On Sun, Nov 15, 2015, Sagi Grimberg wrote:
> On 15/11/2015 19:59, Christoph Hellwig wrote:
>> Without this sg_dma_len will return 0 on architectures tha have
>> the dma_length field.
and what wrong with that?
Christoph, probably typo here? "tha" needs to be "that"
>> Signed-off-by: Christoph H
From: Markus Elfring
Date: Sun, 15 Nov 2015 21:58:45 +0100
The kmem_cache_destroy() function tests whether its argument is NULL
and then returns immediately. Thus the test around the calls is not needed.
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring
--
On 11/15/15 10:06, Christoph Hellwig wrote:
FYI, I sent a patch for the zero S/G length issue. With this xfstests
does fine for ext4 and btrfs. With XFS I still run into corruption
warnings for the slab use after free poison pattern. I suspect that
issue might be related to uniqueue XFS I/O pa
On 15/11/2015 19:59, Christoph Hellwig wrote:
Without this sg_dma_len will return 0 on architectures tha have
the dma_length field.
Signed-off-by: Christoph Hellwig
---
drivers/infiniband/ulp/srp/ib_srp.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/infiniband/ulp/srp/ib_
On Sun, 15 Nov 2015, Matan Barak wrote:
> This series adds support for completion timestamp. In order to
> support this feature, several extended verbs were implemented
> (as instructed in libibverbs).
This is the portion that
implements timestaping for libmlx5 and this patchset depends on anothe
Fold simplified versions of build_phys_page_list and
iwch_register_phys_mem into iwch_get_dma_wr now that no other callers
are left.
Signed-off-by: Christoph Hellwig
---
drivers/infiniband/hw/cxgb3/iwch_mem.c | 71 ---
drivers/infiniband/hw/cxgb3/iwch_provider.c | 75
Just pass and address/size pair instead of an ib_phys_buf array.
Signed-off-by: Christoph Hellwig
---
drivers/infiniband/hw/nes/nes_cm.c| 10 +--
drivers/infiniband/hw/nes/nes_verbs.c | 140 --
drivers/infiniband/hw/nes/nes_verbs.h | 3 +-
3 files changed,
We have stopped using phys MRs in the kernel a while ago, so let's
remove all the cruft used to implement them.
Signed-off-by: Christoph Hellwig
---
drivers/infiniband/hw/cxgb3/iwch_mem.c | 31 ---
drivers/infiniband/hw/cxgb3/iwch_provider.c | 69 --
drivers/infiniband/hw/cxgb3/iwch
This functionality has no users and was only supported by the staged out
EHCA driver.
Signed-off-by: Christoph Hellwig
---
drivers/infiniband/core/verbs.c | 7 -
drivers/staging/rdma/ehca/ehca_iverbs.h | 2 --
drivers/staging/rdma/ehca/ehca_main.c | 1 -
drivers/staging/rdma/ehc
Signed-off-by: Christoph Hellwig
---
drivers/staging/rdma/amso1100/c2_provider.c | 67 ++---
1 file changed, 12 insertions(+), 55 deletions(-)
diff --git a/drivers/staging/rdma/amso1100/c2_provider.c
b/drivers/staging/rdma/amso1100/c2_provider.c
index 6c3e9cc..2de5a8f 10
Remove the unused ib_allow_mw and ib_bind_mw functions, remove the
unused IB_WR_BIND_MW and IB_WC_BIND_MW opcodes and move ib_dealloc_mw
into the uverbs module.
Signed-off-by: Christoph Hellwig
---
Documentation/infiniband/core_locking.txt | 2 -
drivers/infiniband/core/uverbs.h|
Just IB_DEVICE_LOCAL_DMA_LKEY and IB_DEVICE_MEM_MGT_EXTENSIONS for now
as I'm most familar with those.
Signed-off-by: Christoph Hellwig
Reviewed-by: Sagi Grimberg
Reviewed-By: Jason Gunthorpe
---
include/rdma/ib_verbs.h | 18 ++
1 file changed, 18 insertions(+)
diff --git a/in
This series removes huge chunks of code related to old memory
registration methods that we don't use anymore.
This expects my "IB: merge struct ib_device_attr into struct ib_device"
patch to be already applied.
