On Thu 2025-11-13 17:09:46, John Ogness wrote:
This is v2 of a series to address multiple reports [0][1] (+ 2 offlist) of suspend failing when NBCON console drivers are in use. With the help of NXP and NVIDIA we were able to isolate the problem and verify the fix.
v1 is here [2].
The first NBCON drivers appeared in 6.13, so currently there is no LTS kernel that requires this series. But it should go into 6.17.x and 6.18.
The changes since v1:
For printk_trigger_flush() add support for all flush types that are available. This will prevent printk_trigger_flush() from trying to inappropriately queue irq_work after this series is applied.
Add WARN_ON_ONCE() to the printk irq_work queueing functions in case they are called when irq_work is blocked. There should never be (and currently are no) such callers, but these functions are externally available.
John Ogness
[0] https://lore.kernel.org/lkml/80b020fc-c18a-4da4-b222-16da1cab2f4c@nvidia.com [1] https://lore.kernel.org/lkml/DB9PR04MB8429E7DDF2D93C2695DE401D92C4A@DB9PR04M... [2] https://lore.kernel.org/lkml/20251111144328.887159-1-john.ogness@linutronix....
John Ogness (2): printk: Allow printk_trigger_flush() to flush all types printk: Avoid scheduling irq_work on suspend
kernel/printk/internal.h | 8 ++-- kernel/printk/nbcon.c | 9 ++++- kernel/printk/printk.c | 81 ++++++++++++++++++++++++++++++++-------- 3 files changed, 78 insertions(+), 20 deletions(-)
The patchset seems to be ready for linux-next from my POV. I am going to wait few more days for potential feedback. I'll push it later the following week unless anyone complains.
Best Regards, Petr