[PATCH for-next 0/6] minor mlx5 fixes for 3.18

[PATCH for-next 0/6] minor mlx5 fixes for 3.18

[Eli Cohen]

Hi Roland,

the following series of patches contains some trivial fixes. The patches are
rebased on Linus's tree since they refer to mlx5 patches that are not yet on 
your ib-next branch, namely:

4d2f9bb mlx5: Adjust events to use unsigned long param instead of void *
f241e74 mlx5: minor fixes (mainly avoidance of hidden casts)
9603b61 mlx5: Move pci device handling from mlx5_ib to mlx5_core

Eli

Eli Cohen (5):
  IB/mlx5: Clear umr resources after ib_unregister_device
  IB/mlx5: Improve debug prints in mlx5_ib_reg_user_mr
  IB/core: Avoid leakage from kernel to user space
  IB/mlx5: Fix possible array overflow
  IB/mlx5: Remove duplicate code from mlx5_set_path

Yishai Hadas (1):
  IB/mlx5: Modify to work with arbitrary page size

 drivers/infiniband/core/uverbs_main.c |  1 +
 drivers/infiniband/hw/mlx5/main.c     |  8 ++++----
 drivers/infiniband/hw/mlx5/mem.c      | 18 ++++++++++--------
 drivers/infiniband/hw/mlx5/mr.c       |  6 +++---
 drivers/infiniband/hw/mlx5/qp.c       | 21 +++++----------------
 5 files changed, 23 insertions(+), 31 deletions(-)

-- 
2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-next 1/6] IB/mlx5: Clear umr resources after ib_unregister_device

[Eli Cohen]

Some ULPs may make use of resources created in create_umr_res so make sure to
call destroy_umrc_res after returning from ib_unregister_device which makes
sure all ULPs have closed their resources.

Signed-off-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/hw/mlx5/main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index d8907b20522a..ecd3aebc46fe 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -1414,8 +1414,8 @@ err_dealloc:
 static void mlx5_ib_remove(struct mlx5_core_dev *mdev, void *context)
 {
 	struct mlx5_ib_dev *dev = context;
-	destroy_umrc_res(dev);
 	ib_unregister_device(&dev->ib_dev);
+	destroy_umrc_res(dev);
 	destroy_dev_resources(&dev->devr);
 	free_comp_eqs(dev);
 	ib_dealloc_device(&dev->ib_dev);
-- 
2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-next 2/6] IB/mlx5: Improve debug prints in mlx5_ib_reg_user_mr

[Eli Cohen]

Print access flags and error code from ib_umem_get.

Signed-off-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/hw/mlx5/mr.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c
index 80b3c63eab5d..8ee7cb46e059 100644
--- a/drivers/infiniband/hw/mlx5/mr.c
+++ b/drivers/infiniband/hw/mlx5/mr.c
@@ -881,12 +881,12 @@ struct ib_mr *mlx5_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	int order;
 	int err;
 
-	mlx5_ib_dbg(dev, "start 0x%llx, virt_addr 0x%llx, length 0x%llx\n",
-		    start, virt_addr, length);
+	mlx5_ib_dbg(dev, "start 0x%llx, virt_addr 0x%llx, length 0x%llx, access_flags 0x%x\n",
+		    start, virt_addr, length, access_flags);
 	umem = ib_umem_get(pd->uobject->context, start, length, access_flags,
 			   0);
 	if (IS_ERR(umem)) {
-		mlx5_ib_dbg(dev, "umem get failed\n");
+		mlx5_ib_dbg(dev, "umem get failed (%ld)\n", PTR_ERR(umem));
 		return (void *)umem;
 	}
 
-- 
2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-next 3/6] IB/core: Avoid leakage from kernel to user space

[Eli Cohen]

Clear the reserved field of struct ib_uverbs_async_event_desc which is 
copied to user space.

Signed-off-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/core/uverbs_main.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
index c73b22a257fe..b21af1eaf503 100644
--- a/drivers/infiniband/core/uverbs_main.c
+++ b/drivers/infiniband/core/uverbs_main.c
@@ -477,6 +477,7 @@ static void ib_uverbs_async_handler(struct ib_uverbs_file *file,
 
 	entry->desc.async.element    = element;
 	entry->desc.async.event_type = event;
+	entry->desc.async.reserved   = 0;
 	entry->counter               = counter;
 
 	list_add_tail(&entry->list, &file->async_file->event_list);
-- 
2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-next 4/6] IB/mlx5: Fix possible array overflow

[Eli Cohen]

The check to verify that userspace does not provide an invalid index to the
micro UAR was placed too late. Fix this by moving the check before using the
index.