Also available as a git tree:
http://git.infradead.org/users/hch/rdma.git/sh
Signed-off-by: Christoph Hellwig
---
include/rdma/ib_verbs.h | 5 -
1 file changed, 5 deletions(-)
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 0bbbd4b..3253563 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -1145,11 +1145,6 @@ enum ib_access_fl
And simplify the calling convention for full-memory registrations.
Signed-off-by: Christoph Hellwig
---
drivers/staging/rdma/ehca/ehca_classes.h | 5 +-
drivers/staging/rdma/ehca/ehca_mrmw.c| 94 +++-
2 files changed, 46 insertions(+), 53 deletions(-)
diff --git
FYI, I sent a patch for the zero S/G length issue. With this xfstests
does fine for ext4 and btrfs. With XFS I still run into corruption
warnings for the slab use after free poison pattern. I suspect that
issue might be related to uniqueue XFS I/O patterns. One thing that
might be related is th
Without this sg_dma_len will return 0 on architectures tha have
the dma_length field.
Signed-off-by: Christoph Hellwig
---
drivers/infiniband/ulp/srp/ib_srp.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c
b/drivers/infiniband/ulp/srp/ib_srp.c
index 3
On 15/11/2015 14:55, Christoph Hellwig wrote:
On Sun, Nov 15, 2015 at 11:40:02AM +0200, Sagi Grimberg wrote:
I doubt INT_MAX is useful as a budget in any use-case. it can easily
hog the CPU. If the consumer is given access to poll a CQ, it must be
able to provide some way to budget it. Why not
On 15/11/2015 11:04, Or Gerlitz wrote:
On Sun, Nov 15, 2015 at 10:48 AM, Sagi Grimberg
wrote:
Or is correct,
I have attempted to convert iser to use blk_iopoll in the past, however
I've seen inconsistent performance and latency skews (comparing to
tasklets iser is using today). This was mani
On Sun, Nov 15, 2015 at 11:40:02AM +0200, Sagi Grimberg wrote:
> I doubt INT_MAX is useful as a budget in any use-case. it can easily
> hog the CPU. If the consumer is given access to poll a CQ, it must be
> able to provide some way to budget it. Why not expose a budget argument
> to the consumer?
On Sun, Nov 15, 2015 at 10:48:41AM +0200, Sagi Grimberg wrote:
> I have attempted to convert iser to use blk_iopoll in the past, however
> I've seen inconsistent performance and latency skews (comparing to
> tasklets iser is using today). This was manifested in IOPs test cases
> where I ran multipl
Extended poll_cq supports writing only user's required work completion
fields. Adding support for this extended verb.
Signed-off-by: Matan Barak
---
src/cq.c | 699 +
src/mlx5.c | 5 +
src/mlx5.h | 14 ++
3 files changed, 584 inser
Add support for filling the timestamp field in ibv_poll_cq_ex
(if it's required by the user).
Signed-off-by: Matan Barak
---
src/cq.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/src/cq.c b/src/cq.c
index 0185696..5e06990 100644
--- a/src/cq.c
+++ b/src/cq.c
@@ -913,6 +913,11 @@ inli
Always inline isn't supported by every compiler. Adding it to
configure.ac in order to support it only when possible.
Inline other poll_one data path functions in order to eliminate
"ifs".
Signed-off-by: Matan Barak
---
configure.ac | 17 +
src/cq.c | 42 +
The current ibv_poll_cq_ex mechanism needs to query every field
for its existence. In order to avoid this penalty at runtime,
add optimized functions for special cases.
Signed-off-by: Matan Barak
---
src/cq.c| 363 +---
src/mlx5.h | 1
In order to create a CQ which supports timestamp, the user needs
to specify the timestamp flag for ibv_create_cq_ex.
Adding support for ibv_create_cq_ex in the mlx5's vendor library.
Signed-off-by: Matan Barak
---
src/mlx5.c | 1 +
src/mlx5.h | 2 ++
src/verbs.c | 68
In order to query the current HCA's core clock, libmlx5 should
support ibv_query_values verb. Querying the hardware's cycles
register is done by mmaping this register to user-space.
Therefore, when libmlx5 initializes we mmap the cycles register.
This assumes the machine's architecture places the P
Hi Eli,
This series adds support for completion timestamp. In order to
support this feature, several extended verbs were implemented
(as instructed in libibverbs).
ibv_query_device_ex was extended to support reading the
hca_core_clock and timestamp mask. The same verb was extended
with vendor dep
Add timestamp support for query_device extended verb.
