netdev - RE: [EXT] Prestera driver fail to probe twice

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: 
 <BN9PR18MB4251BC09AAD68BDA10AD4C9FDB442@BN9PR18MB4251.namprd18.prod.outlook.com>
Date: Thu, 8 Feb 2024 07:36:54 +0000
From: Elad Nachman <enachman@...vell.com>
To: Jakub Kicinski <kuba@...nel.org>
CC: Köry Maincent <kory.maincent@...tlin.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        Taras Chornyi
	<taras.chornyi@...ision.eu>,
        Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
        Miquel Raynal <miquel.raynal@...tlin.com>
Subject: RE: [EXT] Prestera driver fail to probe twice

Once again, existing deployed firmware does not check the enum below , so this does not ensure backward compatibility.

-----Original Message-----
From: Jakub Kicinski <kuba@...nel.org> 
Sent: Wednesday, February 7, 2024 8:32 PM
To: Elad Nachman <enachman@...vell.com>
Cc: Köry Maincent <kory.maincent@...tlin.com>; netdev@...r.kernel.org; Taras Chornyi <taras.chornyi@...ision.eu>; Thomas Petazzoni <thomas.petazzoni@...tlin.com>; Miquel Raynal <miquel.raynal@...tlin.com>
Subject: Re: [EXT] Prestera driver fail to probe twice

On Wed, 7 Feb 2024 15:06:34 +0000 Elad Nachman wrote:
> MAGIC value is cleared after FW is downloaded to RAM, just before it is executed.
> So checking the MAGIC value to determine if the firmware is running is useless.
> 
> You will get the MAGIC value read correctly only when firmware is 
> during the download process, which does not help you implementing the 
> logic you want.

I'm not aware of other networking drivers which cannot be reloaded / rebound. I think it's part of base expectations for upstream drivers.
Could you work on fixing this?

IIRC we have enum devlink_param_reset_dev_on_drv_probe_value to control whether device is always reset on driver load. You can make resetting an opt-in if you're concerned about backward compatibility.