On 21/05/2013 17:13, Jack Morgenstein wrote:
I just need to verify that the patch can be applied correctly on the upstream
kernel.
The use of RCU (and not spinlock) makes sense from a performance standpoint
in any case. We do NOT want to force mlx4_cq_completion to have a spinlock
which is devic
On 21/05/2013 18:24, Hefty, Sean wrote:
I don't remember this patch at all.
Alex, can you please send Sean this patch
--
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/maj
On Tue, May 21, 2013 at 6:24 PM, Hefty, Sean wrote:
>> One thing seen in the nodes dmesg is a message from an old patch of
>> yours which exists in ofed1.5.3 but didn't hit (or wasn't accepted?)
>> upstream saying "ib_cm: calculated mra timeout 67584 > 8192, decreasing
>> used timeout_ms" does th
On Tue, May 21, 2013 at 1:43 AM, Shawn Bohrer wrote:
> I appologize if I missed it, but did any support for L3/L4 CSUM
> generation get added? Doesn't look like the upstream libibverbs has
> it, and I don't seem to see any patches floating around.
Roland commented that he will make a point relea
On Tue, May 21, 2013 at 1:54 AM, Shawn Bohrer wrote:
> Are there any patches for libibverbs to add
> ibv_create_flow/ibv_destroy_flow? And are there any needed patches
> for libmlx4? I'm building up a stack so we can begin testing this series.
YES there are patches NO I didn't post them here y
On Mon, May 20, 2013 at 7:49 PM, Roland Dreier wrote:
> That's fine, I'll do the releases this week.
cool.
--
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.h
On 24/05/2013 20:43, Liu Ginhann wrote:
Here is version information of our test chassis,
what driver / card are you using?
--
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.or
On 24/05/2013 20:43, Liu Ginhann wrote:
Ibv_reg_mr() call return EFAULT - bad address
Basically, AFAIK, the IB stack should support what you are trying to do,
If you tell from which code e.g in libibverbs/libmlx4 or the kernel this
value is returned, it might be helpful for further assisting
On Sun, May 26, 2013 at 4:52 PM, Liu Ginhann
wrote:
> Will dig into it.Thanks.
does this works if you use get_free_pages in the kernel instead of kmem_cache?
Or.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More
On 29/05/2013 02:10, Roland Dreier wrote:
libibverbs is a library that allows programs to use RDMA "verbs" for
direct access to RDMA (currently InfiniBand and iWARP) hardware from userspace.
Hey, so there's RoCE out there too...
The new stable release, 1.1.7, is available from
libmlx4 rele
On Tue, May 28, 2013 at 8:51 PM, Jeff Squyres (jsquyres)
wrote:
> I ask because, as an MPI guy, I would *love* to see this stuff integrated
> into the kernel and libibverbs.
Hi Jeff,
Have you looked on ODP? see
https://www.openfabrics.org/resources/document-downloads/presentations/doc_downlo
On 30/05/2013 01:56, Jeff Squyres (jsquyres) wrote:
On May 29, 2013, at 4:53 AM, Or Gerlitz wrote:
Have you looked on ODP? see
https://www.openfabrics.org/resources/document-downloads/presentations/doc_download/568-on-demand-paging-for-user-space-networking.html
Is this upstream?
No
Has
iginates
>
> Hank
>
>> -Original Message-
>> From: linux-rdma-ow...@vger.kernel.org [mailto:linux-rdma-
>> ow...@vger.kernel.org] On Behalf Of Or Gerlitz
>> Sent: Monday, May 27, 2013 12:53 AM
>> To: Liu Ginhann
>> Cc: Or Gerlitz; linux-rdma@vger.kernel.org
>&
On 08/05/2013 15:21, Or Gerlitz wrote:
Here's a 2nd batch of iser patches for 3.10, with the hightlight being
a fix to the device removal flow from Roi Dayan. For some reason the race
this patch fixes doesn't hit on IB link layer as of different timings
(e.g more modules that registe
On 04/06/2013 04:24, Jeff Squyres (jsquyres) wrote:
On May 29, 2013, at 1:53 AM, Or Gerlitz wrote:
Have you looked on ODP? see
https://www.openfabrics.org/resources/document-downloads/presentations/doc_download/568-on-demand-paging-for-user-space-networking.html
Is the idea behind ODP that
On 08/06/2013 19:42, Liu Ginhann wrote:
does this works if you use get_free_pages in the kernel instead of
kmem_cache?
