mbox series

[0/3] PRU firmware event/interrupt mapping fixes

Message ID 20210323223839.17464-1-s-anna@ti.com (mailing list archive)
Headers show
Series PRU firmware event/interrupt mapping fixes | expand


Suman Anna March 23, 2021, 10:38 p.m. UTC

The following series includes fixes for various different issues
associated with the PRU firmware event/interrupt mapping configuration
logic added in the same commit c75c9fdac66e ("remoteproc: pru: Add
support for PRU specific interrupt configuration"). The fixes are
agnostic of SoC family.

Following is the summary of issues and fixes:
 - Patch #1 fixes the interrupt node finding logic to always use the
   inherent sibling relationship between a PRU/RTU/Tx_PRU node and its
   corresponding PRUSS INTC node. This fixes the firmware event mappings
   for cases when the PRU nodes do not have an 'interrupt-parent' property
   (this is the norm, the property is neither required nor added in the DT
   nodes normally).
 - Patch #2 fixes a minor issue with returning a success value to the
   caller on a fw event mapping failure.
 - Patch #3 fixes a kernel crash due to switching of firmwares between
   consecutive runs, the first one with events and the second one without
   events. There are no issues when the same firmwares are run or if they
   are run in reverse order.

Patches should apply cleanly on top of the current rproc-fixes branch
commit 9afeefcf06fc ("remoteproc: pru: Fix firmware loading crashes on K3 SoCs")


Suman Anna (3):
  remoteproc: pru: Fixup interrupt-parent logic for fw events
  remoteproc: pru: Fix wrong success return value for fw events
  remoteproc: pru: Fix and cleanup firmware interrupt mapping logic

 drivers/remoteproc/pru_rproc.c | 33 +++++++++++++++++++++++++--------
 1 file changed, 25 insertions(+), 8 deletions(-)