On 04/13/2015 09:27 PM, Jason Gunthorpe wrote:
On Mon, Apr 13, 2015 at 02:25:16PM +0200, Michael Wang wrote:
dev_list = kmalloc(sizeof *dev_list, GFP_KERNEL);
if (!dev_list)
@@ -1673,13 +1671,19 @@ static void ipoib_add_one(struct ib_device *device)
}
for (p = s; p =
ping
On Sun, Feb 22, 2015 at 08:50:27PM -0800, Yuval Shaia wrote:
Current approach force one to implement all ops even when some functions can
use the default implementation.
As a result, for new DMA ops (e.x new arch) many functions just wrap the
default function.
The fix is to check each
On 04/13/2015 09:25 PM, Hefty, Sean wrote:
@@ -1037,17 +1033,13 @@ void rdma_destroy_id(struct rdma_cm_id *id)
mutex_unlock(id_priv-handler_mutex);
if (id_priv-cma_dev) {
-switch (rdma_node_get_transport(id_priv-id.device-
node_type)) {
-case
On 04/13/2015 08:12 PM, ira.weiny wrote:
[snip]
-
-if (rdma_node_get_transport(ib_device-node_type) != RDMA_TRANSPORT_IB)
-return;
+int count = 0;
I'm ok with this as an intermediate patch but going forward if we are going to
have calls like
static inline int
On 04/13/2015 08:45 PM, Hefty, Sean wrote:
@@ -1153,9 +1153,7 @@ static void ib_sa_add_one(struct ib_device *device)
{
struct ib_sa_device *sa_dev;
int s, e, i;
-
-if (rdma_node_get_transport(device-node_type) != RDMA_TRANSPORT_IB)
-return;
+int count = 0;
On 13/04/2015 16:29, Yann Droneaud wrote:
Le jeudi 02 avril 2015 à 18:12 +, Haggai Eran a écrit :
...
I want to add that we would like to see users registering a very large
memory region (perhaps the entire process address space) for local
access, and then enabling remote access only to
On 2/23/2015 6:50 AM, Yuval Shaia wrote:
Current approach force one to implement all ops even when some functions can
use the default implementation.
As a result, for new DMA ops (e.x new arch) many functions just wrap the
default function.
The fix is to check each DMA operation individually
On 04/13/2015 10:33 PM, Jason Gunthorpe wrote:
On Mon, Apr 13, 2015 at 02:36:45PM +0200, Michael Wang wrote:
We have finished introducing the cap_XX(), and raw helper rdma_ib_or_iboe()
is no longer necessary, thus clean it up.
So, the net result is not looking too bad, but I'm confused
On 04/13/2015 09:29 PM, Jason Gunthorpe wrote:
On Mon, Apr 13, 2015 at 06:40:35PM +, Hefty, Sean wrote:
- if (rdma_node_get_transport(ib_device-node_type) !=
RDMA_TRANSPORT_IB)
- return;
+ int count = 0;
I'm ok with this as an intermediate patch but going forward if we are
On 4/13/2015 3:56 PM, Yann Droneaud wrote:
In a call to ib_umem_get(), if address is 0x0 and size is
already page aligned, check added in commit 8494057ab5e4
(IB/uverbs: Prevent integer overflow in ib_umem_get address
arithmetic) will refuse to register a memory region that
could otherwise be
Hi,
Le mardi 14 avril 2015 à 12:20 +0300, Sagi Grimberg a écrit :
On 4/13/2015 3:56 PM, Yann Droneaud wrote:
In a call to ib_umem_get(), if address is 0x0 and size is
already page aligned, check added in commit 8494057ab5e4
(IB/uverbs: Prevent integer overflow in ib_umem_get address
On Mon, Apr 13, 2015 at 8:53 PM, Jason Gunthorpe
jguntho...@obsidianresearch.com wrote:
Or, I was also a bit annoyed by your last email, the tone just seemed
inappropriate/in-gracious for the situation.
