[PATCH v2 rdma-next 0/2] irdma coverity fixes

[PATCH v2 rdma-next 0/2] irdma coverity fixes

[Tatyana Nikolova]

This is a short series of coverity fixes for irdma.

Shiraz Saleem (2):
  RDMA/irdma: Check contents of user-space irdma_mem_reg_req object
  RDMA/irdma: Fix potential overflow expression in irdma_prm_get_pbles

v1->v2
* Add u32 sums for u16 variables to show that the operations are overflow safe.
* Replace shifting ops with DIV_ROUND_UP_ULL macro to get bits_needed

 drivers/infiniband/hw/irdma/pble.h  |  2 +-
 drivers/infiniband/hw/irdma/utils.c |  4 ++--
 drivers/infiniband/hw/irdma/verbs.c | 26 ++++++++++++++++++++------
 3 files changed, 23 insertions(+), 9 deletions(-)

-- 
2.27.0

[PATCH v2 rdma-next 1/2] RDMA/irdma: Check contents of user-space irdma_mem_reg_req object

[Tatyana Nikolova]

From: Shiraz Saleem <shiraz.saleem@intel.com>

The contents of user-space req object is used in array indexing
in irdma_handle_q_mem without checking for valid values.

Guard against bad input on each of these req object pages by
limiting them to number of pages that make up the region.

Reported-by: coverity-bot <keescook+coverity-bot@chromium.org>
Addresses-Coverity-ID: 1505160 ("TAINTED_SCALAR")
Fixes: b48c24c2d710 ("RDMA/irdma: Implement device supported verb APIs")
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com>
---
 drivers/infiniband/hw/irdma/verbs.c | 26 ++++++++++++++++++++------
 1 file changed, 20 insertions(+), 6 deletions(-)

diff --git a/drivers/infiniband/hw/irdma/verbs.c b/drivers/infiniband/hw/irdma/verbs.c
index 5bb46a4d26ff..9712f6902ba8 100644
--- a/drivers/infiniband/hw/irdma/verbs.c
+++ b/drivers/infiniband/hw/irdma/verbs.c
@@ -2358,12 +2358,10 @@ static int irdma_handle_q_mem(struct irdma_device *iwdev,
 	struct irdma_cq_mr *cqmr = &iwpbl->cq_mr;
 	struct irdma_hmc_pble *hmc_p;
 	u64 *arr = iwmr->pgaddrmem;
-	u32 pg_size;
+	u32 pg_size, total;
 	int err = 0;
-	int total;
 	bool ret = true;
 
-	total = req->sq_pages + req->rq_pages + req->cq_pages;
 	pg_size = iwmr->page_size;
 	err = irdma_setup_pbles(iwdev->rf, iwmr, use_pbles);
 	if (err)
@@ -2380,6 +2378,7 @@ static int irdma_handle_q_mem(struct irdma_device *iwdev,
 
 	switch (iwmr->type) {
 	case IRDMA_MEMREG_TYPE_QP:
+		total = req->sq_pages + req->rq_pages;
 		hmc_p = &qpmr->sq_pbl;
 		qpmr->shadow = (dma_addr_t)arr[total];
 
@@ -2406,7 +2405,7 @@ static int irdma_handle_q_mem(struct irdma_device *iwdev,
 		hmc_p = &cqmr->cq_pbl;
 
 		if (!cqmr->split)
-			cqmr->shadow = (dma_addr_t)arr[total];
+			cqmr->shadow = (dma_addr_t)arr[req->cq_pages];
 
 		if (use_pbles)
 			ret = irdma_check_mem_contiguous(arr, req->cq_pages,
@@ -2747,7 +2746,8 @@ static struct ib_mr *irdma_reg_user_mr(struct ib_pd *pd, u64 start, u64 len,
 	struct irdma_mr *iwmr;
 	struct ib_umem *region;
 	struct irdma_mem_reg_req req;
-	u32 stag = 0;
+	u32 total, stag = 0;
+	u8 shadow_pgcnt = 1;
 	bool use_pbles = false;
 	unsigned long flags;
 	int err = -EINVAL;
@@ -2801,7 +2801,13 @@ static struct ib_mr *irdma_reg_user_mr(struct ib_pd *pd, u64 start, u64 len,
 
