From: Bjorn Helgaas <helgaas@kernel.org>
To: Sizhe Liu <liusizhe5@huawei.com>
Cc: bhelgaas@google.com, jonathan.cameron@huawei.com,
shiju.jose@huawei.com, pandoh@google.com,
linux-pci@vger.kernel.org, linuxarm@huawei.com,
prime.zeng@hisilicon.com, fanghao11@huawei.com,
shenyang39@huawei.com
Subject: Re: [PATCH v2] PCI/AER: Fix AER log missing in DPC case
Date: Fri, 6 Feb 2026 14:10:46 -0600 [thread overview]
Message-ID: <20260206201046.GA75132@bhelgaas> (raw)
In-Reply-To: <20260129140103.712011-1-liusizhe5@huawei.com>
On Thu, Jan 29, 2026 at 10:01:03PM +0800, Sizhe Liu wrote:
> In the current DPC error reporting case, some AER log information is missing.
>
> -- Error log abnormal
> pcieport 0000:20:00.0: DPC: containment event, status: 0x1f11: unmasked uncorrectable error detected
> (------ AER error log supposed to be printed here, but missing ------)
> nvme nvme0: frozen state error detected, reset controller
> {4}[Hardware Error]: Hardware error from APEI Generic Hardware Error Source: 0
>
> Cause:
> In aer_print_error(), PCIe AER errors is reported, and is rate-limited
> by info->ratelimit_print[i]. There are two entry points for
> aer_print_error().
>
> 1) Native AER
> aer_isr_one_error_type() -> aer_process_err_devices() ->
> aer_print_error()
> 2) DPC
> dpc_process_error() -> aer_print_error()
>
> The value of info->ratelimit_print[i] is initialized correctly in
> the native AER case:
> aer_isr_one_error_type() -> find_source_device() ->
> find_device_iter() -> add_error_device()
>
> In the DPC case, info->ratelimit_print[i] is not initialized and
> alloc by 0 , so in aer_print_error(), it will directly return at line
> if (!info->ratelimit_print[i])
> This will result in losing the AER log messages in the DPC case.
>
> Solution:
> 1. Move the initialization of info->ratelimit_print[i] to
> aer_ratelimit_print_init().
> 2. Add aer_ratelimit_print_init() in dpc_process_error().
> 3. Replace the initialization by aer_ratelimit_print_init()in
> Native AER case.
I see the problem, and I think you're right that we're not logging any
AER info for DPC events (including events handled via the EDR path,
which also calls dpc_process_error()).
Currently we do the ratelimit init in add_error_device(), which also
includes pci_dev_get() for the device. I don't see a similar
pci_dev_get() anywhere in the DPC path. There is one in the EDR path:
edr_handle_event
acpi_dpc_port_get
pci_dev_get <--
dpc_process_error
aer_get_device_error_info(aer_err_info)
aer_print_error(aer_err_info)
pcie_do_recovery
pci_dev_put
Maybe DPC and EDR should be using add_error_device() directly? It
seems like holding that reference on the device is important.
> Test with AER inject:
> Set the DPC reporting priority in the BIOS and send
> MalfTLP(AER FATAL ERROR) to device.
>
> -- Error log normal
> pcieport 0000:20:00.0: DPC: containment event, status:0x1f11: unmasked uncorrectable error detected
> pcieport 0000:20:00.0: PCIe Bus Error: severity=Uncorrectable (Fatal), type=Transaction Layer, (Receiver ID)
> pcieport 0000:20:00.0: device [19e5:a120] error status/mask=00040000/04580000
> pcieport 0000:20:00.0: [18] MalfTLP (First)
> pcieport 0000:20:00.0: AER: TLP Header: 0x00000000 0x00000000 0x00000000 0x00000000
> nvme nvme0: frozen state error detected, reset controller
> {2}[Hardware Error]: Hardware error from APEI Generic Hardware Error Source: 0
>
> [1] https://lore.kernel.org/linux-pci/20260127035405.712271-1-liusizhe5@huawei.com/
>
> Fixes: a57f2bfb4a58 ("PCI/AER: Ratelimit correctable and non-fatal error logging")
> Signed-off-by: Sizhe Liu <liusizhe5@huawei.com>
> ---
> v2
> - Corrected the format and spelling errors in the commit log.