I made a full clarification and wait to see Sean's response. Sean is
doing excellent work
On Mon, Apr 13, 2015 at 8:53 PM, Jason Gunthorpe
jguntho...@obsidianresearch.com wrote:
As far as patches go, you keep asking for reviews, but where are the
reviews from Mellanox? I count 45 @mellanox folks who contributed to
the kernel, but I wouldn't have guessed that based on mailing
list
Signed-off-by: Fengguang Wu fengguang...@intel.com
---
ipoib_multicast.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
index 856ed6e..f5c5f14 100644
---
tree: git://flatbed.openfabrics.org/~amirv/linux.git for-upstream
head: 66a75344a2e5a8e8483cfd6ab1db1c0fbcb56fda
commit: cbaa5ae8702fea9670d77daf95f35b4d38822ad0 [52/62] IB/ipoib: Add cleanup
to sendonly multicast objects [ib-next:ipoib-s1]
reproduce:
# apt-get install sparse
git checkout
On Mon, Apr 13, 2015 at 7:10 PM, Hefty, Sean sean.he...@intel.com wrote:
Unfortunately, Sean -- you haven't provided for 2 months any
feedback on the patches to the CM/CMA modules which you own/maintain
for supporting name-spaces and such which are part of enabling proper
RDMA support for
On 4/14/2015 2:50 AM, Hefty, Sean wrote:
Yes, this comment probably could use a reword..
+ IB_GID_TYPE_IB= 0,
+ IB_GID_TYPE_ROCE_V2 = 1,
+ IB_GID_TYPE_SIZE
+};
Can you explain the purpose of defining a 'GID type'. A GID is just a
global
address. Why does it
On 4/7/2015 2:22 PM, Sébastien Dugué wrote:
Hello,
When an application using XRCs abruptly terminates, the mmaped pages
of the CQ buffers are leaked.
This comes from the fact that when resources are released in
ib_uverbs_cleanup_ucontext(), we fail to release the CQs because their
On 3/26/2015 1:46 AM, Bart Van Assche wrote:
On 03/25/2015 02:19 PM, Somnath Kotur wrote:
+static void ib_device_complete_cb(struct kref *kref)
+{
+struct ib_device *device = container_of(kref, struct ib_device,
+refcount);
+
+if (device-reg_state =
1) Droped ipvlan/macvlan/dsa patches.
2) Do not check the pointer returned by netdev_priv.
3) Add the FIX tag.
- console log ---
ipoib_get_iflink dev = 88026eb1c000, priv = 88026eb1c7c0, priv-parent
= (null)
ipoib_get_iflink dev =
Le 14/04/2015 18:30, Erez Shitrit a écrit :
On Tue, Apr 14, 2015 at 6:20 PM, Honggang Li ho...@redhat.com wrote:
[snip]
This will make parent interface to return 0 instead of its own ifindex.
I would suggest write something like that:
+ /* parent interface */
+ if
On 04/14/2015 04:18 PM, ira.weiny wrote:
[snip]
/**
- * cap_ib_cm_dev - Check if any port of device has the capability Infiniband
- * Communication Manager.
+ * cap_ib_cm_any_port - Check if any port of the device has Infiniband
+ * Communication Manager (CM) support.
*
* @device:
This struct members other than struct iser_mem_reg are unused,
so remove it altogether.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg sa...@mellanox.com
Signed-off-by: Adir Lev ad...@mellanox.com
---
drivers/infiniband/ulp/iser/iscsi_iser.h | 21 +-
This code was added before we had protection data length
calculation (in iser_send_command), so we needed to calc
the sg data length from the sg itself. This is not needed
anymore.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg sa...@mellanox.com
Signed-off-by: Adir
Buffer length was assigned twice, and no reason to set va to
io_addr and then add the offset, just set va to io_addr + offset.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg sa...@mellanox.com
Signed-off-by: Adir Lev ad...@mellanox.com
---
No need to keep two iser_data_buf structures just in case we use
mem copy. We can avoid that just by adding a pointer to the original
sg. So keep only two iser_data_buf per command (data and protection)
and pass the relevant data_buf to bounce buffer routine.
