On Sun, Nov 15, 2015 at 11:48:08AM -0800, Bart Van Assche wrote:
> Did I understand correctly that page-aligned I/O works fine but I/O that is
> not aligned on a page boundary not ? Have you already had the time to verify
> whether the "IB/srp: Convert to new registration API" patch is the patch
>
On Sun, Nov 15, 2015 at 8:11 PM, Christoph Lameter wrote:
> 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
>
On 15/11/2015 23:10, Or Gerlitz wrote:
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?
Because it's not length 0. It's because I did
On Tue, Oct 20, 2015 at 10:17 PM, Doug Ledford wrote:
> On 10/15/2015 08:01 AM, Matan Barak wrote:
>> When using ifup/ifdown while executing enum_netdev_ipv4_ips,
>> ifa could become invalid and cause use after free error.
>> Fixing it by protecting with RCU lock.
>>
>> Fixes: 03db3a2d81e6 ('IB/co
On Thu, Oct 15, 2015 at 6:07 PM, Matan Barak wrote:
> Hi Doug,
>
> This series adds the support for RoCE v2. In order to support RoCE v2,
> we add gid_type attribute to every GID. When the RoCE GID management
> populates the GID table, it duplicates each GID with all supported types.
> This gives
On 11/15/2015 7:30 AM, 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).
Can you describe what these timestamps are actually for? It's not
clear at all from the comments.
On Mon, Nov 16, 2015 at 5:46 PM, Tom Talpey wrote:
> On 11/15/2015 7:30 AM, 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).
>
>
> Can you describe what these t
From: Jenny Derzhavetz
When all the task data is sent as immeidatedata, we are
allowed to use the local_dma_lkey as it is not sent to
the wire.
Signed-off-by: Jenny Derzhavetz
Signed-off-by: Sagi Grimberg
---
drivers/infiniband/ulp/iser/iscsi_iser.h | 3 ++-
drivers/infiniband/ulp/iser/is
This patchset adds remote invalidation support to iser initiator and
target. The support negotiation for this feature is based on IBTA
annex 12 "Support for iSCSI Extensions for RDMA" carried in rdma_cm
private data.
Remote invalidation allows a peer host to invalidate a remote key
as part of a SE
From: Roi Dayan
destroy workqueue on transport register error
release kmem cache on workqueue alloc error
Signed-off-by: Roi Dayan
Signed-off-by: Sagi Grimberg
---
drivers/infiniband/ulp/iser/iscsi_iser.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers/inf
This is a pre-step before adding remote invalidate
support. Also, every ULP should prefer fastreg over
fmr.
Signed-off-by: Sagi Grimberg
---
drivers/infiniband/ulp/iser/iser_memory.c | 9 -
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/drivers/infiniband/ulp/iser/iser_mem
On 11/15/2015 3:34 AM, Sagi Grimberg wrote:
+
+struct ib_stop_cqe {
+struct ib_cqecqe;
+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);
+
+c
From: Jenny Derzhavetz
iser target does not support zero based virtual addresses and
send with invalidate, so it should declare that it doesn't.
Signed-off-by: Jenny Derzhavetz
Signed-off-by: Sagi Grimberg
---
drivers/infiniband/ulp/isert/ib_isert.c | 6 ++
1 file changed, 6 insertions(+)
With remote invalidate we won't local invalidate
but we still want to increment the rkey.
Signed-off-by: Sagi Grimberg
Signed-off-by: Jenny Derzhavetz
---
drivers/infiniband/ulp/iser/iser_memory.c | 20 ++--
1 file changed, 10 insertions(+), 10 deletions(-)
diff --git a/drivers
From: Jenny Derzhavetz
This parameter is described as "is mr valid indicator".
In other words, it indicates whether memory registration
is valid or not. So intuitive values would be:
mr_valid=True, when memory registration is valid and
mr_valid=False otherwise.
Signed-off-by: Jenny Derzhavetz
S
The iser RDMA_CM negotiation protocol is shared by
the initiator and the target, so have a shared header
for the defines and structure. Move relevant items from
the initiator and target headers.
Signed-off-by: Sagi Grimberg
Signed-off-by: Jenny Derzhavetz
Cc: linux-scsi
---
drivers/infiniband/
From: Jenny Derzhavetz
We'll use remote invalidate, according to negotiation result
during connection establishment. If the initiator declared that
it supports the remote invalidate exception then the target will
use IB_WR_SEND_WITH_INV with the correct rkey for the response.
Signed-off-by: Jenn
From: Jenny Derzhavetz
Declare that we support remote invalidation and be able to detect
the invalidated rkey so we won't invalidate it locally. The spec
mandates that we must not rely on the taget remote invalidate our
rkey so we must check it upon a receive (scsi response) completion.
