c/0x65
 [<ffffffff8104a9d6>] warn_slowpath_common+0xa1/0xbb
 [<ffffffffa036b141>] ? intel_crtc_update_cursor+0x335/0x452 [i915]
 [<ffffffff8104aa36>] warn_slowpath_fmt+0x46/0x48
 [<ffffffffa035cf66>] ? gen2_write32+0x35/0x44 [i915]
 [<ffffffffa036b141>] intel_crtc_update_cursor+0x335/0x452 [i915]
 [<ffffffffa0372afa>] intel_crtc_enable_planes+0xce/0xfd [i915]
 [<ffffffffa03748b4>] intel_crtc_control+0x7d/0xe1 [i915]
 [<ffffffffa0374965>] intel_crtc_update_dpms+0x4d/0x5d [i915]
 [<ffffffffa0379115>] intel_connector_dpms+0x49/0x55 [i915]
 [<ffffffffa02a3c36>] drm_mode_obj_set_property_ioctl+0x103/0x1b3 [drm]
 [<ffffffffa02a3d16>] drm_mode_connector_property_set_ioctl+0x30/0x32 [drm]
 [<ffffffffa0295be5>] drm_ioctl+0x37d/0x421 [drm]
 [<ffffffffa02a3ce6>] ? drm_mode_obj_set_property_ioctl+0x1b3/0x1b3 [drm]
 [<ffffffff81073d35>] ? local_clock+0x19/0x22
 [<ffffffff811799ca>] do_vfs_ioctl+0x430/0x4ef
 [<ffffffff81054ec0>] ? do_sigaltstack+0xbe/0x16a
 [<ffffffff81182c06>] ? __fget_light+0x50/0x78
 [<ffffffff81179ae3>] SyS_ioctl+0x5a/0x7f
 [<ffffffff81429557>] entry_SYSCALL_64_fastpath+0x12/0x6f
---[ end trace 7c3e6448deed5ebc ]---


Thanks,
Bjørn
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: Still time for v4.2? - c0165304e10f ("drm/i915: Only enable cursor if it can be enabled.")

[Maarten Lankhorst]

Hey,

Op 25-08-15 om 09:45 schreef Bjørn Mork:
> Hello, I see that I consistently get the warning below on v4.2-rc8.  I
> believe you already fixed this a long time ago by commit c0165304e10f
> ("drm/i915: Only enable cursor if it can be enabled."), which I assume is
> a stable candidate for v4.2.y.
>
> But wouldn't it look better if you managed to squeeze this into v4.2?
> The warning will probably cause a few unnecessary worries and reports
> from users...
>
> Sample WARNING from v4.2-rc8, minus the irrelevant list of modules:
> ------------[ cut here ]------------
> WARNING: CPU: 1 PID: 2675 at drivers/gpu/drm/i915/intel_display.c:9669 intel_crtc_update_cursor+0x335/0x452 [i915]()
> Missing switch case (0) in i9xx_update_cursor
> CPU: 1 PID: 2675 Comm: Xorg Tainted: G        W       4.2.0-rc8 #339
> Hardware name: LENOVO 2776LEG/2776LEG, BIOS 6EET55WW (3.15 ) 12/19/2011
>  0000000000000009 ffff88022ca07b68 ffffffff81423887 0000000000000006
>  ffff88022ca07bb8 ffff88022ca07ba8 ffffffff8104a9d6 0000000000070088
>  ffffffffa036b141 ffff8802317c9000 000000000053e000 ffff880231790000
What are you doing that triggers this? Works for me on v4.2..

~Maarten
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: Still time for v4.2? - c0165304e10f ("drm/i915: Only enable cursor if it can be enabled.")

