Messages in this thread | | | Date | Tue, 10 Feb 2009 13:10:34 -0600 | From | Steve Wise <> | Subject | Re: [PATCH 2.6.30] RDMA/cxgb3: Remove modulo math. |
| |
Randy Dunlap wrote: > Steve Wise wrote: > >> From: Steve Wise <swise@opengridcomputing.com> >> >> Removes the need for special u64 math on i386 systems. >> >> Fixes i386 build break in linux-next introduced by >> commit 1e27e8cee0698259ccb1fe6abeaf4b48969c0945. >> >> Signed-off-by: Steve Wise <swise@opengridcomputing.com> >> > > Yes, that works, thanks. But this patch should go into 2.6.29, not > just 2.6.30. > > > I thought the commit that caused this was:
1e27e8cee0698259ccb1fe6abeaf4b48969c0945
And that was going in 2.6.30. (I thought).
>> --- >> >> drivers/infiniband/hw/cxgb3/iwch_qp.c | 8 ++++---- >> 1 files changed, 4 insertions(+), 4 deletions(-) >> >> diff --git a/drivers/infiniband/hw/cxgb3/iwch_qp.c b/drivers/infiniband/hw/cxgb3/iwch_qp.c >> index 2cf6f13..5bb299a 100644 >> --- a/drivers/infiniband/hw/cxgb3/iwch_qp.c >> +++ b/drivers/infiniband/hw/cxgb3/iwch_qp.c >> @@ -232,8 +232,8 @@ static int iwch_sgl2pbl_map(struct iwch_dev *rhp, struct ib_sge *sg_list, >> return -EINVAL; >> } >> offset = sg_list[i].addr - mhp->attr.va_fbo; >> - offset += ((u64) mhp->attr.va_fbo) % >> - (1UL << (12 + mhp->attr.page_size)); >> + offset += mhp->attr.va_fbo & >> + ((1UL << (12 + mhp->attr.page_size)) - 1); >> pbl_addr[i] = ((mhp->attr.pbl_addr - >> rhp->rdev.rnic_info.pbl_base) >> 3) + >> (offset >> (12 + mhp->attr.page_size)); >> @@ -263,8 +263,8 @@ static int build_rdma_recv(struct iwch_qp *qhp, union t3_wr *wqe, >> wqe->recv.sgl[i].len = cpu_to_be32(wr->sg_list[i].length); >> >> /* to in the WQE == the offset into the page */ >> - wqe->recv.sgl[i].to = cpu_to_be64(((u32) wr->sg_list[i].addr) % >> - (1UL << (12 + page_size[i]))); >> + wqe->recv.sgl[i].to = cpu_to_be64(((u64) wr->sg_list[i].addr) & >> + ((1UL << (12 + page_size[i]))-1)); >> >> /* pbl_addr is the adapters address in the PBL */ >> wqe->recv.pbl_addr[i] = cpu_to_be32(pbl_addr[i]); >> > > >
| |