Oliver Neukum schrieb:
Am Sonntag, 21. Juni 2009 18:51:44 schrieb Josua Dietze:
I have accumulated some information about these weird hybrid
devices, mostly wireless modems. I am now convinced that ambiguities
and side effects can be handled when switching these devices in the
usb-storage code.
The question is whether this needs to be done in kernel space at all.
We do need this code in kernel space if we want to do reset_resume()
for those devices, but we'd need a lot more infrastructure which we don't
have right now.
Now I'm confused: I thought it was agreed to add switching functions
to the kernel ...
If this is not wanted, why are there such functions for Option,
Huawei and Sierra devices?
If there is a decision to do switching in userspace, fine; but then
let it stay there. As an author of an userspace switching tool I'd
prefer clear standards.
On several occasions I was surprised by kernel updates which
interfered with what I was doing (like a well-meant IGNORE_DEVICE in
"uusual_devs"). My mistake was of course not to be on this list ...
I don't think reset_resume() has any relevance in connection with
switching devices. The ones I'm referring to just "disappear" from
the system and come back as a "new device" when switched, most of
them with a different ID. So usb-storage just needs to send the bulk
message and do nothing because the device is gone in a second
anyway. That's how "option_ms" and "sierra_ms" do it.
"usb_stor_huawei_e220_init" in "initializers" might be different,
but I don't see any special handling there either.
All these devices have been able to be switched in userspace for
quite a while ...
Josua Dietze
--
Man is the only creature on earth enabled to take a
warm meal while flying! Loriot
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
