On Wed, 29 Dec 2004, Pete Zaitcev wrote:
[] For example a multithreaded application which at first make a ioctl which uses exclusive_access and second tread will do some polling without exclusive_access. But in case the second thread try first a non existing ioctl, it will get into lock and cant send signal to first tread to interrupt his ioctl... It is not a good example, but I hope you understand what i mean.
This sounds like a poor example to me. If an application wants to test existing capabilities, it can do that before forking. We don't have to accommodate this sort of behaviour, so Sergey's patch looks good enough.
Well, Sergey's patch sets mutex also for nonexisting ioctl, which isn't *right*.
Meelis, does it solve your modem_run problem? Yes or no?
I haven't tested Sergey's patch, but my patch is reported to work also with eciadsl:
http://www.mail-archive.com/speedtouch%40ml.free.fr/msg06848.html
-- Kaupo Arulo
------------------------------------------------------- The SF.Net email is sponsored by: Beat the post-holiday blues Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek. It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