I tried get_free_pages, kmalloc, kmem_cache_alloc. None of them work, it failed
with the same error - EFAULT bad pointer.
After code walk through, I believe it failed in ib_u
which is returned to the caller within struct ib_flow and used
later for detaching that flow.
Signed-off-by: Hadar Hen Zion
Signed-off-by: Or Gerlitz
---
drivers/infiniband/hw/mlx4/main.c | 246 +
1 files changed, 246 insertions(+), 0 deletions(-)
diff --git
ATTR_MC_DEFAULT - same as IB_FLOW_ATTR_ALL_DEFAULT but only for
multicast
IB_FLOW_ATTR_SNIFFER - sniffer rule, receive all port traffic
ALL_DEFAULT and MC_DEFAULT rules options are valid only for Ethernet link type.
Signed-off-by: Hadar Hen Zion
Signed-off-by: Or Gerlitz
---
drivers/infin
Hi Roland, all
These patches add Flow Steering support to the kernel IB core, to uverbs and
to the mlx4 IB (verbs) driver along with one patch to uverbs which adds
some code to support extensions.
IB/core: Add receive Flow Steering support
IB/core: Infra-structure to support verbs extension
From: Hadar Hen Zion
Implement ib_uverbs_create_flow and ib_uverbs_destroy_flow to
support flow steering for user space applications.
Signed-off-by: Hadar Hen Zion
Signed-off-by: Or Gerlitz
---
drivers/infiniband/core/uverbs.h |3 +
drivers/infiniband/core/uverbs_cmd.c | 206
differently.
Signed-off-by: Igor Ivanov
Signed-off-by: Hadar Hen Zion
Signed-off-by: Or Gerlitz
---
drivers/infiniband/core/uverbs_main.c | 29 -
include/uapi/rdma/ib_user_verbs.h | 10 ++
2 files changed, 34 insertions(+), 5 deletions(-)
diff --git a
On 13/06/2013 12:53, Sebastian Riemer wrote:
On 12.06.2013 15:24, Bart Van Assche wrote:
The SRP initiator implements host reset by reconnecting to the SRP
target. That means that communication with the target is possible
as soon as host reset finished. Hence skip the host settle delay.
Signed-
From: Moni Shoua
Currently, the mlx4 driver set RoCE (IBoE) gids to encode related
Ethernet netdevice interface MAC address and possibly VLAN id.
Change this scheme such that gids encode interface IP addresses
(both IP4 and IPv6).
Signed-off-by: Moni Shoua
Signed-off-by: Or Gerlitz
Currently, the IB stack (core + drivers) handle RoCE (IBoE) gids as
they encode related Ethernet net-device interface MAC address and
possibly VLAN id.
This series changes RoCE GIDs to encode IP addresses (IPv4 + IPv6)
of the that Ethernet interface, under the following reasoning:
1. There are e
differently.
Signed-off-by: Igor Ivanov
Signed-off-by: Hadar Hen Zion
Signed-off-by: Or Gerlitz
---
drivers/infiniband/core/uverbs_main.c | 29 -
include/uapi/rdma/ib_user_verbs.h | 10 ++
2 files changed, 34 insertions(+), 5 deletions(-)
diff --git a
extend the Infiniband address structures (e.g. ib_ah_attr) with layer 2
address parameters, namely mac and vlan.
Signed-off-by: Moni Shoua
Signed-off-by: Or Gerlitz
---
drivers/infiniband/core/cm.c |3 ++
drivers/infiniband/core/cma.c | 39 ++
drivers
From: Matan Barak
Add support for RoCE (IBoE) IP based addressing extensions towards user space.
Extend INIT_QP_ATTR and QUERY_ROUTE ucma commands.
Extend MODIFY_QP and CREATE_AH uverbs commands.
Signed-off-by: Matan Barak
Signed-off-by: Or Gerlitz
---
drivers/infiniband/core/ucma.c
Jason Gunthorpe wrote:
> Can you talk abit about compatibility please? What happens when nodes
> with this patch are on the same network as nodes without it?
