RE: [EXTERNAL] Re: [PATCH v2 2/5] net: marvell: prestera: enlarge fw restart time
From: Elad Nachman
Date: Thu Mar 21 2024 - 13:24:56 EST
> -----Original Message-----
> From: Andrew Lunn <andrew@xxxxxxx>
> Sent: Thursday, March 21, 2024 2:10 AM
> To: Elad Nachman <enachman@xxxxxxxxxxx>
> Cc: Taras Chornyi <taras.chornyi@xxxxxxxxxxx>; davem@xxxxxxxxxxxxx;
> edumazet@xxxxxxxxxx; kuba@xxxxxxxxxx; pabeni@xxxxxxxxxx;
> kory.maincent@xxxxxxxxxxx; thomas.petazzoni@xxxxxxxxxxx;
> miquel.raynal@xxxxxxxxxxx; przemyslaw.kitszel@xxxxxxxxx;
> dkirjanov@xxxxxxx; netdev@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> Subject: [EXTERNAL] Re: [PATCH v2 2/5] net: marvell: prestera: enlarge fw
> restart time
>
> Prioritize security for external emails: Confirm sender and content safety
> before clicking links or opening attachments
>
> ----------------------------------------------------------------------
> On Wed, Mar 20, 2024 at 07:20:05PM +0200, Elad Nachman wrote:
> > From: Elad Nachman <enachman@xxxxxxxxxxx>
> >
> > Increase firmware restart timeout, as current timeout value of 5
> > seconds was too small, in actual life it can take up to 30 seconds for
> > firmware to shutdown and for the firmware loader to shift to the ready
> > to receive firmware code state.
>
> So this means the probe can be waiting 30 seconds?
>
> This seems like a really good reason not to reset the hardware during -
> EPROBE_DEFER.
>
> Andrew
Unfortunately that's how the firmware loader on the firmware cpu state machine works.
The so called HW reset will force exit of the current firmware and will move the state machine
To the state that it can receive the next firmware.
Unfortunately, there is no ABI interface to verify which firmware is already loaded, and then supporting
Warm boot reading of the values back to the kernel.
Since many of these firmware binaries are secure-boot protected, upgrading is very tricky.
Elad.