[PATCH] HID: playstation: ratelimit input CRC check failed messages

[PATCH] HID: playstation: ratelimit input CRC check failed messages

[Mauricio Faria de Oliveira]

The kernel log of a Steam Deck was mostly filled with CRC check errors
from what is apparently a bogus controller, with 8174 messages in <200
seconds, averaging ~43 messages/second:

    $ cat dmesg-pstore.202610250608-0 \
      | grep -c "input CRC's check failed"
    8174

    $ cat dmesg-pstore.202610250608-0 \
      | grep -m1 "input CRC's check failed"
    <3>[    8.715455] playstation 0005:054C:05C4.0006:
                      DualShock4 input CRC's check failed

    $ tac dmesg-pstore.202610250608-0 \
      | grep -m1 "input CRC's check failed"
    <3>[  196.187459] playstation 0005:054C:05C4.0006:
                      DualShock4 input CRC's check failed

    $ echo $((8174 / (196-8)))
    43

Let's ratelimit the input CRC check failed messages for the exceptional,
high-frequency error case, which doesn't impact the occasional, low-freq
error case (as ratelimiting does not trigger) or the non-error case.

CRC check failed messages for driver-initiated requests are kept as is.

Testing with a UHID driver to synthetically reproduce input CRC check
errors (generated by an LLM) shows that it reduces the duplicate lines
and still provides useful information (number of errors/same message).

For example, triggering such error every 25ms (40 times/second), prints
10 messages and silently counts another 187 in 5 seconds, ie, 197 / 5 =
39.4 times/second (aligned with the synthetic reproducer's rate):

    [    9.554184] playstation ...: Registered DualShock4 controller ...
    [   10.053936] playstation ...: DualShock4 input CRC's check failed
    [   10.078162] playstation ...: DualShock4 input CRC's check failed
    [   10.103447] playstation ...: DualShock4 input CRC's check failed
    [   10.128739] playstation ...: DualShock4 input CRC's check failed
    [   10.154016] playstation ...: DualShock4 input CRC's check failed
    [   10.179198] playstation ...: DualShock4 input CRC's check failed
    [   10.204474] playstation ...: DualShock4 input CRC's check failed
    [   10.229762] playstation ...: DualShock4 input CRC's check failed
    [   10.254904] playstation ...: DualShock4 input CRC's check failed
    [   10.280441] playstation ...: DualShock4 input CRC's check failed
    [   15.070695] dualshock4_parse_report: 187 callbacks suppressed
    ...

This reduces the space consumed in the kernel log during 5 seconds from
~200 lines to only 11 lines (5.5%).

Signed-off-by: Mauricio Faria de Oliveira <mfo@igalia.com>
---
 drivers/hid/hid-playstation.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/hid/hid-playstation.c b/drivers/hid/hid-playstation.c
index f9dc9378cf777d33aab3cdb87302c6e48f406d2f..b053037a3f0b95b6986c7e53e787e005676ed565 100644
--- a/drivers/hid/hid-playstation.c
+++ b/drivers/hid/hid-playstation.c
@@ -824,6 +824,7 @@ static int ps_get_report(struct hid_device *hdev, u8 report_id, u8 *buf,
 		u32 report_crc = get_unaligned_le32(&buf[crc_offset]);
 
 		if (!ps_check_crc32(PS_FEATURE_CRC32_SEED, buf, crc_offset, report_crc)) {
+			/* hid_err_ratelimited() not used on driver-initiated requests */
 			hid_err(hdev, "CRC check failed for reportID=%d\n", report_id);
 			return -EILSEQ;
 		}
@@ -1459,7 +1460,7 @@ static int dualsense_parse_report(struct ps_device *ps_dev, struct hid_report *r
 		u32 report_crc = get_unaligned_le32(&data[size - 4]);
 
 		if (!ps_check_crc32(PS_INPUT_CRC32_SEED, data, size - 4, report_crc)) {
-			hid_err(hdev, "DualSense input CRC's check failed\n");
+			hid_err_ratelimited(hdev, "DualSense input CRC's check failed\n");
 			return -EILSEQ;
 		}
 
@@ -2425,7 +2426,7 @@ static int dualshock4_parse_report(struct ps_device *ps_dev, struct hid_report *
 
 		/* Last 4 bytes of input report contains CRC. */
 		if (!ps_check_crc32(PS_INPUT_CRC32_SEED, data, size - 4, report_crc)) {
-			hid_err(hdev, "DualShock4 input CRC's check failed\n");
+			hid_err_ratelimited(hdev, "DualShock4 input CRC's check failed\n");
 			return -EILSEQ;
 		}
 

---
base-commit: 83f32c773255720a661aa45d83e36278135f23f5
change-id: 20261027-hid-playstation-ratelimit-input-crc-error-26f871f1c9d2

Best regards,
-- 
Mauricio Faria de Oliveira <mfo@igalia.com>

Re: [PATCH] HID: playstation: ratelimit input CRC check failed messages

[Roderick Colenbrander]

Hi Mauricio,

This looks reasonable to reduce log spam.
Signed-off-by: roderick.colenbrander@sony.com

