On Thu, Jan 2, 2020 at 11:34 AM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote: > > On Thu, Jan 02, 2020 at 11:24:09AM -0500, Marcelo Diop-Gonzalez wrote: > > Hi, I noticed couple minor bugs while looking around in > > drivers/staging/vc04_services/*: > > > > 1. When opening "/dev/vchiq", vchiq_debugfs_add_instance() adds a debugfs > > directory named after the tgid of the process who opened the file. But > > there will be collisions if you try opening it multiple times from the same > > process. > > Odd, but shouldn't cause any problems as no code paths should care if > debugfs works or not, it's only for debugging :) > > > 2. In vchiq_dump_platform_instances(), called from vchiq_read(), each > > instance's ->mark member is used to track whether it's been printed > > already, but two concurrent readers could stomp on each other (causing one > > of them to see fewer instances than actually exist), since a field in the > > vchiq_instance struct itself is being modified with no > > locking/synchronization > > > > I didn't want to send any patches out since I'm not sure if/how people > > would want to fix these, but thought I'd point them out. > > Patches are always best, please feel free to submit to help resolve > these. > sounds good! > thanks, > > greg k-h _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
