Am 28.06.2016 um 22:09 schrieb Alex Williamson: > On Tue, 28 Jun 2016 20:40:37 +0200 > Heiner Kallweit <[email protected]> wrote: > >> The init function ensures that iommu_group_kset can't be NULL. >> Therefore this check isn't needed. > > Have you seen your previous patch? Wasn't this claim that the init > function ensures iommu_group_kset more true when iommu_init() called > BUG_ON(!iommu_group_kset) than it is when we can get by with just > returning -ENOMEM? Thanks, > Uh, you're right. I was in "module mode" where initcall retvals are assessed. Here the initcall retval is ignored. If we leave the BUG_ON in however then the check is actually unneeded as the system is halted after the BUG_ON.
By the way: This check is in iommu_group_get_by_id() and this function doesn't seem to have any user in the kernel. It was added about three years ago. Is the original intention of this function obsolete meanwhile? Thanks, Heiner > Alex > >> Signed-off-by: Heiner Kallweit <[email protected]> >> --- >> drivers/iommu/iommu.c | 3 --- >> 1 file changed, 3 deletions(-) >> >> diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c >> index b36ec9c..b36e24d 100644 >> --- a/drivers/iommu/iommu.c >> +++ b/drivers/iommu/iommu.c >> @@ -235,9 +235,6 @@ struct iommu_group *iommu_group_get_by_id(int id) >> struct iommu_group *group; >> const char *name; >> >> - if (!iommu_group_kset) >> - return NULL; >> - >> name = kasprintf(GFP_KERNEL, "%d", id); >> if (!name) >> return NULL; > > _______________________________________________ iommu mailing list [email protected] https://lists.linuxfoundation.org/mailman/listinfo/iommu
