]> www.infradead.org Git - users/jedix/linux-maple.git/commit
accel/habanalabs: no CPUCP prints on heartbeat failure
authorOhad Sharabi <osharabi@habana.ai>
Sun, 18 Feb 2024 09:54:29 +0000 (11:54 +0200)
committerOfir Bitton <obitton@habana.ai>
Sun, 23 Jun 2024 06:45:51 +0000 (09:45 +0300)
commitecda35d461f491f7aa2a0f2060a2c36a5950c3b7
tree3f511fcf186f38620b327c07f602f1046e05d385
parent467cfe945656df044c8cf9121e5cdbe5b977b497
accel/habanalabs: no CPUCP prints on heartbeat failure

If we detected heartbet event while some daemon in the background send
(via driver interface) CPUCP messages the dmesg will be flooded.

Instead, a slight refactor in hl_fw_send_cpu_message() returns -EAGAIN
when CPU is disabled (i.e. heartbeat failure) and only then.

Later, all calling functions that may be invoked by user space can issue
prints only if the error code is not -EAGAIN.

Signed-off-by: Ohad Sharabi <osharabi@habana.ai>
Reviewed-by: Ofir Bitton <obitton@habana.ai>
Signed-off-by: Ofir Bitton <obitton@habana.ai>
drivers/accel/habanalabs/common/debugfs.c
drivers/accel/habanalabs/common/device.c
drivers/accel/habanalabs/common/firmware_if.c
drivers/accel/habanalabs/common/hwmon.c
drivers/accel/habanalabs/gaudi/gaudi.c
drivers/accel/habanalabs/gaudi2/gaudi2.c
drivers/accel/habanalabs/goya/goya.c