There are cases where the EFI runtime services may end up in a funny
state, e.g., due to a crash in the variable services, and this affects
other EFI runtime services as well.

That means that, even though GetTime() should not return an error, there
are cases where it might, and there is no point in logging such an
occurrence multiple times.

This works around an issue where user space -apparently- keeps hitting
on /dev/rtc if it fails to read the h/w clock, resulting in a tsunami of
log spam and a non-responsive system as a result.

Cc: Pierre Gondois <pierre.gondois@arm.com>
Cc: Alexandru Elisei <alexandru.elisei@arm.com>
Link: https://lore.kernel.org/all/Y2o1hdZK9GGDVJsS@monolith.localdoman/
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
 drivers/rtc/rtc-efi.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/rtc/rtc-efi.c b/drivers/rtc/rtc-efi.c
index e991cccdb6e9cee4..174959f783b8309b 100644
--- a/drivers/rtc/rtc-efi.c
+++ b/drivers/rtc/rtc-efi.c
@@ -164,7 +164,7 @@ static int efi_read_time(struct device *dev, struct rtc_time *tm)
 
 	if (status != EFI_SUCCESS) {
 		/* should never happen */
-		dev_err(dev, "can't read time\n");
+		dev_err_once(dev, "can't read time\n");
 		return -EINVAL;
 	}
 
-- 
2.39.1

On Fri, 17 Feb 2023 15:23:38 +0100, Ard Biesheuvel wrote:
> There are cases where the EFI runtime services may end up in a funny
> state, e.g., due to a crash in the variable services, and this affects
> other EFI runtime services as well.
> 
> That means that, even though GetTime() should not return an error, there
> are cases where it might, and there is no point in logging such an
> occurrence multiple times.
> 
> [...]

Applied, thanks!

[1/1] rtc: efi: Avoid spamming the log on RTC read failure
      commit: 668a2abf91143caa226b3ccd0bd4d79ea85935a6

Best regards,

-- 
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com