[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20190429185611.121751-1-helgaas@kernel.org>
Date: Mon, 29 Apr 2019 13:56:10 -0500
From: Bjorn Helgaas <helgaas@...nel.org>
To: Alexandru Gagniuc <mr.nuke.me@...il.com>,
Lukas Wunner <lukas@...ner.de>,
Alex Williamson <alex.williamson@...hat.com>,
Austin Bolen <austin_bolen@...l.com>,
Alexandru Gagniuc <alex_gagniuc@...lteam.com>,
Keith Busch <keith.busch@...el.com>,
Shyam Iyer <Shyam_Iyer@...l.com>,
Sinan Kaya <okaya@...nel.org>,
Linus Torvalds <torvalds@...ux-foundation.org>,
linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: [PATCH] Revert "PCI/LINK: Report degraded links via link bandwidth
Alex W. reported a case where the new link bandwidth notification logging,
e.g.:
vfio-pci 0000:07:00.0: 32.000 Gb/s available PCIe bandwidth, limited by 2.5 GT/s x16 link at 0000:00:02.0 (capable of 64.000 Gb/s with 5 GT/s x16 link)
is somewhat surprising. The message is great if link bandwidth was reduced
because of a hardware problem, but in this case it's caused by a GPU doing
link state management to conserve power, so the message is to be expected
and is likely to confuse the user.
We haven't figured out a good way to handle both the "hardware failure" and
the "active link management" cases yet, so revert this for now.
Bjorn
Powered by blists - more mailing lists