]> git.ipfire.org Git - thirdparty/kernel/stable.git/commit
ACPI: APEI: GHES: Add NVIDIA vendor CPER record handler
authorKai-Heng Feng <kaihengf@nvidia.com>
Mon, 30 Mar 2026 09:41:57 +0000 (17:41 +0800)
committerRafael J. Wysocki <rafael.j.wysocki@intel.com>
Mon, 6 Apr 2026 14:48:58 +0000 (16:48 +0200)
commitd7610855b0b5e934a35dedb02047a2419bf00770
tree9b0dda579e56d27c6eb44413c2ac9f80c71b4e48
parent35bdb5dbacf3ab4e4ee970ec2df2aa972ebbc21f
ACPI: APEI: GHES: Add NVIDIA vendor CPER record handler

Add support for decoding NVIDIA-specific CPER sections delivered via
the APEI GHES vendor record notifier chain. NVIDIA hardware generates
vendor-specific CPER sections containing error signatures and diagnostic
register dumps. This implementation registers a notifier_block with the
GHES vendor record notifier and decodes these sections, printing error
details via dev_info().

The driver binds to ACPI device NVDA2012, present on NVIDIA server
platforms. The NVIDIA CPER section contains a fixed header with error
metadata (signature, error type, severity, socket) followed by
variable-length register address-value pairs for hardware diagnostics.

This work is based on libcper [1].

Example output:
nvidia-ghes NVDA2012:00: NVIDIA CPER section, error_data_length: 544
nvidia-ghes NVDA2012:00: signature: CMET-INFO
nvidia-ghes NVDA2012:00: error_type: 0
nvidia-ghes NVDA2012:00: error_instance: 0
nvidia-ghes NVDA2012:00: severity: 3
nvidia-ghes NVDA2012:00: socket: 0
nvidia-ghes NVDA2012:00: number_regs: 32
nvidia-ghes NVDA2012:00: instance_base: 0x0000000000000000
nvidia-ghes NVDA2012:00: register[0]: address=0x8000000100000000 value=0x0000000100000000

https://github.com/openbmc/libcper/commit/683e055061ce [1]
Reviewed-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Signed-off-by: Kai-Heng Feng <kaihengf@nvidia.com>
[ rjw: Changelog edits ]
Link: https://patch.msgid.link/20260330094203.38022-4-kaihengf@nvidia.com
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
MAINTAINERS
drivers/acpi/apei/Kconfig
drivers/acpi/apei/Makefile
drivers/acpi/apei/ghes-nvidia.c [new file with mode: 0644]