[Bjørn Mork]

Maarten Lankhorst <maarten.lankhorst@linux.intel.com> writes:

> Hey,
>
> Op 25-08-15 om 09:45 schreef Bjørn Mork:
>> Hello, I see that I consistently get the warning below on v4.2-rc8.  I
>> believe you already fixed this a long time ago by commit c0165304e10f
>> ("drm/i915: Only enable cursor if it can be enabled."), which I assume is
>> a stable candidate for v4.2.y.
>>
>> But wouldn't it look better if you managed to squeeze this into v4.2?
>> The warning will probably cause a few unnecessary worries and reports
>> from users...
>>
>> Sample WARNING from v4.2-rc8, minus the irrelevant list of modules:
>> ------------[ cut here ]------------
>> WARNING: CPU: 1 PID: 2675 at drivers/gpu/drm/i915/intel_display.c:9669 intel_crtc_update_cursor+0x335/0x452 [i915]()
>> Missing switch case (0) in i9xx_update_cursor
>> CPU: 1 PID: 2675 Comm: Xorg Tainted: G        W       4.2.0-rc8 #339
>> Hardware name: LENOVO 2776LEG/2776LEG, BIOS 6EET55WW (3.15 ) 12/19/2011
>>  0000000000000009 ffff88022ca07b68 ffffffff81423887 0000000000000006
>>  ffff88022ca07bb8 ffff88022ca07ba8 ffffffff8104a9d6 0000000000070088
>>  ffffffffa036b141 ffff8802317c9000 000000000053e000 ffff880231790000
> What are you doing that triggers this? Works for me on v4.2..

That's a really good question.  I saw the warnings a couple of times
after resuming from suspend (S3) and made a too hasty assumption that
resuming would always trigger it.

But now that I'm trying to provoke it, I cannot reproduce it whatever I
do...

FWIW, my suspend scripts run "xscreensaver-command -lock" to make sure
the screen is at least locked if someone else should wake the PC, and
xscreensaver is set up to blank and power off the screen.  I don't know
if this affects the results.


Bjørn
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: Still time for v4.2? - c0165304e10f ("drm/i915: Only enable cursor if it can be enabled.")

[Bjørn Mork]

Just for the record: I still get these quite often on resuming from
suspend-to-ram.  But I can't reliably provoke them for some reason, so
the exact trigger mechanism is unknown.  Related to timing issues caused
by other devices, maybe?  Or some odd Lenovo firmware thingy?

Does of course not matter much to me.  Just wanted to let you know in
case you see similar reports.


