Re: [PATCH v2 0/2] ATA port deferred qc fixes

[Niklas Cassel <cassel@kernel.org>]

On Thu, Apr 02, 2026 at 02:22:49AM -0700, Tommy Kelly wrote:
> The device/driver/subsystem tree looks like rk3568-dwc-ahci -> scsi -> sda.
> The platform does not (yet) support FIS-Based Switching (FBS), so it is
> using Command-Based Switching (CBS). FBS support may land in Linux 7.1.

I still do not have access to a PMP.
Damien said he might give me access to one.

FWIW, since you are using rk368, I saw that:

commit 5918bf2a17f4689abfb94d341c5240088f8bd16e
Author: Heiko Stuebner <heiko@sntech.de>
Date:   Sun Feb 1 20:18:01 2026 +0100

    arm64: dts: rockchip: Add port subnodes to RK356x SATA controllers


Has now landed in Linus's tree, which adds PMP FBS support for this platform.


What we know:
1) There is a design flaw in the current deferral code which will make
   PMP FBS defer commands to all drives when sending a non-NCQ command.
   (Basically giving the FBS the same performance as CBS).

2) There is a bug that you found that seems to cause indefinite requeues on
   PMP CBS. This manifests as:
systemd[1]: dev-sdc.device: Job dev-sdc.device/start timed out.
systemd[1]: Timed out waiting for device dev-sdc.device - /dev/sdc.

   We don't know if this bug also affects PMP FBS or not.
   Out of curiousity, do you still see this bug if running with the above
   patch cherry-picked or not?


Kind regards,
Niklas