Received-SPF: pass (google.com: domain of linux-wireless+bounces-4136-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
From: Javier Martinez Canillas <javierm@redhat.com>
To: Kalle Valo <kvalo@kernel.org>
Cc: linux-kernel@vger.kernel.org, Nishanth Menon <nm@ti.com>, Breno Leitao
 <leitao@debian.org>, Li Zetao <lizetao1@huawei.com>,
 linux-wireless@vger.kernel.org
Subject: Re: [PATCH] wlcore: sdio: warn only once for
 wl12xx_sdio_raw_{read,write}() failures
In-Reply-To: <8734tdaal1.fsf@kernel.org>
References: <20240227002059.379267-1-javierm@redhat.com>
 <8734tdaal1.fsf@kernel.org>
Date: Wed, 28 Feb 2024 09:24:03 +0100
Message-ID: <877cipngik.fsf@minerva.mail-host-address-is-not-set>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain

Kalle Valo <kvalo@kernel.org> writes:

Hello Kalle,

Thanks for your feedback.

> Javier Martinez Canillas <javierm@redhat.com> writes:
>
>> Report these failures only once, instead of keep logging the warnings for
>> the same condition every time that a SDIO read or write is attempted. This
>> behaviour is spammy and unnecessarily pollutes the kernel log buffer.
>
> Removing error messages is not usually a good idea, it would be much

This patch is not removing error messages though, just limiting to print
only since IMO there is no need to constantly keep printing the same error
message over and over.

> better to fix the root cause.
>

Agreed and I'm trying to figure out the cause. But to do that, I need a
usable serial console and it's barely usable with all the warns and stack
traces printed while I'm trying to type commands.

>> For example, on an AM625 BeaglePlay board where accessing a SDIO WiFi chip
>> fails with an -110 error:
>>
>>   $ dmesg | grep "sdio write\|read failed (-110)" | wc -l
>>   39
>
> -110 is -ETIMEDOUT. Why is it timing out?
>

If I knew it then I wouldn't have to type this patch :) In theory it
should work according to Nishanth (Cc'ed) since I've both the firmware
and the required patches for the bootloader to set some clocks early.

But it's not working for me... I don't know what's missing for me.

>> Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>

[...]

>> -	if (WARN_ON(ret))
>> -		dev_err(child->parent, "sdio read failed (%d)\n", ret);
>> +	if (WARN_ON_ONCE(ret))
>> +		dev_err_once(child->parent, "sdio read failed (%d)\n", ret);
>
> WARN_ON() feels excessive here, maybe remove that entirely? But

Agreed and I'm on board to drop it.

> dev_err_ratelimited() feels more approriate than printing the error just
> once.
>

Works for me. Thanks!

-- 
Best regards,

Javier Martinez Canillas
Core Platforms
Red Hat