Reported by: Shachar Raindel <raindel-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/hw/mlx5/main.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index ecd3aebc46fe..a24431746377 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -650,13 +650,13 @@ static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vm
 			return -EINVAL;
 
 		idx = get_index(vma->vm_pgoff);
+		if (idx >= uuari->num_uars)
+			return -EINVAL;
+
 		pfn = uar_index2pfn(dev, uuari->uars[idx].index);
 		mlx5_ib_dbg(dev, "uar idx 0x%lx, pfn 0x%llx\n", idx,
 			    (unsigned long long)pfn);
 
-		if (idx >= uuari->num_uars)
-			return -EINVAL;
-
 		vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
 		if (io_remap_pfn_range(vma, vma->vm_start, pfn,
 				       PAGE_SIZE, vma->vm_page_prot))
-- 
2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-next 5/6] IB/mlx5: Remove duplicate code from mlx5_set_path

[Eli Cohen]

Some of the fields where set twice. Re-organize to avoid that.

Signed-off-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/hw/mlx5/qp.c | 21 +++++----------------
 1 file changed, 5 insertions(+), 16 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index 8c574b63d77b..34b92fc345b7 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -1302,6 +1302,11 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, const struct ib_ah_attr *ah,
 	path->rlid	= cpu_to_be16(ah->dlid);
 
 	if (ah->ah_flags & IB_AH_GRH) {
+		if (ah->grh.sgid_index >= dev->mdev->caps.port[port - 1].gid_table_len) {
+			pr_err(KERN_ERR "sgid_index (%u) too large. max is %d\n",
+			       ah->grh.sgid_index, dev->mdev->caps.port[port - 1].gid_table_len);
+			return -EINVAL;
+		}
 		path->grh_mlid |= 1 << 7;
 		path->mgid_index = ah->grh.sgid_index;
 		path->hop_limit  = ah->grh.hop_limit;
@@ -1317,22 +1322,6 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, const struct ib_ah_attr *ah,
 	path->static_rate = err;
 	path->port = port;
 
-	if (ah->ah_flags & IB_AH_GRH) {
-		if (ah->grh.sgid_index >= dev->mdev->caps.port[port - 1].gid_table_len) {
-			pr_err(KERN_ERR "sgid_index (%u) too large. max is %d\n",
-			       ah->grh.sgid_index, dev->mdev->caps.port[port - 1].gid_table_len);
-			return -EINVAL;
-		}
-
-		path->grh_mlid |= 1 << 7;
-		path->mgid_index = ah->grh.sgid_index;
-		path->hop_limit  = ah->grh.hop_limit;
-		path->tclass_flowlabel =
-			cpu_to_be32((ah->grh.traffic_class << 20) |
-				    (ah->grh.flow_label));
-		memcpy(path->rgid, ah->grh.dgid.raw, 16);
-	}
-
 	if (attr_mask & IB_QP_TIMEOUT)
 		path->ackto_lt = attr->timeout << 3;
 
-- 
2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH for-next 6/6] IB/mlx5: Modify to work with arbitrary page size

[Eli Cohen]

From: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

When dealing with umem objects, the driver assumed host page sizes defined by
PAGE_SHIFT. Modify the code to use arbitrary page shift provided by umem->page_shift
to support different page sizes.

Signed-off-by: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/hw/mlx5/mem.c | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/mem.c b/drivers/infiniband/hw/mlx5/mem.c
index a3e81444c825..dae07eae9507 100644
--- a/drivers/infiniband/hw/mlx5/mem.c
+++ b/drivers/infiniband/hw/mlx5/mem.c
@@ -55,16 +55,17 @@ void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
 	u64 pfn;
 	struct scatterlist *sg;
 	int entry;
+	unsigned long page_shift = ilog2(umem->page_size);
 
-	addr = addr >> PAGE_SHIFT;
+	addr = addr >> page_shift;
 	tmp = (unsigned long)addr;
 	m = find_first_bit(&tmp, sizeof(tmp));
 	skip = 1 << m;
 	mask = skip - 1;
 	i = 0;
 	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
-		len = sg_dma_len(sg) >> PAGE_SHIFT;
-		pfn = sg_dma_address(sg) >> PAGE_SHIFT;
+		len = sg_dma_len(sg) >> page_shift;
+		pfn = sg_dma_address(sg) >> page_shift;
 		for (k = 0; k < len; k++) {
 			if (!(i & mask)) {
 				tmp = (unsigned long)pfn;
@@ -103,14 +104,15 @@ void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
 