>
> drivers/pci/pci.h | 1 +
> drivers/pci/pcie/aer.c | 35 +++++++++++++++++++++++------------
> drivers/pci/pcie/dpc.c | 1 +
> 3 files changed, 25 insertions(+), 12 deletions(-)
>
> diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h
> index 0e67014aa001..0cbcbcd52354 100644
> --- a/drivers/pci/pci.h
> +++ b/drivers/pci/pci.h
> @@ -748,6 +748,7 @@ struct aer_err_info {
>
> int aer_get_device_error_info(struct aer_err_info *info, int i);
> void aer_print_error(struct aer_err_info *info, int i);
> +void aer_ratelimit_print_init(struct pci_dev *dev, struct aer_err_info *e_info, int idx);
>
> int pcie_read_tlp_log(struct pci_dev *dev, int where, int where2,
> unsigned int tlp_len, bool flit,
> diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
> index e0bcaa896803..b73915b63327 100644
> --- a/drivers/pci/pcie/aer.c
> +++ b/drivers/pci/pcie/aer.c
> @@ -925,6 +925,28 @@ int cper_severity_to_aer(int cper_severity)
> EXPORT_SYMBOL_GPL(cper_severity_to_aer);
> #endif
>
> +/**
> + * aer_ratelimit_print_init - set flag whether error message is printed
> + * @dev: pointer to pci_dev to be rate-limited
> + * @e_info: pointer to error info
> + * @idx: index for ratelimit_print array
> + */
> +void aer_ratelimit_print_init(struct pci_dev *dev, struct aer_err_info *e_info, int idx)
> +{
> + /*
> + * Ratelimit AER log messages. "dev" is either the source
> + * identified by the root's Error Source ID or it has an unmasked
> + * error logged in its own AER Capability. Messages are emitted
> + * when "ratelimit_print[i]" is non-zero. If we will print detail
> + * for a downstream device, make sure we print the Error Source ID
> + * from the root as well.
> + */
> + if (aer_ratelimit(dev, e_info->severity)) {
> + e_info->ratelimit_print[idx] = 1;
> + e_info->root_ratelimit_print = 1;
> + }
> +}
> +
> void pci_print_aer(struct pci_dev *dev, int aer_severity,
> struct aer_capability_regs *aer)
> {
> @@ -990,18 +1012,7 @@ static int add_error_device(struct aer_err_info *e_info, struct pci_dev *dev)
> e_info->dev[i] = pci_dev_get(dev);
> e_info->error_dev_num++;
>
> - /*
> - * Ratelimit AER log messages. "dev" is either the source
> - * identified by the root's Error Source ID or it has an unmasked
> - * error logged in its own AER Capability. Messages are emitted
> - * when "ratelimit_print[i]" is non-zero. If we will print detail
> - * for a downstream device, make sure we print the Error Source ID
> - * from the root as well.
> - */
> - if (aer_ratelimit(dev, e_info->severity)) {
> - e_info->ratelimit_print[i] = 1;
> - e_info->root_ratelimit_print = 1;
> - }
> + aer_ratelimit_print_init(dev, e_info, i);
> return 0;
> }
>
> diff --git a/drivers/pci/pcie/dpc.c b/drivers/pci/pcie/dpc.c
> index fc18349614d7..d17adc642781 100644
> --- a/drivers/pci/pcie/dpc.c
> +++ b/drivers/pci/pcie/dpc.c
> @@ -275,6 +275,7 @@ void dpc_process_error(struct pci_dev *pdev)
> status);
> if (dpc_get_aer_uncorrect_severity(pdev, &info) &&
> aer_get_device_error_info(&info, 0)) {
> + aer_ratelimit_print_init(pdev, &info, 0);
> aer_print_error(&info, 0);
> pci_aer_clear_nonfatal_status(pdev);
> pci_aer_clear_fatal_status(pdev);
> --
> 2.33.0
>
next prev parent reply other threads:[~2026-02-06 20:10 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-01-29 14:01 [PATCH v2] PCI/AER: Fix AER log missing in DPC case Sizhe Liu
2026-02-06 10:00 ` Sizhe LIU
2026-02-06 20:10 ` Bjorn Helgaas [this message]
2026-02-10 8:00 ` Sizhe LIU
2026-02-11 12:18 ` Sizhe Liu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260206201046.GA75132@bhelgaas \
--to=helgaas@kernel.org \
--cc=bhelgaas@google.com \
--cc=fanghao11@huawei.com \
--cc=jonathan.cameron@huawei.com \
--cc=linux-pci@vger.kernel.org \
--cc=linuxarm@huawei.com \
--cc=liusizhe5@huawei.com \
--cc=pandoh@google.com \
--cc=prime.zeng@hisilicon.com \
--cc=shenyang39@huawei.com \
--cc=shiju.jose@huawei.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox