Re: [PATCH tty-next v5 5/6] serial: 8250: Switch to nbcon console

[Petr Mladek <pmladek@suse.com>]

On Mon 2025-01-27 14:54:25, Jon Hunter wrote:
> Hi John,
> 
> On 20/01/2025 16:34, Thierry Reding wrote:
> > On Mon, Jan 20, 2025 at 05:23:26PM +0100, Thierry Reding wrote:
> > > On Thu, Jan 16, 2025 at 10:41:08AM +0000, Jon Hunter wrote:
> > > > 
> > > > On 16/01/2025 10:38, John Ogness wrote:
> > > > > On 2025-01-16, Jon Hunter <jonathanh@nvidia.com> wrote:
> > > > > > > Do you at least know if it is failing to suspend or failing to resume
> > > > > > > (based on power consumption)?
> > > > > > 
> > > > > > 
> > > > > > Unfortunately, I don't. These are farm boards and so nothing local I can
> > > > > > get my hands on. For some reason all the serial console logs are not
> > > > > > available and so I am going to talk to the farm team about fixing that
> > > > > > because we should at least have serial logs.
> > > > > 
> > > > > Can you confirm that the board is actually booting? The suspend code for
> > > > > 8250_tegra.c is quite simple. I am wondering if the farm tests are
> > > > > failing somewhere else, such as the atomic printing during early boot.
> > > > 
> > > > 
> > > > Yes they are all booting fine. I have an independent boot test and that is
> > > > passing. It is just the suspend test that is failing.
> > > 
> > > I was able to capture logs, but unfortunately they don't provide much
> > > insight either. On the first try it doesn't suspend and goes back to
> > > userspace after a second or so:
> > > 
> > > --- >8 ---
> > > -sh-5.1# rtcwake --device /dev/rtc1 --mode mem --seconds 5
> > > rtcwake: assuming RTC uses UTC ...
> > > rtcwake: wakeup from "mem" using /dev/rtc1 at Thu Jan  1 00:01:00 1970
> > > [   36.332486] PM: suspend entry (deep)
> > > [   36.332832] Filesystems sync: 0.000 seconds
> > > [   36.369331] +1.8V_RUN_CAM: disabling
> > > [   36.373884] +2.8V_RUN_CAM: disabling
> > > [   36.375571] +1.2V_RUN_CAM_FRONT: disabling
> > > [   36.380359] +1.05V_RUN_CAM_REAR: disabling
> > > [   36.387399] +3.3V_RUN_TOUCH: disabling
> > > [   36.390808] +2.8V_RUN_CAM_AF: disabling
> > > [   36.393621] +1.8V_RUN_VPP_FUSE: disabling
> > > [   36.408218] Freezing user space processes
> > > [   36.413660] Freezing user space processes completed (elapsed 0.005 seconds)
> > > [   36.413680] OOM killer disabled.
> > > [   36.413693] Freezing remaining freezable tasks
> > > [   36.415033] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
> > > [   36.428474] drm drm: [drm:drm_client_dev_suspend] fbdev: ret=0
> > > [   36.428527] drm drm: [drm:drm_atomic_state_init] Allocated atomic state 2e5cd010
> > > [   36.428547] drm drm: [drm:drm_atomic_get_crtc_state] Added [CRTC:47:crtc-0] 6a6be0ef state to 2e5cd010
> > > [   36.428561] drm drm: [drm:drm_atomic_get_crtc_state] Added [CRTC:63:crtc-1] 00d818c2 state to 2e5cd010
> > > [   36.428574] drm drm: [drm:drm_atomic_get_plane_state] Added [PLANE:32:plane-0] 4e145b7d state to 2e5cd010
> > > [   36.428587] drm drm: [drm:drm_atomic_get_plane_state] Added [PLANE:36:plane-1] dbf67d12 state to 2e5cd010
> > > [   36.428597] drm drm: [drm:drm_atomic_get_plane_state] Added [PLANE:40:plane-2] 763d8809 state to 2e5cd010