Thanks,
Roderick Colenbrander



On Tue, Oct 27, 2026 at 5:36 PM Mauricio Faria de Oliveira
<mfo@igalia.com> wrote:
>
> The kernel log of a Steam Deck was mostly filled with CRC check errors
> from what is apparently a bogus controller, with 8174 messages in <200
> seconds, averaging ~43 messages/second:
>
>     $ cat dmesg-pstore.202610250608-0 \
>       | grep -c "input CRC's check failed"
>     8174
>
>     $ cat dmesg-pstore.202610250608-0 \
>       | grep -m1 "input CRC's check failed"
>     <3>[    8.715455] playstation 0005:054C:05C4.0006:
>                       DualShock4 input CRC's check failed
>
>     $ tac dmesg-pstore.202610250608-0 \
>       | grep -m1 "input CRC's check failed"
>     <3>[  196.187459] playstation 0005:054C:05C4.0006:
>                       DualShock4 input CRC's check failed
>
>     $ echo $((8174 / (196-8)))
>     43
>
> Let's ratelimit the input CRC check failed messages for the exceptional,
> high-frequency error case, which doesn't impact the occasional, low-freq
> error case (as ratelimiting does not trigger) or the non-error case.
>
> CRC check failed messages for driver-initiated requests are kept as is.
>
> Testing with a UHID driver to synthetically reproduce input CRC check
> errors (generated by an LLM) shows that it reduces the duplicate lines
> and still provides useful information (number of errors/same message).
>
> For example, triggering such error every 25ms (40 times/second), prints
> 10 messages and silently counts another 187 in 5 seconds, ie, 197 / 5 =
> 39.4 times/second (aligned with the synthetic reproducer's rate):
>
>     [    9.554184] playstation ...: Registered DualShock4 controller ...
>     [   10.053936] playstation ...: DualShock4 input CRC's check failed
>     [   10.078162] playstation ...: DualShock4 input CRC's check failed
>     [   10.103447] playstation ...: DualShock4 input CRC's check failed
>     [   10.128739] playstation ...: DualShock4 input CRC's check failed
>     [   10.154016] playstation ...: DualShock4 input CRC's check failed
>     [   10.179198] playstation ...: DualShock4 input CRC's check failed
>     [   10.204474] playstation ...: DualShock4 input CRC's check failed
>     [   10.229762] playstation ...: DualShock4 input CRC's check failed
>     [   10.254904] playstation ...: DualShock4 input CRC's check failed
>     [   10.280441] playstation ...: DualShock4 input CRC's check failed
>     [   15.070695] dualshock4_parse_report: 187 callbacks suppressed
>     ...
>
> This reduces the space consumed in the kernel log during 5 seconds from
> ~200 lines to only 11 lines (5.5%).
>
> Signed-off-by: Mauricio Faria de Oliveira <mfo@igalia.com>
> ---
>  drivers/hid/hid-playstation.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/hid/hid-playstation.c b/drivers/hid/hid-playstation.c
> index f9dc9378cf777d33aab3cdb87302c6e48f406d2f..b053037a3f0b95b6986c7e53e787e005676ed565 100644
> --- a/drivers/hid/hid-playstation.c
> +++ b/drivers/hid/hid-playstation.c
> @@ -824,6 +824,7 @@ static int ps_get_report(struct hid_device *hdev, u8 report_id, u8 *buf,
>                 u32 report_crc = get_unaligned_le32(&buf[crc_offset]);
>
>                 if (!ps_check_crc32(PS_FEATURE_CRC32_SEED, buf, crc_offset, report_crc)) {
> +                       /* hid_err_ratelimited() not used on driver-initiated requests */
>                         hid_err(hdev, "CRC check failed for reportID=%d\n", report_id);
>                         return -EILSEQ;
>                 }
> @@ -1459,7 +1460,7 @@ static int dualsense_parse_report(struct ps_device *ps_dev, struct hid_report *r
>                 u32 report_crc = get_unaligned_le32(&data[size - 4]);
>
>                 if (!ps_check_crc32(PS_INPUT_CRC32_SEED, data, size - 4, report_crc)) {
> -                       hid_err(hdev, "DualSense input CRC's check failed\n");
> +                       hid_err_ratelimited(hdev, "DualSense input CRC's check failed\n");
>                         return -EILSEQ;
>                 }
>
> @@ -2425,7 +2426,7 @@ static int dualshock4_parse_report(struct ps_device *ps_dev, struct hid_report *
>
>                 /* Last 4 bytes of input report contains CRC. */
>                 if (!ps_check_crc32(PS_INPUT_CRC32_SEED, data, size - 4, report_crc)) {
> -                       hid_err(hdev, "DualShock4 input CRC's check failed\n");
> +                       hid_err_ratelimited(hdev, "DualShock4 input CRC's check failed\n");
>                         return -EILSEQ;
>                 }
>
>
> ---
> base-commit: 83f32c773255720a661aa45d83e36278135f23f5
> change-id: 20261027-hid-playstation-ratelimit-input-crc-error-26f871f1c9d2
>
> Best regards,
> --
> Mauricio Faria de Oliveira <mfo@igalia.com>
>
>