This patch does not change any
Signed-off-by: Sagi Grimberg sa...@mellanox.com
---
drivers/infiniband/ulp/iser/iscsi_iser.h |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h
b/drivers/infiniband/ulp/iser/iscsi_iser.h
index 5fd0963..a39645a 100644
---
In case the user unloaded ib_iser while ep_connect is in
progress, we need to destroy the endpoint although ep_disconnect
wasn't invoked (we detect this by the iser conn state != DOWN).
However, if we got an REJECTED/UNREACHABLE CM event we move the
connection state to DOWN which will prevent us
Fast registration and local invalidate work requests can
also fail. We should call error completion handler for them.
Reported-by: Roi Dayan r...@mellanox.com
Signed-off-by: Sagi Grimberg sa...@mellanox.com
---
drivers/infiniband/ulp/iser/iser_verbs.c | 11 ++-
1 files changed, 6
In some rare cases, IO operations may be not aligned to page
boundaries. This prevents iser from performing fast memory
registration. In order to overcome that iser uses a bounce
buffer to carry the transaction. We basically allocate a buffer
in the size of the transaction and perform a copy.
The
This length miss-calculation may cause a silent data corruption
in the DIX case and cause the device to reference unmapped area.
Fixes: d77e65350f2d ('libiscsi, iser: Adjust data_length to include protection
information')
Signed-off-by: Sagi Grimberg sa...@mellanox.com
---
Instead of passing ib_sge as output variable, we pass the mem_reg
pointer to have the routines fill the rkey as well. This reduces
code duplication and extra assignments. This is a preparation step
to unify some registration logics together. Also, pass iser_fast_reg_mr
the fastreg descriptor
In singleton scatterlists, DMA memory registration code
is taken both for Fastreg and FMR code paths. Move it to
a function.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg sa...@mellanox.com
Signed-off-by: Adir Lev ad...@mellanox.com
---
Le 14/04/2015 17:26, Patrick McHardy a écrit :
On 14.04, Honggang Li wrote:
[snip]
- return vlan-lowerdev-ifindex;
+ if (vlan vlan-lowerdev)
+ return vlan-lowerdev-ifindex;
That is completely useless. vlan (=netdev_priv) can not be NULL as
netdev_priv() never
related qib. In order to check upstream linux-next fix
the race condition or not, I build linux-next-20150414 on two machines. Both
machines panic at modprobe ib_ipoib. Do you means I need to report a
bug? But I do not know report to who or where.
Here is the tag:
Fixes: 5aa7add8f14b
Starting monitoring for VG vg_rdma01: 3 logical volume(s) in volume
group vg_rdma01 monitored
[ OK ]
Starting cgconfig service: Failed to parse /etc/cgconfig.conf or
/etc/cgconfig.d[FAILED]
Loading OpenIB kernel modules:
BUG: unable to handle kernel NULL pointer dereference at
0120
This is a part of the GID meta info. The user should be able to choose
between RoCE V1 (which is represented here by IB_GID_TYPE_IB) and RoCE
V2 - just as a user could choose between IPv6 and IPv4.
IPv4 and IPv6 are different protocols, not different formats for the same
address. How does
From: Honggang Li ho...@redhat.com
Date: Mon, 13 Apr 2015 17:21:58 +0800
If CONFIG_ARCH_DMA_ADDR_T_64BIT enabled for x86 systems and physical
memory is more than 4GB, dma_map_page may return a valid memory
address which greater than 0x. As a result, the mlx5 device page
allocator RB
-20150414 on two machines. Both
machines panic at modprobe ib_ipoib. Do you means I need to report a
bug? But I do not know report to who or where.
Here is the tag:
Fixes: 5aa7add8f14b (infiniband/ipoib: implement ndo_get_iflink)
--
To unsubscribe from this list: send the line unsubscribe linux-rdma
today when I'm tracing a race
condition issue related qib. In order to check upstream linux-next fix
the race condition or not, I build linux-next-20150414 on two machines.
