linux-kernel - Re: [PATCH] PCI: pciehp: Do not turn off slot if presence comes up after link

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 for Android: free password hash cracker in your pocket

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

Date:   Tue, 12 Feb 2019 11:37:00 +0100
From:   Lukas Wunner <lukas@...ner.de>
To:     Alex_Gagniuc@...lteam.com
Cc:     mr.nuke.me@...il.com, bhelgaas@...gle.com, Austin.Bolen@...l.com,
        keith.busch@...el.com, Shyam.Iyer@...l.com, gustavo@...eddedor.com,
        linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] PCI: pciehp: Do not turn off slot if presence comes up
 after link

On Tue, Feb 12, 2019 at 09:30:31AM +0100, Lukas Wunner wrote:
> Instead of modifying the logic in pciehp_handle_presence_or_link_change(),
> you could amend pcie_wait_for_link() to poll PDS until it's set, in
> addition to DLLLA.  The rationale would be that although the link is up,
> the hot-added device cannot really be considered accessible until PDS
> is also set.  Unfortunately we cannot do this for all devices because
> (as I've said before), some broken devices hardwire PDS to zero.  But
> it should be safe to constrain it to those which can disable in-band
> presence.

An alternative approach would of course be to only enable one of PDC or
DLLSC events if in-band presence is disabled.  In the case of DLLSC,
you'd also need to modify pciehp_reset_slot() in addition to
pcie_enable_notification().

Thanks,

Lukas