Dan Williams wrote: > Initial tests with the CXL CPER implementation identified that error > reports were being duplicated in the log and the trace event [1]. Then > it was discovered that the notification handler took sleeping locks > while the GHES event handling runs in spin_lock_irqsave() context [2] > > Given multiple bugs to fix and how late it is in the development cycle, > remove the CXL hookup for now and try again during the next merge > window. > > Cc: Ard Biesheuvel <ardb@xxxxxxxxxx> > Cc: Rafael J. Wysocki <rafael@xxxxxxxxxx> > Cc: Ira Weiny <ira.weiny@xxxxxxxxx> > Cc: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx> > Link: http://lore.kernel.org/r/20240108165855.00002f5a@xxxxxxxxxx [1] Stephen noticed that the fix for this already went upstream as: 54ce1927eb78 ("cxl/cper: Fix errant CPER prints for CXL events") ...and it collides with this removal. I really do not want to have a fire drill to fix locking this late in the cycle, so I still think trying again for v6.9 is more comfortable. That also allows collecting Smita's work as well. I will fix up the changelog to: --- Initial tests with the CXL CPER implementation identified that error reports were being duplicated in the log and the trace event [1]. Then it was discovered that the notification handler took sleeping locks while the GHES event handling runs in spin_lock_irqsave() context [2] While the duplicate reporting was fixed in v6.8-rc4, the fix for the sleeping-lock-vs-atomic collision would enjoy more time to settle and gain some test cycles. Given how late it is in the development cycle, remove the CXL hookup for now and try again during the next merge window. Note that end result is that v6.8 does not emit CXL CPER payloads to the kernel log, but this is in line with the CXL trend to move error reporting to trace events instead of the kernel log. ---