Both
machines panic at modprobe ib_ipoib. Do you means I need to report a
bug? But I do not know report to who or where
On Tue, Apr 14, 2015 at 6:20 PM, Honggang Li ho...@redhat.com wrote:
[...]
Hi,
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c
b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 657b89b..11ea6e2 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++
On Tue, Apr 14, 2015 at 01:43:11PM -0400, ira.weiny wrote:
A failure to init port 2 ends up ends up killing port 1 and releasing the
device associated resources.
Yes, that is the only reasonable thing that could happen.
init failure should only be possible under exceptional cases (OOM).
The
Le 14/04/2015 18:01, Yann Droneaud a écrit :
[snip]
Here is the tag:
Fixes: 5aa7add8f14b (infiniband/ipoib: implement ndo_get_iflink)
Pardon me, but this patch was never submitted to
linux-rdma@vger.kernel.org for review !?
Sorry for that, I missed it. Only Roland Dreier was CCed.
--
To
But that moves us in the wrong direction. If we later support port 2
without
port 1 that code will be broken.
I agree that the code will be broken, but supporting that model requires a lot
more work in how the ib_cm listens across devices.
--
To unsubscribe from this list: send the line
On Tue, Apr 14, 2015 at 11:25:15AM -0600, Jason Gunthorpe wrote:
On Tue, Apr 14, 2015 at 10:18:07AM -0400, ira.weiny wrote:
After more thought and reading other opinions, I must agree we should not
have cap_foo_dev.
I looked at it a bit, and I think Sean has also basically said, CM
does
On Tue, Apr 14, 2015 at 10:18:07AM -0400, ira.weiny wrote:
After more thought and reading other opinions, I must agree we should not
have cap_foo_dev.
I looked at it a bit, and I think Sean has also basically said, CM
does not support certain mixed port combinations. iWarp and IB simply
cannot
Yes, that is the only reasonable thing that could happen.
init failure should only be possible under exceptional cases (OOM).
The only system response is to call ib_umad_add_one again - so of
course the first call had to completely clean up everything it did.
A reasonable follow up change
On Tue, Apr 14, 2015 at 06:02:47PM +, Hefty, Sean wrote:
Yes, that is the only reasonable thing that could happen.
init failure should only be possible under exceptional cases (OOM).
The only system response is to call ib_umad_add_one again - so of
course the first call had to
On Mon, Apr 13, 2015 at 02:01:38PM -0600, Jason Gunthorpe wrote:
On Mon, Apr 13, 2015 at 03:46:03PM -0400, ira.weiny wrote:
This doesn't quite look right, it should be 'goto error1'
Looks like you replied to the wrong patch. ?? I don't see error1 in
ipoib_add_one.
For the ib_cm
On Tue, Apr 14, 2015 at 05:37:57PM +0200, Andrew Lunn wrote:
And I suspect that it is the same for ipvlan and dsa.
I agree about DSA. I don't see any way this could happen.
Andrew
I only keep the ipoib patch and drop the rest patches.
thanks
--
To unsubscribe from this list: send
On Tue, Apr 14, 2015 at 11:20:17PM +0800, Honggang Li wrote:
Signed-off-by: Honggang Li ho...@redhat.com
---
net/dsa/slave.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index 827cda56..070b599 100644
--- a/net/dsa/slave.c
+++
On 14.04, Honggang LI wrote:
On Tue, Apr 14, 2015 at 04:26:27PM +0100, Patrick McHardy wrote:
That is completely useless. vlan (=netdev_priv) can not be NULL as
netdev_priv() never returns NULL and vlan-lowerdev is always valid
because a macvlan wouldn't make much sense otherwise.
OK,
On 04/14/2015 05:40 PM, ira.weiny wrote:
On Tue, Apr 14, 2015 at 04:32:57PM +0200, Michael Wang wrote:
On 04/14/2015 04:18 PM, ira.weiny wrote:
[snip]
/**
- * cap_ib_cm_dev - Check if any port of device has the capability
Infiniband
- * Communication Manager.
