Re: [PATCH 5.10 00/11] 5.10.123-rc1 review

[Jason A. Donenfeld]

Hi Jon,

On Thu, Jun 16, 2022 at 09:48:37AM +0100, Jon Hunter wrote:
> No new regressions for Tegra. I am seeing the following kernel warning 
> that is causing a boot test to fail, but this has been happening for a 
> few releases now (I would have reported it earlier but we have been 
> having some infrastructure issues) ...
> 
>   WARNING KERN urandom_read_iter: 82 callbacks suppressed
> 
> This appears to be introduced by commit "random: convert to using 
> fops->read_iter()" [0]. Interestingly, I am not seeing this in the 
> mainline as far as I can tell and so I am not sure if there is something 
> else that is missing?
> 
> 
> Test results for stable-v5.10:
>      10 builds:	10 pass, 0 fail
>      28 boots:	28 pass, 0 fail
>      75 tests:	74 pass, 1 fail
> 
> Linux version:	5.10.123-rc1-gf67ea0f67087
> Boards tested:	tegra124-jetson-tk1, tegra186-p2771-0000,
>                  tegra194-p2972-0000, tegra194-p3509-0000+p3668-0000,
>                  tegra20-ventana, tegra210-p2371-2180,
>                  tegra210-p3450-0000, tegra30-cardhu-a04
> 
> Test failures:	tegra194-p2972-0000: boot.py
> 
> Tested-by: Jon Hunter <jonathanh@xxxxxxxxxx>
> 
> Jon
> 
> [0] 
> https://lore.kernel.org/lkml/20220527084907.568432116@xxxxxxxxxxxxxxxxxxx/

Please CC me on RNG issues.

I'm surprised that this message results in a failure. It's not a
WARN_ON() or a BUG() that's being triggered here. This is just the
simple `pr_warn("%s: %d callbacks suppressed\n")` in lib/ratelimit.c,
which really shouldn't be causing your CI to fail. Sounds like your
harness could use some adjusting.

Nonetheless, you have found a 4 year old bug in the urandom warning
accounting that was recently made more easily triggerable by a newer
commit, though not the one you mentioned. I'll fix this up and keep you
CC'd on the patch, which should make it into stable as well.

Jason