Hello, the attached patch adds support for the Genesys 7-in-1 Card Reader/Writer. This resolves situations where the generic usb-storage-scsi-bulk device support is disabled due to selection of the ub module.
MAybe that patch is really really unneccssary, but read on!
The "Low Performance USB Block driver" look still eperimental
to me in some respect, acording to its sources:
ub.c:10: * TODO (sorted by decreasing priority)
<list with 15 items>
ub.c:40: #define UB_MAJOR 125 /* Stolen from Experimental range for a
week - XXX */
ub.c:330: * Find me a way to identify "next free minor" for add_disk(),
ub.c:331: * and the array disappears the next day. However, the number of
ub.c:332: * hosts has something to do with the naming and /proc/partitions.
ub.c:333: * This has to be thought out in detail before changing.
ub.c:334: * If UB_MAX_HOST was 1000, we'd use a bitmap. Or a better data
structure.
ub.c:1359: * XXX sd.c sets capacity to zero in such case. However, it
doesn't
ub.c:1360: * work for us. In case of zero capacity, block layer refuses to
ub.c:1361: * have the /dev/uba opened (why?) Set capacity to some random
value.
other than that i feel a bit irritated due to the ub device node naming.
Documentation/devices.txt says /dev/ubdX (user-mode block device)
is major=98 and for user mode linux while it never mentions anything
that looks like /dev/uba or something, whatever this "a" means.
I really wonder how such a yet unfinished module
ever made it to merge into a stable kernel series.
At least when i do have the choice between tow modules
then i want to be able to use both alternating.
In other words i am sure you will suggest me a
"mknod /dev/uba b 125 0" or alikes and all should work,
but having such strong side effects like turning off
features of other moules plus adding new device nodes
is really something worth a node in the config.
current:
CONFIG_BLK_DEV_UB:
x This driver supports certain USB attached storage devices
x such as flash keys. If unsure, say N.
Lets consider those lack of documentation a bug.
I would suggest that the ub driver has to report any
registering and unregistering, and even the failure
for such a device node in the kernel log.
That will trigger anybodys attention that looks for
messages of usb-storage reporting the new scsi deivce.
Of course those ub change will have noticeable influx
on any module autoloader setup thus breaking the
ability to switch between old and new 2.6 kernels
without touching any other files of the machine setup.
If i was customed that usb-storage was autoloaded
when mounting my sd-card then it will now break,
and if i have changed to ub devices, than accessing
/dev/ub... will break when loaded an odler kernel.
-Alex.
--
Superg�nstige DSL-Tarife + WLAN-Router f�r 0,- EUR*
Jetzt zu GMX wechseln und sparen http://www.gmx.net/de/go/dsl
linux-2.6.9-rc2-usb7i1cardreader.diff
Description: Binary data
