On Wed, Apr 05, 2017 at 05:04:35PM +0200, SF Markus Elfring wrote: > >> @@ -270,13 +268,14 @@ int qib_init_pportdata(struct qib_pportdata *ppd, > >> struct qib_devdata *dd, > >> if (!ppd->congestion_entries) > >> goto bail_1; > >> > >> - size = sizeof(struct cc_table_shadow); > >> - ppd->ccti_entries_shadow = kzalloc(size, GFP_KERNEL); > >> + ppd->ccti_entries_shadow = kzalloc(sizeof(*ppd->ccti_entries_shadow), > >> + GFP_KERNEL); > >> if (!ppd->ccti_entries_shadow) > >> goto bail_2; > >> > >> - size = sizeof(struct ib_cc_congestion_setting_attr); > >> - ppd->congestion_entries_shadow = kzalloc(size, GFP_KERNEL); > >> + ppd->congestion_entries_shadow = kzalloc(sizeof(*ppd > >> + > >> ->congestion_entries_shadow), > >> + GFP_KERNEL); > > > > Not related to this patch but is related to your patch-set - can you check > > the array allocations in lines 264 and 268? > > Do you refer to source code places here which are affected by the update step > "[PATCH 4/5] IB/qib: Use kcalloc() in qib_init_pportdata()"?
Oops, please ignore. > > > > Besides that: > > Reviewed-by: Yuval Shaia <yuval.sh...@oracle.com> > > Do you find the proposed change for the shown data types really acceptable > in these function calls? I found that the fix brings no harm to the existing code. > > Regards, > Markus