+ *
On 14.04, Honggang Li wrote:
Signed-off-by: Honggang Li ho...@redhat.com
---
drivers/net/macvlan.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c
index b227a13..1e59f39 100644
--- a/drivers/net/macvlan.c
+++
On 04/14/2015 05:50 PM, ira.weiny wrote:
On Tue, Apr 14, 2015 at 10:35:34AM +0200, Michael Wang wrote:
On 04/13/2015 09:25 PM, Hefty, Sean wrote:
@@ -1037,17 +1033,13 @@ void rdma_destroy_id(struct rdma_cm_id *id)
mutex_unlock(id_priv-handler_mutex);
if (id_priv-cma_dev) {
-
On Sun, Apr 12, 2015 at 1:51 AM, Or Gerlitz gerlitz...@gmail.com wrote:
On Thu, Apr 9, 2015 at 8:41 PM, Roland Dreier rol...@kernel.org wrote:
On Tue, Apr 7, 2015 at 10:12 AM, Jason Gunthorpe
jguntho...@obsidianresearch.com wrote:
I don't think you understand how deep the problem Or is
On 14/04/2015 15:50, Sagi Grimberg wrote:
On 4/14/2015 3:00 PM, Yann Droneaud wrote:
Le mardi 14 avril 2015 à 12:20 +0300, Sagi Grimberg a écrit :
On 4/13/2015 3:56 PM, Yann Droneaud wrote:
...
diff --git a/drivers/infiniband/core/umem.c
b/drivers/infiniband/core/umem.c
index
On Tue, Apr 14, 2015 at 04:26:27PM +0100, Patrick McHardy wrote:
That is completely useless. vlan (=netdev_priv) can not be NULL as
netdev_priv() never returns NULL and vlan-lowerdev is always valid
because a macvlan wouldn't make much sense otherwise.
OK, please drop this patch.
thanks
--
On Tue, Apr 14, 2015 at 10:35:34AM +0200, Michael Wang wrote:
On 04/13/2015 09:25 PM, Hefty, Sean wrote:
@@ -1037,17 +1033,13 @@ void rdma_destroy_id(struct rdma_cm_id *id)
mutex_unlock(id_priv-handler_mutex);
if (id_priv-cma_dev) {
- switch
to check upstream linux-next fix
the race condition or not, I build linux-next-20150414 on two machines. Both
machines panic at modprobe ib_ipoib. Do you means I need to report a
bug? But I do not know report to who or where.
Here is the tag:
Fixes: 5aa7add8f14b (infiniband/ipoib: implement
No need to pass that, we can take it from the task.
In a later stage, this function will be invoked
according to a device capability.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg sa...@mellanox.com
Signed-off-by: Adir Lev ad...@mellanox.com
---
Instead of open-coding connection fastreg pool get/put,
we introduce iser_reg_desc[get|put] helpers.
We aren't setting these static as this will be a per-device
routine later on. Also, cleanup iser_unreg_rdma_mem_fastreg
a bit.
This patch does not change any functionality.
Signed-off-by: Sagi
Memory regions are resources that are saved
in the device caches. Increase the probability for
a cache hit by adding the MRU descriptor to pool
head.
