Skip to contentSkip to conversation

ti_usb_3410_5052 breakage in 2.6.24-rc1

linux-usb-devel Alan Stern 1 participant 1 message Nov 2, 2007 1 message not in the archive

lore
Share & tools
lore

Alan Stern

author
#1 lore raw
In reply to a message not in the archive
Quoted 13 lines · Steven King
Quoted 3 lines

But you didn't post the log messages, so I haven't seen them. In addition, that "wrong number of endpoints" message occurs long after the underlying problem; it's probably not connected.

Okay, attached is full .config, dmesg, /proc/bus/usb/devices and lsusb for 2.6.23.1 (aka working dongle), 2.6.24-rc1 (aka non-working dongle), 2.6.24-rc1-r (2.6.24-rc1 with commit 063a2da8f01806906f7d7b1a1424b9afddebc443 backed out, aka working dongle 2) and finally 2.6.24-rc1-X (2.6.24-rc1 with my patch to ti_usb_3410_5052, aka working dongle 3). Normally I have a udev rule to automagically select configuration 2, for these tests I disabled that. The {dmesg, proc.bus.usb.devices, lsusb}-2.6.xxx-1 is after I did 'echo 2 > /sys/.../bConfigurationValue'.

Quoted 2 lines

Clean fresh build for each of the kernels, the dongle was plugged after the system was fully up and X had started.

Okay, this explains everything. In 2.6.23 (or 2.6.24-rc1 with the commit backed out), usb-serial binds to the "wrong" interface and performs a firmware download. Up to that point the device has just the one configuration, but with the new firmware it has two. Then you install the second config and the driver can bind successfully.

With 2.6.24-rc1, the driver never gets a chance to update the firmware because the invalid endpoint matching prevents it.

Oliver, what do you think should be done to solve this?

Alan Stern