Re: [PATCH net] bgmac: fix *initial* chip reset to support BCM5358 - Linux regression tracking (Thorsten Leemhuis)

From: "Linux regression tracking (Thorsten Leemhuis)"  <regressions@leemhuis.info>
To: "Florian Fainelli" <f.fainelli@gmail.com>,
	"Rafał Miłecki" <rafal@milecki.pl>,
	"Ricardo Cañuelo" <ricardo.canuelo@collabora.com>,
	"Rafał Miłecki" <zajec5@gmail.com>
Cc: "David S . Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	Arnd Bergmann <arnd@arndb.de>,
	netdev@vger.kernel.org, bcm-kernel-feedback-list@broadcom.com,
	Jon Mason <jdmason@kudzu.us>,
	Linux kernel regressions list <regressions@lists.linux.dev>
Subject: Re: [PATCH net] bgmac: fix *initial* chip reset to support BCM5358
Date: Fri, 14 Apr 2023 16:04:10 +0200	[thread overview]
Message-ID: <d6990d00-6fd5-cd89-755d-d7f566c574fa@leemhuis.info> (raw)
In-Reply-To: <b7b11a57-9512-cda9-1b15-5dd5aa12f162@gmail.com>

On 05.04.23 14:42, Florian Fainelli wrote:
> On 4/4/2023 6:52 AM, Rafał Miłecki wrote:

>> On 4.04.2023 15:46, Ricardo Cañuelo wrote:
>>> On mar 07-02-2023 23:53:27, Rafał Miłecki wrote:
>>>> While bringing hardware up we should perform a full reset including the
>>>> switch bit (BGMAC_BCMA_IOCTL_SW_RESET aka SICF_SWRST). It's what
>>>> specification says and what reference driver does.
>>>>
>>>> This seems to be critical for the BCM5358. Without this hardware
>>>> doesn't
>>>> get initialized properly and doesn't seem to transmit or receive any
>>>> packets.
>>>>
>>>> Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
>>>
>>> KernelCI found this patch causes a regression in the
>>> bootrr.deferred-probe-empty test [1] on sun8i-h3-libretech-all-h3-cc
>>> [2], see the bisection report for more details [3]
>>>
>>> Does it make sense to you?
>>
>> It doesn't seem to make any sense. I guess that on your platform
>> /sys/kernel/debug/devices_deferred
>> is not empty anymore?
>>
>> Does your platform use Broadcom Ethernet controller at all?
> 
> I do not believe it does, however according to the log, the driver is
> enabled:
> 
> <6>[    1.819466] bgmac_bcma: Broadcom 47xx GBit MAC driver loaded
> 
> but it should not be probing any device since you don't have any
> internal BCMA bus to match gigabit devices with. Later in the log we see:
> 
> 1c22c00.codec    sun4i-codec: Failed to register our card
> 
> and most likely as you already wrote the deferred device list might not
> be empty.

What happened to this? It seems there wasn't any progress since above
mail week. But well, seems to be a odd issue anyway (is that one of
those issues that CI systems find, but don't cause practical issues in
the field?). Hence: can somebody with more knowledge about this please
tell if it this is something I can likely drop from the list of tacked
regressions?

Ciao, Thorsten (wearing his 'the Linux kernel's regression tracker' hat)
--
Everything you wanna know about Linux kernel regression tracking:
https://linux-regtracking.leemhuis.info/about/#tldr
If I did something stupid, please tell me, as explained on that page.

#regzbot poke