[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240807072908.1da91994@kernel.org>
Date: Wed, 7 Aug 2024 07:29:08 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Yunsheng Lin <linyunsheng@...wei.com>
Cc: <netdev@...r.kernel.org>, <davem@...emloft.net>, <edumazet@...gle.com>,
<pabeni@...hat.com>, <ilias.apalodimas@...aro.org>, Jesper Dangaard Brouer
<hawk@...nel.org>, Alexander Duyck <alexander.duyck@...il.com>, Yonglong
Liu <liuyonglong@...wei.com>
Subject: Re: [RFC net] net: make page pool stall netdev unregistration to
avoid IOMMU crashes
On Wed, 7 Aug 2024 19:00:35 +0800 Yunsheng Lin wrote:
> > Note that page pool pages may last forever, we have seen it happen
> > e.g. when application leaks a socket and page is stuck in its rcv queue.
>
> We saw some page_pool pages might last forever too, but were not sure
> if it was the same reason as above? Are there some cmds/ways to debug
> if a application leaks a socket and page is stuck in its rcv queue?
I used drgn to scan all sockets to find the page.
> > Hopefully this is fine in this particular case, as we will only stall
> > unregistering of devices which want the page pool to manage the DMA
> > mapping for them, i.e. HW backed netdevs. And obviously keeping
> > the netdev around is preferable to a crash.
Powered by blists - more mailing lists