[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20251007162111.GA3604844@ziepe.ca>
Date: Tue, 7 Oct 2025 13:21:11 -0300
From: Jason Gunthorpe <jgg@...pe.ca>
To: Gerd Bayer <gbayer@...ux.ibm.com>
Cc: Tariq Toukan <tariqt@...dia.com>, Saeed Mahameed <saeedm@...dia.com>,
Leon Romanovsky <leon@...nel.org>, Shay Drori <shayd@...dia.com>,
Mark Bloch <mbloch@...dia.com>, Andrew Lunn <andrew+netdev@...n.ch>,
"David S . Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
Alex Vesker <valex@...lanox.com>,
Feras Daoud <ferasda@...lanox.com>, netdev@...r.kernel.org,
linux-rdma@...r.kernel.org, linux-kernel@...r.kernel.org,
Niklas Schnelle <schnelle@...ux.ibm.com>,
linux-s390@...r.kernel.org
Subject: Re: [PATCH net v2] net/mlx5: Avoid deadlock between PCI error
recovery and health reporter
On Tue, Oct 07, 2025 at 04:48:26PM +0200, Gerd Bayer wrote:
> - task: kmcheck
> mlx5_unload_one() tries to acquire devlink lock while the PCI error
> recovery code has set pdev->block_cfg_access by way of
> pci_cfg_access_lock()
This seems wrong, arch code shouldn't invoke the driver's error
handler while hodling pci_dev_lock().
Or at least if we do want to do this the locking should be documented
and some lockdep map should be added to pci_cfg_access_lock() and the
normal AER path..
Jason
Powered by blists - more mailing lists