On Thu, May 18, 2017 at 03:29:45PM +0200, Johannes Thumshirn wrote: > On 05/18/2017 03:19 PM, Christoph Hellwig wrote: > > All SG_IO test should also apply to block device nodes that support > > the ioctl.. > > > > But these are not necessarily SG_IO tests, are they? > > The test included is doesn't hit the SG_IO path in the sg driver, but > the sg_read path. > > Of cause we can make a generic sg_io tool but IMHO it's quite convenient > to just throw in the reproducers we get. > > And as it exercises the /dev/sg character device I decided to make a new > group. > > Johannes
Looking at this some more, it seems like the syzkaller reproducer always bangs on /dev/sg0. How hard would it be to adapt it to run on the sg device for every test device instead?