 	switch (req.reg_type) {
 	case IRDMA_MEMREG_TYPE_QP:
-		use_pbles = ((req.sq_pages + req.rq_pages) > 2);
+		total = req.sq_pages + req.rq_pages + shadow_pgcnt;
+		if (total > iwmr->page_cnt) {
+			err = -EINVAL;
+			goto error;
+		}
+		total = req.sq_pages + req.rq_pages;
+		use_pbles = (total > 2);
 		err = irdma_handle_q_mem(iwdev, &req, iwpbl, use_pbles);
 		if (err)
 			goto error;
@@ -2814,6 +2820,14 @@ static struct ib_mr *irdma_reg_user_mr(struct ib_pd *pd, u64 start, u64 len,
 		spin_unlock_irqrestore(&ucontext->qp_reg_mem_list_lock, flags);
 		break;
 	case IRDMA_MEMREG_TYPE_CQ:
+		if (iwdev->rf->sc_dev.hw_attrs.uk_attrs.feature_flags & IRDMA_FEATURE_CQ_RESIZE)
+			shadow_pgcnt = 0;
+		total = req.cq_pages + shadow_pgcnt;
+		if (total > iwmr->page_cnt) {
+			err = -EINVAL;
+			goto error;
+		}
+
 		use_pbles = (req.cq_pages > 1);
 		err = irdma_handle_q_mem(iwdev, &req, iwpbl, use_pbles);
 		if (err)
-- 
2.27.0

[PATCH v2 rdma-next 2/2] RDMA/irdma: Fix potential overflow expression in irdma_prm_get_pbles

[Tatyana Nikolova]

From: Shiraz Saleem <shiraz.saleem@intel.com>

Coverity reports a signed 32-bit overflow on "1 << pprm->pble_shift" when
used expression to compute bits_needed that expects 64bit, unsigned.

Fix this by using the 1ULL in the left shift operator and convert
mem_size to u64.

Reported-by: coverity-bot <keescook+coverity-bot@chromium.org>
Addresses-Coverity-ID: 1505157 ("Integer handling issues")
Fixes: 915cc7ac0f8e ("RDMA/irdma: Add miscellaneous utility definitions")
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com>
---
 drivers/infiniband/hw/irdma/pble.h  | 2 +-
 drivers/infiniband/hw/irdma/utils.c | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/hw/irdma/pble.h b/drivers/infiniband/hw/irdma/pble.h
index e4e635dc4fd9..e1b3b8118a2c 100644
--- a/drivers/infiniband/hw/irdma/pble.h
+++ b/drivers/infiniband/hw/irdma/pble.h
@@ -121,7 +121,7 @@ enum irdma_status_code irdma_prm_add_pble_mem(struct irdma_pble_prm *pprm,
 					      struct irdma_chunk *pchunk);
 enum irdma_status_code
 irdma_prm_get_pbles(struct irdma_pble_prm *pprm,
-		    struct irdma_pble_chunkinfo *chunkinfo, u32 mem_size,
+		    struct irdma_pble_chunkinfo *chunkinfo, u64 mem_size,
 		    u64 **vaddr, u64 *fpm_addr);
 void irdma_prm_return_pbles(struct irdma_pble_prm *pprm,
 			    struct irdma_pble_chunkinfo *chunkinfo);
diff --git a/drivers/infiniband/hw/irdma/utils.c b/drivers/infiniband/hw/irdma/utils.c
index ea1df5918c11..5bbe44e54f9a 100644
--- a/drivers/infiniband/hw/irdma/utils.c
+++ b/drivers/infiniband/hw/irdma/utils.c
@@ -2314,7 +2314,7 @@ enum irdma_status_code irdma_prm_add_pble_mem(struct irdma_pble_prm *pprm,
  */
 enum irdma_status_code
 irdma_prm_get_pbles(struct irdma_pble_prm *pprm,
-		    struct irdma_pble_chunkinfo *chunkinfo, u32 mem_size,
+		    struct irdma_pble_chunkinfo *chunkinfo, u64 mem_size,
 		    u64 **vaddr, u64 *fpm_addr)
 {
 	u64 bits_needed;
@@ -2326,7 +2326,7 @@ irdma_prm_get_pbles(struct irdma_pble_prm *pprm,
 	*vaddr = NULL;
 	*fpm_addr = 0;
 
-	bits_needed = (mem_size + (1 << pprm->pble_shift) - 1) >> pprm->pble_shift;
+	bits_needed = DIV_ROUND_UP_ULL(mem_size, BIT_ULL(pprm->pble_shift));
 
 	spin_lock_irqsave(&pprm->prm_lock, flags);
 	while (chunk_entry != &pprm->clist) {
-- 
2.27.0

Re: [PATCH v2 rdma-next 0/2] irdma coverity fixes

[Jason Gunthorpe]

On Fri, Jun 25, 2021 at 11:23:27AM -0500, Tatyana Nikolova wrote:
> This is a short series of coverity fixes for irdma.
> 
> Shiraz Saleem (2):
>   RDMA/irdma: Check contents of user-space irdma_mem_reg_req object
>   RDMA/irdma: Fix potential overflow expression in irdma_prm_get_pbles

Applied to for-next, thanks

Jason