On 2015/10/13, 16:19, "Aya Mahfouz" <mahfouz.saif.elyazal@xxxxxxxxx> wrote: >container_of0 can be replaced by the Linux kernel macro container_of. >The only difference is that container_of0 tries to evade type casting >when the pointer is erroneous or null. All uses of container_of0 have >been replaced by container_of and then the macro was removed in the last >patch. Thank you for your patch. However, it should be noted that this blanket replacement of container_of0() isn't necessarily safe. If the caller isn't explicitly checking for IS_ERR(ptr) or ptr == NULL then the container_of() macro can transform the err/NULL pointer into another pointer that is NOT err/NULL and cause the kernel to oops when the bad pointer is used, even if there are later IS_ERR() or NULL checks. I'm not against removing container_of0() completely, but not without a clear review of the callers to determine whether err/NULL pointers are not actually possible in this callpath, or explicit checks in the callers for err/NULL pointers before container_of() is used. Did you do that? Cheers, Andreas >Aya Mahfouz (7): > staging: lustre: lcommon_cl.c: replace container_of0 by container_of > staging: lustre: llite_nfs.c: replace container_of0 by container_of > staging: lustre: lu_object.c: replace container_of0 by container_of > staging: lustre: echo_client.c: replace container_of0 by container_of > staging: lustre: osc_io.c: replace container_of0 by container_of > staging: lustre: osc_object.c: replace container_of0 by container_of > staging: lustre: remove container_of0 and __container_of > > drivers/staging/lustre/include/linux/libcfs/libcfs.h | 11 ----------- > drivers/staging/lustre/lustre/lclient/lcommon_cl.c | 10 +++++----- > drivers/staging/lustre/lustre/llite/llite_nfs.c | 2 +- > drivers/staging/lustre/lustre/obdclass/lu_object.c | 6 +++--- > drivers/staging/lustre/lustre/obdecho/echo_client.c | 2 +- > drivers/staging/lustre/lustre/osc/osc_io.c | 4 ++-- > drivers/staging/lustre/lustre/osc/osc_object.c | 2 +- > 7 files changed, 13 insertions(+), 24 deletions(-) > >-- >2.4.2 > > >-- >Kind Regards, >Aya Saif El-yazal Mahfouz > Cheers, Andreas -- Andreas Dilger Lustre Software Architect Intel High Performance Data Division _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
