On Mon, Jul 23, 2018 at 04:09:29PM -0600, Jason Gunthorpe wrote: > On Wed, Jul 11, 2018 at 04:20:43PM -0600, Jason Gunthorpe wrote: > > From: Jason Gunthorpe <jgg@xxxxxxxxxxxx> > > > > This is just a little series to make the kabi specs language cleaner and > > clearer. > > > > We have several places that use flags, and they should all follow the same > > basic rules for consistency: > > - Must be validated, unsupported bits must cause EINVAL > > - Must be passed inline as a u32 or u64, ie cannot get EFAULT > > - Are always optional, missing flags means flag == 0 > > - Are always be stored in a u32 or u64 after fetching from userspace > > - Are linked to an enum defining the bits located in a uapi header > > > > The new specs attribute and the uverbs_get_flagsXX helper enforce most of > > these rules. > > > > Jason Gunthorpe (2): > > IB/uverbs: Move ib_access_flags and ib_read_counters_flags to uapi > > IB/uverbs: Add UVERBS_ATTR_FLAGS_IN to the specs language > > Applied to for-next 0-day says the second patch has a build failure if CONFIG_INFINIBAND_USER_ACCESS is disabled, I'm going to drop the second patch and re-spin it for now. Jason -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