Signed-off-by: Sagi Grimberg sa...@mellanox.com
---
drivers/infiniband/ulp/iser/iser_memory.c |2 +-
1 files changed, 1 insertions(+), 1
Make iser_[create|destroy]_fastreg_desc shorter, more
readable and easily extendable.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg sa...@mellanox.com
Signed-off-by: Adir Lev ad...@mellanox.com
---
drivers/infiniband/ulp/iser/iser_verbs.c | 118
Hi Roland,
This set contains bug fixes as well as code refactoring patches
- Patches 1-3: Bug fixes (stable material)
- Patches 4-6: Bounce buffer related code cleanups
- Patches 7-12,16: Refactoring
- Patches 13-15: Minor optimizations
- Patch 17:Version bump
- Patch 18:
No need to keep lkey, va, len variables, we can keep
them as struct ib_sge. This will help when we change the
memory registration logic.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg sa...@mellanox.com
Signed-off-by: Adir Lev ad...@mellanox.com
---
As memory registration/de-registration methods, lets
move them to their natural location. While we're at it,
make iser_reg_page_vec routine static.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg sa...@mellanox.com
---
drivers/infiniband/ulp/iser/iscsi_iser.h |4
No need for these two separate. Keep it in a single routine
like in the fastreg case. This will also make iser_reg_page_vec
closer to iser_fast_reg_mr arguments. This is a preparation
step for registration flow refactor.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg
Signed-off-by: Honggang Li ho...@redhat.com
---
drivers/net/ipvlan/ipvlan_main.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ipvlan/ipvlan_main.c b/drivers/net/ipvlan/ipvlan_main.c
index 77b92a0..479fcf7 100644
--- a/drivers/net/ipvlan/ipvlan_main.c
+++
Signed-off-by: Honggang Li ho...@redhat.com
---
drivers/net/macvlan.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c
index b227a13..1e59f39 100644
--- a/drivers/net/macvlan.c
+++ b/drivers/net/macvlan.c
@@ -998,7 +998,9 @@
Starting monitoring for VG vg_rdma01: 3 logical volume(s) in volume
group vg_rdma01 monitored
[ OK ]
Starting cgconfig service: Failed to parse /etc/cgconfig.conf or
/etc/cgconfig.d[FAILED]
Loading OpenIB kernel modules:
BUG: unable to handle kernel NULL pointer dereference at 0120
Signed-off-by: Honggang Li ho...@redhat.com
---
net/dsa/slave.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index 827cda56..070b599 100644
--- a/net/dsa/slave.c
+++ b/net/dsa/slave.c
@@ -59,7 +59,10 @@ static int
On 04/14/15 15:23, Matan Barak wrote:
+mutex_lock(cache-lock);
+
+for (ix = 0; ix cache-sz; ix++)
+if (cache-data_vec[ix].attr.ndev == ndev)
+write_gid(ib_dev, port, cache, ix, zgid, zattr);
+
+mutex_unlock(cache-lock);
+return 0;
The traditional Linux
On Tue, Apr 14, 2015 at 04:32:57PM +0200, Michael Wang wrote:
On 04/14/2015 04:18 PM, ira.weiny wrote:
[snip]
/**
- * cap_ib_cm_dev - Check if any port of device has the capability
Infiniband
- * Communication Manager.
+ * cap_ib_cm_any_port - Check if any port of the device
On Tue, Apr 14, 2015 at 05:35:12PM +0200, Nicolas Dichtel wrote:
Le 14/04/2015 17:26, Patrick McHardy a écrit :
On 14.04, Honggang Li wrote:
[snip]
- return vlan-lowerdev-ifindex;
+ if (vlan vlan-lowerdev)
+ return vlan-lowerdev-ifindex;
That is completely useless. vlan
The four callback functions deferference the pointers without test.
As ipoib_get_iflink failed, it is a good reason to fix the other three
functions. Those functions will return zero on error, as valid ifindex
greater than zero.
I understand that those simple patches should be emerged into a
If CONFIG_ARCH_DMA_ADDR_T_64BIT enabled for x86 systems and physical
memory is more than 4GB, dma_map_page may return a valid memory
address which greater than 0x. As a result, the mlx5 device page
allocator RB tree will be initialized with valid addresses greater than
0xfff.
