Re: cam_da too noisy about SYNCHRONIZE CACHE

From: Warner Losh <imp_at_bsdimp.com>
Date: Wed, 28 Jan 2026 21:52:56 UTC 
On Tue, Jan 27, 2026 at 1:15 AM Poul-Henning Kamp <phk@phk.freebsd.dk>
wrote:

> I have a Rpi3B which runs 15.0 on a Seagate USB-SATA gadget.
>
> Every few hours usually around XX:X0:30 it spits out:
>
>         Jan 27 00:40:30 rpi3b kernel: (da0:umass-sim0:0:0:0): MODE SENSE
> for CACHE page command failed.
>         Jan 27 00:40:30 rpi3b kernel: (da0:umass-sim0:0:0:0): Mode page 8
> missing, disabling SYNCHRONIZE CACHE
>         Jan 27 00:40:30 rpi3b kernel: (da0:umass-sim0:0:0:0): Devices
> already quirked for NO_SYNC_CACHE, maybe remove quirk table
>
> The relevant code in cam_da is:
>
>                         if (mark_bad) {
>         bad:
>                                 xpt_print(done_ccb->ccb_h.path,
>                                     "Mode page 8 missing, disabling
> SYNCHRONIZE CACHE\n");
>                                 if (softc->quirks & DA_Q_NO_SYNC_CACHE)
>                                         xpt_print(done_ccb->ccb_h.path,
>             "Devices already quirked for NO_SYNC_CACHE, maybe remove quirk
> table\n");
>                                 softc->quirks |= DA_Q_NO_SYNC_CACHE;
>                                 softc->disk->d_flags &=
> ~DISKFLAG_CANFLUSHCACHE;
>                         }
>
> I can understand emitting the message the first time after a reboot, but
> that program logic makes absolutely no sense to me ?
>

That should have all been behind bootverbose. My bad for letting it linger
for so long. I added this verbosity when I was worried about people
complaining that I broke their drive. I fixed it too well: They've only
complained about the messages...

Warner