------------[ cut here ]------------
WARNING: CPU: 1 PID: 2682 at drivers/gpu/drm/i915/intel_display.c:9669 intel_crtc_update_cursor+0x335/0x452 [i915]()
Missing switch case (0) in i9xx_update_cursor
Modules linked in: cpuid nbd xt_nat xt_tcpudp ipt_REJECT nf_reject_ipv4 drbg ansi_cprng ctr ccm nfnetlink_queue nfnetlink_log nfnetlink bridge xt_conntrack ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack cdc_mbim cdc_wdm cdc_ncm usbnet cdc_acm mii usb_serial_simple usbserial xt_multiport iptable_filter ip_tables bnep xt_hl nf_log_ipv6 nf_log_common xt_LOG binfmt_misc ip6table_filter ip6_tables x_tables nfsd nfs_acl nfs lockd grace fscache sunrpc 8021q garp mrp stp llc tun fuse loop iTCO_wdt iTCO_vendor_support snd_hda_codec_conexant snd_hda_codec_generic coretemp kvm_intel kvm psmouse evdev serio_raw uvcvideo videobuf2_vmalloc videobuf2_memops arc4 videobuf2_core v4l2_common videodev btusb btrtl btbcm btintel bluetooth
 i2c_i801 lpc_ich iwlmvm mfd_core mac80211 iwlwifi snd_hda_intel snd_hda_codec snd_hwdep snd_hda_core cfg80211 snd_pcm_oss snd_mixer_oss snd_pcm snd_timer wmi i915 i2c_algo_bit thinkpad_acpi nvram snd soundcore acpi_cpufreq rfkill ac drm_kms_helper battery drm button video processor ext4 crc16 jbd2 mbcache microcode sg sd_mod sr_mod cdrom ahci libahci libata scsi_mod ehci_pci uhci_hcd ehci_hcd usbcore usb_common e1000e ptp thermal pps_core thermal_sys [last unloaded: nbd]
CPU: 1 PID: 2682 Comm: Xorg Tainted: G        W       4.2.0 #340
Hardware name: LENOVO 2776LEG/2776LEG, BIOS 6EET55WW (3.15 ) 12/19/2011
 0000000000000009 ffff8800b69c3b68 ffffffff814238a7 0000000000000006
 ffff8800b69c3bb8 ffff8800b69c3ba8 ffffffff8104a9d6 0000000000070088
 ffffffffa03a5141 ffff8802317f2000 0000000001154000 ffff880231790000
Call Trace:
 [<ffffffff814238a7>] dump_stack+0x4c/0x65
 [<ffffffff8104a9d6>] warn_slowpath_common+0xa1/0xbb
 [<ffffffffa03a5141>] ? intel_crtc_update_cursor+0x335/0x452 [i915]
 [<ffffffff8104aa36>] warn_slowpath_fmt+0x46/0x48
 [<ffffffffa0396f66>] ? gen2_write32+0x35/0x44 [i915]
 [<ffffffffa03a5141>] intel_crtc_update_cursor+0x335/0x452 [i915]
 [<ffffffffa03acafa>] intel_crtc_enable_planes+0xce/0xfd [i915]
 [<ffffffffa03ae8b4>] intel_crtc_control+0x7d/0xe1 [i915]
 [<ffffffffa03ae965>] intel_crtc_update_dpms+0x4d/0x5d [i915]
 [<ffffffffa03b3115>] intel_connector_dpms+0x49/0x55 [i915]
 [<ffffffffa02fcc36>] drm_mode_obj_set_property_ioctl+0x103/0x1b3 [drm]
 [<ffffffffa02fcd16>] drm_mode_connector_property_set_ioctl+0x30/0x32 [drm]
 [<ffffffffa02eebe5>] drm_ioctl+0x37d/0x421 [drm]
 [<ffffffffa02fcce6>] ? drm_mode_obj_set_property_ioctl+0x1b3/0x1b3 [drm]
 [<ffffffff81073d35>] ? local_clock+0x19/0x22
 [<ffffffff811799ca>] do_vfs_ioctl+0x430/0x4ef
 [<ffffffff81054ec0>] ? do_sigaltstack+0xbe/0x16a
 [<ffffffff81182c06>] ? __fget_light+0x50/0x78
 [<ffffffff81179ae3>] SyS_ioctl+0x5a/0x7f
 [<ffffffff81429597>] entry_SYSCALL_64_fastpath+0x12/0x6f
---[ end trace 98ba9856e0d6d32c ]---



Bjørn
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: Still time for v4.2? - c0165304e10f ("drm/i915: Only enable cursor if it can be enabled.")

[Maarten Lankhorst]

Op 08-09-15 om 09:31 schreef Bjørn Mork:
> Just for the record: I still get these quite often on resuming from
> suspend-to-ram.  But I can't reliably provoke them for some reason, so
> the exact trigger mechanism is unknown.  Related to timing issues caused
> by other devices, maybe?  Or some odd Lenovo firmware thingy?
>
> Does of course not matter much to me.  Just wanted to let you know in
> case you see similar reports.
>
Does it occur when forcing suspend with dpms? xset dpms force off.


_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: Still time for v4.2? - c0165304e10f ("drm/i915: Only enable cursor if it can be enabled.")

[Bjørn Mork]

Maarten Lankhorst <maarten.lankhorst@linux.intel.com> writes:

> Op 08-09-15 om 09:31 schreef Bjørn Mork:
>> Just for the record: I still get these quite often on resuming from
>> suspend-to-ram.  But I can't reliably provoke them for some reason, so
>> the exact trigger mechanism is unknown.  Related to timing issues caused
>> by other devices, maybe?  Or some odd Lenovo firmware thingy?
>>
>> Does of course not matter much to me.  Just wanted to let you know in
>> case you see similar reports.
>>
> Does it occur when forcing suspend with dpms? xset dpms force off.

Nope



Bjørn
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx