Greg --
Also, I still plan to backport the USB serial locking from 2.6 to 2.4, if you will accept that.
This late in the 2.4 development cycle, I don't really know. If it fixes real bugs, probably. Otherwise I doubt it.
As we have talked about before, the semaphore in usbserial write leads to an oops if the semaphore is held and write is called on interrupt time. Remember, write can be called from an interrupt under certain circumstances and you can't use semaphores then.
This is easy to reproduce in some of the drivers I have tested and other have complained about this bug on linux-usb-devel--use a loopback plug on a port, set stty ocrnl, cat a big file to the port, oops scheduling in interrupt. I think setting tty->low_latency makes this bug more likely. We give customers who have trouble with this oops a hacked usbserial with the semaphore removed.
The semaphore is ok in open, close, and any other function that is not called in interrupts. I will have to check just which usbserial functions that currently use the semaphore can be called in interrupts. The USB serial drivers have to do their own spin locking if needed in these cases.
-- Al
------------------------------------------------------- This SF.Net email is sponsored by: SourceForge.net Broadband Sign-up now for SourceForge Broadband and get the fastest 6.0/768 connection for only $19.95/mo for the first 3 months! http://ads.osdn.com/?ad_id=2562&alloc_id=6184&op=click _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
