Commit 2ff7d09a authored Jun 01, 2011 by Steve Wise Committed by Roland Dreier Jun 17, 2011

RDMA/cxgb4: Don't exceed hw IQ depth limit for user CQs



Memory allocated for user CQs gets rounded up to the next page
boundary.  And after rounding, we recalculate the resulting IQ depth
and we need to make sure we don't exceed the HW limits.

This bug can result a much smaller CQ allocated than was expected if
the HW size field is exceeded, resulting in CQ overflow failures.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>

parent 2c53b436

drivers/infiniband/hw/cxgb4/cq.c

+4 −0

Original line number	Original line	Diff line number	Diff line
	@@ -801,6 +801,10 @@ struct ib_cq c4iw_create_cq(struct ib_device ibdev, int entries,
	if (ucontext) {		if (ucontext) {
	memsize = roundup(memsize, PAGE_SIZE);		memsize = roundup(memsize, PAGE_SIZE);
	hwentries = memsize / sizeof *chp->cq.queue;		hwentries = memsize / sizeof *chp->cq.queue;
			while (hwentries > T4_MAX_IQ_SIZE) {
			memsize -= PAGE_SIZE;
			hwentries = memsize / sizeof *chp->cq.queue;
			}
	}		}
	chp->cq.size = hwentries;		chp->cq.size = hwentries;
	chp->cq.memsize = memsize;		chp->cq.memsize = memsize;