 		*ncont = 0;
 	}
-	*shift = PAGE_SHIFT + m;
+	*shift = page_shift + m;
 	*count = i;
 }
 
 void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
 			  int page_shift, __be64 *pas, int umr)
 {
-	int shift = page_shift - PAGE_SHIFT;
+	unsigned long umem_page_shift = ilog2(umem->page_size);
+	int shift = page_shift - umem_page_shift;
 	int mask = (1 << shift) - 1;
 	int i, k;
 	u64 cur = 0;
@@ -121,11 +123,11 @@ void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
 
 	i = 0;
 	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
-		len = sg_dma_len(sg) >> PAGE_SHIFT;
+		len = sg_dma_len(sg) >> umem_page_shift;
 		base = sg_dma_address(sg);
 		for (k = 0; k < len; k++) {
 			if (!(i & mask)) {
-				cur = base + (k << PAGE_SHIFT);
+				cur = base + (k << umem_page_shift);
 				if (umr)
 					cur |= 3;
 
@@ -134,7 +136,7 @@ void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
 					    i >> shift, be64_to_cpu(pas[i >> shift]));
 			}  else
 				mlx5_ib_dbg(dev, "=====> 0x%llx\n",
-					    base + (k << PAGE_SHIFT));
+					    base + (k << umem_page_shift));
 			i++;
 		}
 	}
-- 
2.1.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH for-next 3/6] IB/core: Avoid leakage from kernel to user space

[Yann Droneaud]

Hi,

Le dimanche 14 septembre 2014 à 16:47 +0300, Eli Cohen a écrit :
> Clear the reserved field of struct ib_uverbs_async_event_desc which is 
> copied to user space.
> 

This one could be applied right now on v3.17 and others.

> Signed-off-by: Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Reviewed-by: Yann Droneaud <ydroneaud-RlY5vtjFyJ3QT0dZR+AlfA@public.gmane.org>
Cc: stable-u79uwXL29TY76Z2rM5mHXA@public.gmane.org

Regards.

> ---
>  drivers/infiniband/core/uverbs_main.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
> index c73b22a257fe..b21af1eaf503 100644
> --- a/drivers/infiniband/core/uverbs_main.c
> +++ b/drivers/infiniband/core/uverbs_main.c
> @@ -477,6 +477,7 @@ static void ib_uverbs_async_handler(struct ib_uverbs_file *file,
>  
>  	entry->desc.async.element    = element;
>  	entry->desc.async.event_type = event;
> +	entry->desc.async.reserved   = 0;
>  	entry->counter               = counter;
>  
>  	list_add_tail(&entry->list, &file->async_file->event_list);


--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[-stable] commit 377b513485fd ("IB/core: Avoid leakage from kernel to user space")

[Yann Droneaud]

Hi,

Please add commit 377b513485fd ("IB/core: Avoid leakage from kernel to
user space") to -stable. It can be applied to v2.6.32 and later.

Regards.

-- 
Yann Droneaud
OPTEYA

Re: [-stable] commit 377b513485fd ("IB/core: Avoid leakage from kernel to user space")

[Luis Henriques]

On Fri, Mar 27, 2015 at 01:42:44PM +0100, Yann Droneaud wrote:
> Hi,
> 
> Please add commit 377b513485fd ("IB/core: Avoid leakage from kernel to
> user space") to -stable. It can be applied to v2.6.32 and later.
> 
> Regards.
> 
> -- 
> Yann Droneaud
> OPTEYA
> 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe stable" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Thanks, I'm queuing it for the 3.16 kernel.

Cheers,
--
Luís
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [-stable] commit 377b513485fd ("IB/core: Avoid leakage from kernel to user space")

[Jiri Slaby]

On 03/27/2015, 01:42 PM, Yann Droneaud wrote:
> Hi,
> 
> Please add commit 377b513485fd ("IB/core: Avoid leakage from kernel to
> user space") to -stable. It can be applied to v2.6.32 and later.

Now pushed to 3.12. Thanks.

-- 
js
suse labs

Re: [-stable] commit 377b513485fd ("IB/core: Avoid leakage from kernel to user space")

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 340 bytes --]

On Fri, 2015-03-27 at 13:42 +0100, Yann Droneaud wrote:
> Hi,
> 
> Please add commit 377b513485fd ("IB/core: Avoid leakage from kernel to
> user space") to -stable. It can be applied to v2.6.32 and later.

Queued up for 3.2, thanks.

Ben.

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 811 bytes --]

Re: [-stable] commit 377b513485fd ("IB/core: Avoid leakage from kernel to user space")

[Zefan Li]

On 2015/3/27 20:42, Yann Droneaud worte:
> Hi,
> 
> Please add commit 377b513485fd ("IB/core: Avoid leakage from kernel to
> user space") to -stable. It can be applied to v2.6.32 and later.
> 

Queued up for 3.4. Thanks!

Re: [-stable] commit 377b513485fd ("IB/core: Avoid leakage from kernel to user space")

[Zefan Li]

On 2015/3/27 20:42, Yann Droneaud worte:
> Hi,
> 
> Please add commit 377b513485fd ("IB/core: Avoid leakage from kernel to
> user space") to -stable. It can be applied to v2.6.32 and later.
> 

Queued up for 3.4. Thanks!