Signed-o
We don't need iser_proto.h anymore, remove it and
move (non-protocol) declarations to ib_isert.h
Signed-off-by: Sagi Grimberg
Signed-off-by: Jenny Derzhavetz
---
drivers/infiniband/ulp/isert/ib_isert.c| 1 -
drivers/infiniband/ulp/isert/ib_isert.h| 31
drivers/infi
On 11/15/2015 12:05 PM, Christoph Hellwig wrote:
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
For the cxgb3/4 parts:
Reviewed-by: Steve Wise
--
To unsubscribe from this list: send the line "
On Sun, Nov 15, 2015 at 01:30:56PM +0200, Hagay Abramovsky wrote:
> From: Haggai Abramonvsky
>
> +
> +int32_t mlx5_store_uidx(struct mlx5_context *ctx, void *rsc)
> +{
> + int32_t tind;
> + int32_t ret = -1;
You don't need this variable. Just initialize uidx and return uidx
always.
> +
I think Doug is removing amso1100 now so this patch isn't needed.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
On 11/15/2015 12:05 PM, Christoph Hellwig wrote:
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 -
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
Will the user-space drivers posting via uverbs (qib, hfi, rxe) need the
post_send interface?
--
To
On Mon, Nov 16, 2015 at 07:00:06PM +0200, Sagi Grimberg wrote:
>
>> 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
>
> Will the user-space
On 11/15/2015 09:59 AM, 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
Hi Bart,
the code in this area changed enough since 4.3 that it won't easily
apply. But a backport would still be very useful!
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.
On 11/16/2015 09:22 AM, Christoph Hellwig wrote:
the code in this area changed enough since 4.3 that it won't easily
apply. But a backport would still be very useful!
In that case:
Reviewed-by: Bart Van Assche
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the bo
> -Original Message-
> From: Steve Wise [mailto:sw...@opengridcomputing.com]
> Sent: Monday, November 16, 2015 10:38 AM
> To: Sagi Grimberg; Christoph Hellwig; linux-rdma@vger.kernel.org
> Cc: bart.vanass...@sandisk.com; ax...@fb.com; linux-s...@vger.kernel.org;
> linux-ker...@vger.kerne
On 16/11/2015 19:02, Christoph Hellwig wrote:
On Mon, Nov 16, 2015 at 07:00:06PM +0200, Sagi Grimberg wrote:
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: Christ
After looking at the nes driver, I don't see any common way to support drain
w/o some serious driver mods. Since SRP is the only
user, perhaps we can ignore iWARP for this function...
But iser/isert essentially does it too (and I think xprtrdma will have
it soon)...
the modify_qp is invoked
Before we expose any new opcodes to userspace we need to actually
move the defintions of user visible opcodes to a uapi header. Without
that we're doomed to regularly break the ABI.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.
Expecting half of the queue to be empty before reopening netif_queue seems
too high.
With this fix threshold will be 90%.
Suggested-By: Ajaykumar Hotchandani
Signed-off-by: Yuval Shaia
---
drivers/infiniband/ulp/ipoib/ipoib.h |4
drivers/infiniband/ulp/ipoib/ipoib_cm.c |8 ++
> -Original Message-
> From: linux-rdma-ow...@vger.kernel.org
> [mailto:linux-rdma-ow...@vger.kernel.org] On Behalf Of Sagi Grimberg
> Sent: Monday, November 16, 2015 12:38 PM
> To: Steve Wise; 'Christoph Hellwig'; linux-rdma@vger.kernel.org
> Cc: bart.vanass...@sandisk.com; ax...@fb.com
On Sat, Nov 14, 2015 at 12:41:02PM +0530, Sudip Mukherjee wrote:
> On Fri, Nov 13, 2015 at 11:09:22AM -0500, Jubin John wrote:
> > On Fri, Nov 13, 2015 at 06:45:58PM +0530, Sudip Mukherjee wrote:
> > > On Wed, Nov 11, 2015 at 02:33:21AM -0500, Jubin John wrote:
> > > > This patch fixes the checkpat
On Wed, Nov 11, 2015 at 11:22:15AM +0300, Dan Carpenter wrote:
> On Wed, Nov 11, 2015 at 02:33:30AM -0500, Jubin John wrote:
> > @@ -8288,6 +8367,21 @@ static int init_cntrs(struct hfi1_devdata *dd)
> > dd->ndevcntrs++;
> > index++;
> >
From: Dean Luick
Correctly set half-swing for integrated devices. A0 needs all fields set for
CcePcieCtrl. B0 and later only need a few fields set.
