linux-kernel - Re: [PATCH v3] nvme/pci: Add quick suspend quirk for Sc7280 Platform

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]

Date:   Wed, 16 Feb 2022 09:12:50 -0800
From:   Christoph Hellwig <hch@...radead.org>
To:     "Nitin Rawat (QUIC)" <quic_nitirawa@...cinc.com>
Cc:     Christoph Hellwig <hch@...radead.org>,
        Vidya Sagar <vidyas@...dia.com>,
        Keith Busch <kbusch@...nel.org>, Jens Axboe <axboe@...com>,
        Sagi Grimberg <sagi@...mberg.me>,
        "linux-nvme@...ts.infradead.org" <linux-nvme@...ts.infradead.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "Sajida Bhanu (Temp) (QUIC)" <quic_c_sbhanu@...cinc.com>
Subject: Re: [PATCH v3] nvme/pci: Add quick suspend quirk for Sc7280 Platform

On Wed, Feb 16, 2022 at 04:07:05PM +0000, Nitin Rawat (QUIC) wrote:
> Hi Christoph/Keith/Rafael
> 
> Since we are giving control to PCIe (NVMe power rails control), it can vary from platform to platform.
> More over, PCIe driver doesn't control these power rails directly from PCIe driver, they tie nvme supply with one of the pcie supply and control them together. 
> So i think it would be better to either have quirk based on platform or always setting simple suspend and platform which needs full suspend can update it through some means. 
> 
> Based on below link, Looks like this can be across platform ...vidya also mention similar concern for tegra platform.  

Here is the thing:  nothing here is really NVMe specific.  We do have
a bunch of drivers including nvme that would love to not do a full
firmware shutdown on a suspend.  For storage devices this is especially
important as each shutdown reduces media life time.  Until very recently
this has been perfectly fine, but now various platforms show up that
want to completely disable power to PCIe slots on a system suspend.

The drivers need information from the PCI / PM core when a ->suspend
call requires the device to got into D3 and when not, and we can't just
do that with quirks in the various drivers.