However,
On Mon, Apr 13, 2015 at 01:14:40PM -0400, Doug Ledford wrote:
On Apr 13, 2015, at 12:42 PM, Jason Gunthorpe
jguntho...@obsidianresearch.com wrote:
On Sun, Apr 12, 2015 at 07:16:33PM +0200, Yann Droneaud wrote:
That would make sense to have someone like Doug or Jason being the
On Tue, Apr 14, 2015 at 10:23:16PM +0300, Eli Cohen wrote:
On Mon, Apr 13, 2015 at 05:21:58PM +0800, Honggang Li wrote:
@@ -241,7 +243,7 @@ static void free_4k(struct mlx5_core_dev *dev, u64 addr)
static int alloc_system_page(struct mlx5_core_dev *dev, u16 func_id)
{
struct page
On 4/15/2015 6:19 AM, Honggang Li wrote:
Fix(bf0bf77 mlx5: Support communicating arbitrary host page size to firmware)
This isn't the way to write the fix note, do it that way
Fixes: bf0bf77f6519 ('mlx5: Support communicating arbitrary host page
size to firmware')
Signed-off-by: Honggang
Acked-by: Eli Cohen e...@mellanox.com
On Wed, Apr 15, 2015 at 11:19:17AM +0800, Honggang Li wrote:
If CONFIG_ARCH_DMA_ADDR_T_64BIT enabled for x86 systems and physical
memory is more than 4GB, dma_map_page may return a valid memory
address which greater than 0x. As a result, the mlx5
On 4/14/2015 6:08 PM, Sagi Grimberg wrote:
Hi Roland,
This set contains bug fixes as well as code refactoring patches
- Patches 1-3: Bug fixes (stable material)
- Patches 4-6: Bounce buffer related code cleanups
- Patches 7-12,16: Refactoring
- Patches 13-15: Minor optimizations
-
On 4/14/2015 11:41 PM, Jason Gunthorpe wrote:
Erez, you basically rewrote this, please make a proper patch with the Fixes and
Reported-By credit for Honggang. Lets merge this through Dave M's tree right
away.
Agree, Erez, add proper Fixes: XXX note and send a patch to netdev
against
On 4/13/2015 8:14 PM, Doug Ledford wrote:
Same here (having been asked privately to look into doing this). FWIW, Red Hat
has agreed that they are willing to allow me to transition my job to a full
time upstream focused position (as of now, I work on upstream around my
deliverables inside the
-Original Message-
From: Hefty, Sean [mailto:sean.he...@intel.com]
Sent: Tuesday, April 14, 2015 11:02 PM
To: Matan Barak; Somnath Kotur; rol...@kernel.org
Cc: linux-rdma@vger.kernel.org
Subject: RE: [PATCH v3 for-next 01/33] IB/core: Add RoCE GID cache
This is a part of the
On Mon, Apr 13, 2015 at 05:21:58PM +0800, Honggang Li wrote:
@@ -241,7 +243,7 @@ static void free_4k(struct mlx5_core_dev *dev, u64 addr)
static int alloc_system_page(struct mlx5_core_dev *dev, u16 func_id)
{
struct page *page;
- u64 addr;
+ u64 addr = 0;
int err;
On Tue, Apr 14, 2015 at 07:30:03PM +0300, Erez Shitrit wrote:
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c
b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 657b89b..11ea6e2 100644
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -846,7 +846,10 @@ static int
On Tue, Apr 14, 2015 at 11:52:11AM +0300, Sagi Grimberg wrote:
On 2/23/2015 6:50 AM, Yuval Shaia wrote:
Current approach force one to implement all ops even when some functions can
use the default implementation.
As a result, for new DMA ops (e.x new arch) many functions just wrap the
On 3/26/2015 1:42 AM, Bart Van Assche wrote:
On 03/25/2015 02:19 PM, Somnath Kotur wrote:
+if (cache-data_vec[ix].attr.ndev
+cache-data_vec[ix].attr.ndev != old_net_dev)
A few lines earlier the memory old_net_dev points at was freed. If two
instances of this function run
On 4/14/2015 3:00 PM, Yann Droneaud wrote:
Hi,
Le mardi 14 avril 2015 à 12:20 +0300, Sagi Grimberg a écrit :
On 4/13/2015 3:56 PM, Yann Droneaud wrote:
In a call to ib_umem_get(), if address is 0x0 and size is
already page aligned, check added in commit 8494057ab5e4
(IB/uverbs: Prevent
90 matches
Mail list logo