Reviewed-by: Stuart Summers
Signed-off-by: Dean Luick
Signed-off-by: Ira Weiny
---
Changes from V1:
Add comments concerning the very l
From: Ira Weiny
It was identified that hfi1_ioctl may sleep with a spin lock held.
This was identified publicly here:
http://www.spinics.net/lists/linux-rdma/msg29926.html
As well as by our internal development.
This series cleans up the code and parameter checks, as well as fixing the
lockin
From: Jubin John
Correct the checks on diag.c with the latest checkpatch
Reviewed-by: Dennis Dalessandro
Reviewed-by: Mike Marciniszyn
Signed-off-by: Jubin John
Signed-off-by: Ira Weiny
---
drivers/staging/rdma/hfi1/diag.c | 81 ++--
1 file changed, 37 in
From: Ira Weiny
physState, linkState, and devState should be phys_state, link_state, and
dev_state
Signed-off-by: Dennis Dalessandro
Signed-off-by: Ira Weiny
---
drivers/staging/rdma/hfi1/diag.c | 24
1 file changed, 12 insertions(+), 12 deletions(-)
diff --git a/dri
From: Dennis Dalessandro
This patch avoids issues while calling into copy from/to user while holding the
lock by only taking the lock when it is absolutely required.
The only commands which require the snoop lock are: *Set Filter *Clear Filter
*Clear Queue
Reviewed-by: Mike Marciniszyn
Signed-
From: Ira Weiny
Set link state is not supported remove from the switch statement and allow the
default to return -ENOTTY
Signed-off-by: Dennis Dalessandro
Signed-off-by: Ira Weiny
---
Changes from v1:
Just remove the check for SETLINKSTATE and allow the switch to default
to -E
From: Dennis Dalessandro
Now that the spinlock is not taken throughout hfi1_ioctl it is safe to return
early rather than setting a variable and falling through the switch.
Reviewed-by: Mike Marciniszyn
Signed-off-by: Dennis Dalessandro
Signed-off-by: Ira Weiny
---
Changes from V1:
Sq
From: Ira Weiny
Rather than have a switch in a large else clause make the parameter checks
return immediately.
Signed-off-by: Dennis Dalessandro
Signed-off-by: Ira Weiny
---
drivers/staging/rdma/hfi1/diag.c | 348 +++
1 file changed, 173 insertions(+), 175
From: Ira Weiny
Final clean up of the if/then/else clause for the parameter checks of
hfi1_ioctl
Signed-off-by: Dennis Dalessandro
Signed-off-by: Ira Weiny
---
Changes from v1:
SETLINKSTATE is removed
drivers/staging/rdma/hfi1/diag.c | 23 +--
1 file changed, 13
Rebased this series on top of latest staging-next branch
Changes in v2:
01/13: Updated commit message with more information about changes in patch
04/13: Updated patch subject with "hfi1" instead of "hfi"
07/13: Refreshed patch based on new hfi patches in staging-next
12/13: Changed logic based on
From: Vennila Megavannan
Removed the RxCtxRHQS counter being dumped into dev_cntrs
Reviewed-by: Dennis Dalessandro
Signed-off-by: Vennila Megavannan
Signed-off-by: Jubin John
---
drivers/staging/rdma/hfi1/chip.c |2 --
drivers/staging/rdma/hfi1/chip.h |1 -
driver
From: Dean Luick
A link downgrade can race with link up. Avoid the race
in two ways. First, by having the downgrade application logic
take the link state mutex for all of its checking. Second, by
waiting for the link to move out of the going up state.
Reviewed-by: Dennis Dalessandro
Signed-off-
This patch fixes the checkpatch issue:
CHECK: Prefer using the BIT macro
Use of BIT macro for HDRQ_INCREMENT in chip.h causes a change in
format specifier for error message in init.c in order to avoid a
build warning.
Reviewed-by: Dean Luick
Reviewed-by: Ira Weiny
Reviewed-by: Mike Marciniszyn
From: Ira Weiny
Rcv bubbles were improperly calculated for HFIs, fix that here.
Reviewed-by: Mike Marciniszyn
Reviewed-by: Arthur Kepner
Signed-off-by: Ira Weiny
Signed-off-by: Jubin John
---
drivers/staging/rdma/hfi1/mad.c | 64 +-
1 files changed, 2 i
Change the default number of krcvqs to number of numa nodes + 1
based on the performance data collected.
Reviewed-by: Mike Marciniszyn
Signed-off-by: Jubin John
---
drivers/staging/rdma/hfi1/chip.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/rdma/hf
From: Mark F. Brown
The krcvqs parameter is displayed incorrectly in sysfs.