> > > [   36.428608] drm drm: [drm:drm_atomic_get_plane_state] Added [PLANE:44:plane-3] b6eabcf1 state to 2e5cd010
> > > [   36.428617] drm drm: [drm:drm_atomic_get_plane_state] Added [PLANE:48:plane-4] 7863878c state to 2e5cd010
> > > [   36.428628] drm drm: [drm:drm_atomic_get_plane_state] Added [PLANE:52:plane-5] 54b8029c state to 2e5cd010
> > > [   36.428638] drm drm: [drm:drm_atomic_get_plane_state] Added [PLANE:56:plane-6] 364063af state to 2e5cd010
> > > [   36.428648] drm drm: [drm:drm_atomic_get_plane_state] Added [PLANE:60:plane-7] e1c11dfb state to 2e5cd010
> > > [   36.428662] drm drm: [drm:drm_atomic_get_connector_state] Added [CONNECTOR:65:HDMI-A-1] 5cb32770 state to 2e5cd010
> > > [   36.428674] drm drm: [drm:drm_atomic_state_init] Allocated atomic state 832943c7
> > > [   36.428682] drm drm: [drm:drm_atomic_get_crtc_state] Added [CRTC:47:crtc-0] f09cf73d state to 832943c7
> > > [   36.428691] drm drm: [drm:drm_atomic_add_affected_planes] Adding all current planes for [CRTC:47:crtc-0] to 832943c7
> > > [   36.428700] drm drm: [drm:drm_atomic_add_affected_connectors] Adding all current connectors for [CRTC:47:crtc-0] to 832943c7
> > > [   36.428711] drm drm: [drm:drm_atomic_get_crtc_state] Added [CRTC:63:crtc-1] 2700922c state to 832943c7
> > > [   36.428720] drm drm: [drm:drm_atomic_add_affected_planes] Adding all current planes for [CRTC:63:crtc-1] to 832943c7
> > > [   36.428727] drm drm: [drm:drm_atomic_add_affected_connectors] Adding all current connectors for [CRTC:63:crtc-1] to 832943c7
> > > [   36.428737] drm drm: [drm:drm_atomic_check_only] checking 832943c7
> > > [   36.428759] drm drm: [drm:drm_atomic_commit] committing 832943c7
> > > [   36.428881] drm drm: [drm:drm_atomic_state_default_clear] Clearing atomic state 832943c7
> > > [   36.428897] drm drm: [drm:__drm_atomic_state_free] Freeing atomic state 832943c7
> > > [   36.429085] r8169 0000:01:00.0 eth0: Link is Down
> > > [   36.713236] Disabling non-boot CPUs ...
> > > -sh-5.1#
> > > --- >8 ---
> > > 
> > > A second attempt soft-hangs:
> > > 
> > > --- >8 ---
> > > -sh-5.1# rtcwake --device /dev/rtc1 --mode mem --seconds 5
> > > rtcwake: assuming RTC uses UTC ...
> > > rtcwake: wakeup from "mem" using /dev/rtc1 at Thu Jan  1 00:01:10 1970
> > > --- >8 ---
> > > 
> > > Where "soft-hang" means it doesn't do anything after this and I can't
> > > SIGINT out of it or anything. However, the serial seems to still be
> > > responsive.
> > 
> > To clarify, this was on top of next-20250120 and reverting the patches
> > that Jon mentioned suspend/resume is fixed for me as well.
> > 
> > I do have a local device that I can test on, so if there's any patches
> > you want me to try, or any options to enable to get more information,
> > please let me know.
> 
> 
> Any feedback on this? Our boards are still broken with this change.

AFAIK, this patch has been reverted in linux-next last week, see
https://lore.kernel.org/r/84wmenqm03.fsf@jogness.linutronix.de

John had a training. I believe that he would look at these
suspend-related problems before sending another revision
of the patchset.

Best Regards,
Petr