HEADSUP: USB da(4) quirks disabled for 4.9 and 5.2
Yoshihiko Sarumaru
mistral at imasy.or.jp
Sun Oct 5 08:36:14 PDT 2003
Hello,
This is a more than 1 month old issue but ...,
On 22 Aug, nate wrote:
> The problem is that a lot of these quirks were added with no documentation
> and a cut/paste of both NO_6_BYTE and NO_SYNCHRONIZE_CACHE quirks even
> though only one of them was needed. Now that no USB devices should
> receive a 6 byte command, that only leaves the NO_SYNCHRONIZE_CACHE
> quirks. I am temporarily disabling them to flush out which ones are
> REALLY needed. They have been disabled since 2003/07/29 and 2003/08/07
> for -current and -stable, respectively. I have had 0 messages saying that
> anything was broken by this.
I agree for this point of view.
On 22 Aug, nate wrote on the other mail:
> > It appears to need it;
> >
> > umass0: Unsupported UFI command 0x35
> > (da0:umass-sim0:0:0:0): Synchronize cache failed, status == 0x6,
> > scsi status == 0x0
>
> Unless that causes an actual problem (i.e. crash, data loss), it is not
> necessary to have a quirk for it. There are two kinds of "unsupported"
> commands: those that cause the device to hang or crash and those that
> successfully reject commands they don't want with the proper return value.
> Quirks will only be added for the former.
I don't agree for this because this quirk is turned out to have
a reason to exist and can be documented.
My internal memory stick slot (MSC-U01) always complains above
three-line dmesg whenever performing umount, and my console get
dirty. That's not good for our health (it doesn't happen until now).
So I want to quirks for "Sony", "MS*", "*" will be kept in
future releases. I feel this quirk is needed even thougn any
data corruption will be occured.
--
Yoshihiko Sarumaru
mail: mistral at imasy.or.jp web: http://www.imasy.or.jp/~mistral/
More information about the freebsd-stable
mailing list