The workaround is to set the param type as uint.
Reviewed-by: Mike Marciniszyn
Reviewed-by: Mitko Haralanov
Signed-off-by: Mark F. Brown
Signed-off-by: Jubin John
---
drivers/staging/rdma/hfi1/hfi.h |2 +-
driv
Space between concantenated string elements is more human
readable and fixes the checkpatch issue:
CHECK: Concatenated strings should use spaces between elements
Reviewed-by: Mike Marciniszyn
Signed-off-by: Jubin John
---
drivers/staging/rdma/hfi1/chip.c | 10 +-
drivers/staging/rdm
From: Dean Luick
Read an EFI variable for the device description. Create the
infrastructure for additional variable reads.
Reviewed-by: Easwar Hariharan
Signed-off-by: Dean Luick
Signed-off-by: Jubin John
---
drivers/staging/rdma/hfi1/Makefile |2 +-
drivers/staging/rdma/hfi1/chip.c |
From: Dean Luick
The variable "index" increments the same as dd->ndevcntrs.
Just use the later. Remove uneeded usage of "index" in the
fill loop - it is not used there or later in the function.
Reviewed-by: Dennis Dalessandro
Signed-off-by: Dean Luick
Signed-off-by: Jubin John
---
drivers/s
From: Mike Marciniszyn
The current is_bx() will incorrectly match on other steppings.
is_a0() is removed in favor of is_ax().
Reviewed-by: Mark Debbage
Signed-off-by: Mike Marciniszyn
Signed-off-by: Jubin John
---
drivers/staging/rdma/hfi1/chip.c | 44 ++---
From: Harish Chegondi
This would reduce L2 cache misses on s_sde in the _hfi1_schedule_send
function when invoked from post_send thereby improving performance of
post_send.
Reviewed-by: Mike Marciniszyn
Signed-off-by: Harish Chegondi
Signed-off-by: Jubin John
---
drivers/staging/rdma/hfi1/ve
From: Vennila Megavannan
Added the following per sdma engine stats:
- SendDmaDescFetchedCnt
- software maintained count of SDMA interrupts
(SDmaInt, SDmaIdleInt, SDmaProgressInt)
- software maintained counts of SDMA error cases
Reviewed-by: Dennis Dalessandro
Signed-off-by: Mike
From: Dean Luick
Add a new EPROM partition, adjusting partition placement.
Add EPROM range commands as a supserset of the partition
commands. Remove old partition commands.
Enhance EPROM erase, creating a range function and using the
largest erase (sub) commands when possible.
Reviewed-by: De
On Mon, Nov 16, 2015 at 8:54 PM, Yuval Shaia wrote:
> Expecting half of the queue to be empty before reopening netif_queue seems
> too high.
> With this fix threshold will be 90%.
Is this backed by tests? why not 75% or 25%?
The origin reason for keeping the queue closed till half of the
complet
On 11/16/2015 6:37 PM, Sagi Grimberg wrote:
destroy workqueue on transport register error
release kmem cache on workqueue alloc error
Sagi and Co
Can you make sure to avoid skipping usage of capital letters when
beginning a sentence
and use punctuations: comma/s, period/s?
Or.
--
To unsubs
On 11/16/2015 6:37 PM, Sagi Grimberg wrote:
This is a pre-step before adding remote invalidate support.
Wouldn't that introduce performance regression on ConnectX3 devices?
Or.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger
On 11/16/2015 6:37 PM, Sagi Grimberg wrote:
This parameter is described as "is mr valid indicator".
In other words, it indicates whether memory registration
is valid or not. So intuitive values would be:
mr_valid=True, when memory registration is valid and
mr_valid=False otherwise.
and what was
On 11/16/2015 6:37 PM, Sagi Grimberg wrote:
--- a/drivers/infiniband/ulp/iser/iser_memory.c
+++ b/drivers/infiniband/ulp/iser/iser_memory.c
@@ -250,7 +250,7 @@ iser_reg_dma(struct iser_device *device, struct
iser_data_buf *mem,
struct scatterlist *sg = mem->sg;
reg->sge.lkey = devi
On 11/16/2015 6:37 PM, Sagi Grimberg wrote:
-/* From iscsi_iser.h */
-
-struct iser_hdr {
- u8 flags;
- u8 rsvd[3];
- __be32 write_stag; /* write rkey */
- __be64 write_va;
- __be32 read_stag; /* read rkey */
- __be64 read_va;
-} __packed;
-
-/*C
65 matches
Mail list logo