On Tue, Jul 10, 2018 at 01:23:35PM +0000, Ruhl, Michael J wrote: > >-----Original Message----- > >From: Leon Romanovsky [mailto:leon@xxxxxxxxxx] > >Sent: Tuesday, July 10, 2018 6:32 AM > >To: Doug Ledford <dledford@xxxxxxxxxx>; Jason Gunthorpe > ><jgg@xxxxxxxxxxxx> > >Cc: Leon Romanovsky <leonro@xxxxxxxxxxxx>; RDMA mailing list <linux- > >rdma@xxxxxxxxxxxxxxx>; Huy Nguyen <huyn@xxxxxxxxxxxx>; Ruhl, Michael J > ><michael.j.ruhl@xxxxxxxxx> > >Subject: [PATCH rdma-next v3 2/2] RDMA/umem: Refactor exit paths in > >ib_umem_get > > > >From: Leon Romanovsky <leonro@xxxxxxxxxxxx> > > > >Simplify exit paths in ib_umem_get. > > > >Signed-off-by: Leon Romanovsky <leonro@xxxxxxxxxxxx> > >--- > > drivers/infiniband/core/umem.c | 44 +++++++++++++++++++------------------ > >----- > > 1 file changed, 20 insertions(+), 24 deletions(-) > > <...> > >- } > >- > > if (vma_list) > > free_page((unsigned long) vma_list); > > free_page((unsigned long) page_list); > > It appears that free_page() will protect against a 0 address. > > With that in mind, is the umem_kfree: target really necessary? IMHO, it is easier for reader to see symmetric exit points. > > Either way: > > Reviewed-By: Michael J. Ruhl <michael.j.ruhl@xxxxxxxxx> > > Mike
Attachment:
signature.asc
Description: PGP signature