This is necessary in order to support hca_core_clock and
timestamp_mask. In addition, hca_core_clock_offset is added
to the vendor specific part in order to map the cycles register
correctly.
Signed-off-by: Matan Barak
---
src/mlx5-abi.h | 9
From: Haggai Abramovsky
This series adds support for CQE (cookie) versions in libmlx5
Non-default cookie versions are a must for RAW PACKET Eth QPs.
The code is capable to work over kernels which don't have this
support yet for the sake of creating other QP types.
Haggai
Changes from v0:
From: Haggai Abramonvsky
Add new struct member, mlx5_resource, for each tracked object, mlx5_srq
and mlx5_qp, to allow retrieving the object in the poll_cq once we need
it.
Change-Id: Id7d1356cc866fa0aa06d256df9f3612dc1988e6a
Signed-off-by: Haggai Abramovsky
---
src/mlx5.h | 18 ++
From: Haggai Abramonvsky
>From now on, if the kernel supports CQE version 1, the library will
choose to work with it.
Change-Id: I22aff64a5f98dde8cb1b6d771424df52ac5d682b
Signed-off-by: Haggai Abramovsky
---
src/mlx5.c | 3 +++
src/verbs.c | 9 +++--
2 files changed, 6 insertions(+), 6 de
From: Haggai Abramonvsky
When wokring with CQE version 1, the library allocates a user-index for
each new QP/XSRQ, and this user-index is passed to the kernel.
Also in the destruction of a QP/XSRQ, the library needs to free the
user-index, so it can be reused.
In this stage, the library still d
From: Haggai Abramonvsky
When working with CQE version 1, the CQE format is different, therefore
we need a new poll_cq function that is compatible with the new CQE
format.
When allocating the user-context, we decide on the CQE version, and
according to this, we choose which poll_cq function t
From: Haggai Abramonvsky
Add new database that stores all the QPs and XSRQs context. Insertions
and deletions to the database are done using the object's user-index.
This database will allow us to retrieve the objects; QPs and XSRQs; by
their user-index in the poll_one.
Change-Id: I3b87e79be35c
+/**
+ * ib_process_direct_cq - process a CQ in caller context
+ * @cq:CQ to process
+ *
+ * This function is used to process all outstanding CQ entries on a
+ * %IB_POLL_DIRECT CQ. It does not offload CQ processing to a different
+ * context and does not ask from completion in
+
+struct ib_stop_cqe {
+ struct ib_cqe cqe;
+ struct completion done;
+};
+
+static void ib_stop_done(struct ib_cq *cq, struct ib_wc *wc)
+{
+ struct ib_stop_cqe *stop =
+ container_of(wc->wr_cqe, struct ib_stop_cqe, cqe);
+
+ complete(&stop->done);
+}
+
On 11/13/2015 3:46 PM, Christoph Hellwig wrote:
From: Sagi Grimberg
Care to sparse some text here to assist a reviewer and future bisections?!
I have asked multiple times to avoid empty change-logs for patches in
this driver.
Signed-off-by: Sagi Grimberg
Signed-off-by: Christoph Hellwig
-
On 11/13/2015 3:46 PM, Christoph Hellwig wrote:
From: Sagi Grimberg
Makes better sense and we'll need it later with CQ abstraction.
iser switch login bufs to void
Sagi, few quick comments on this patch, please address for next version..
The 2nd sentence of the change-log needs better phrasing
On Sun, Nov 15, 2015 at 10:48 AM, Sagi Grimberg
wrote:
> Or is correct,
>
> I have attempted to convert iser to use blk_iopoll in the past, however
> I've seen inconsistent performance and latency skews (comparing to
> tasklets iser is using today). This was manifested in IOPs test cases
> where I
On Fri, Nov 13, 2015 at 3:46 PM, Christoph Hellwig wrote:
The new name is irq_poll as iopoll is already taken. Better suggestions
welcome.
Sagi (or Christoph if you can address that),
@ some pointer over the last 18 months there was a port done at
mellanox for iser to use blk-iopoll and AF
Hello Hal,
With which SRP target has this behavior been observed ? Has this patch
been tested with the LIO SRP target ?
Hi Bart,
This issue was detected when testing a new array with SRP support.
This does not involve LIO as the Linux CM stack does not behave
in the way described in this pat
42 matches
Mail list logo