The CM on the passive side would send a reject with the reason being
"invalid gid" so this will not go unnoticed.
> Does this patch remo
On Thu, Jun 13, 2013 at 8:09 PM, Jason Gunthorpe
wrote:
> On Thu, Jun 13, 2013 at 06:01:44PM +0300, Or Gerlitz wrote:
>> From: Matan Barak
>>
>> Add support for RoCE (IBoE) IP based addressing extensions towards
>> user space.
>>
>> Extend INIT_QP_ATTR and
Move the various definitions and mad structures needed for software
implementation of IBA PM agent from the ipath and qib drivers into
a single include file, which in turn could be used by more consumers.
Signed-off-by: Or Gerlitz
---
Roland, this adds a section which slipped away when I
On 7/5/2011 6:49 PM, Mike Marciniszyn wrote:
I just realized your submissions were in my "junk" folder.
I would prefer that the other patch get accepted first.
So I can generate my patch on top of yours, and re-submit,
no problem, just lets see what Roland prefers, please send me pointer to
th
Roland Dreier wrote:
> What's the other patch in question here?
I don't know, waiting to hear from Mike, Mike?
Or.
--
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/majord
Roland Dreier wrote:
I can rereview the patch intoto if this is needed.
If you wouldn't mind, that would be great. I would like to get a solution for
qib in the tree but as I said I don't like stealing reserved bits, and I'm
not confident I know the tradeoffs here.
Roland, MiKe
Reading furthe
Hi Alex,
I just noted that opensm logs can go anywhere in size when one of my
nodes had no space left on device after the opensm log level was raised...
I see that the rpm provided with RHEL6 is setting an
/etc/logrotate.d/opensm
entry which is a copy of ./scripts/opensm.logrotate from the ope
extended device cap flags field instead of a field per
extended capability. Initial patch done by Eli Cohen
Signed-off-by: Or Gerlitz
changes from v0:
fixed sparse warnings
---
drivers/net/mlx4/en_ethtool.c |9 +
drivers/net/mlx4/en_main.c |3 ++-
drivers/net/mlx4
Justin P. Mattock wrote:
Below you will find a set of patches, that update the copyright info from Sun
Microsystems to Oracle.
should make changes.. Please let me know if I missed anything and/or
You have sent these patches last year, and the approach was rejected
(e.g see http://marc.info/?l=
Can you describe the problem that the current code causes? Does it lead to a
kernel crash?
Maybe the issue is that port space ipoib should never be associated with a
non-IB device, with a check added to cma_acquire_dev().
Sean, Moni,
I believe that crashing the kernel from user space shou
On Mon, Jun 20, 2011 at 9:18 PM, Roland Dreier wrote:
> Or Gerlitz wrote:
>> Roland, My patch series went through Sean's review (and passed... patch 7/9
>> needed
>> fixes and I posted V1 for it) any more feedback you need to accept them or
>> you just want to
Roland, the enclosed patch series enhances the user space IB
stack to support IBoE. This work is based on earlier patches
done by Eli Cohen and posted during
August/2010 see some pointers below.
One notable change from the previous post, is keeping the UD related
resolution between GID to Etherne
their IB code path.
The solution was suggested by: Roland Dreier
and Jason Gunthorpe
Signed-off-by: Or Gerlitz
Signed-off-by: Eli Cohen
---
include/infiniband/verbs.h | 21 +
man/ibv_query_port.3 |4
2 files changed, 25 insertions(+), 0 deletions(-)
diff
new
library will not touch that field so it will be resolved as unspecified,
and old library running over new kernel will ignore the setting done by
the kernel.
The solution was suggested by: Roland Dreier
and Jason Gunthorpe
Signed-off-by: Or Gerlitz
Signed-off-by: Eli Cohen
---
include
Add IB GID change event which is generated by the kernel
IBoE stack when the HW driver updates the GID table.
Signed-off-by: Or Gerlitz
Signed-off-by: Eli Cohen
---
examples/asyncwatch.c |2 ++
include/infiniband/verbs.h |3 ++-
man/ibv_get_async_event.3 |2 ++
3 files
Since IBoE requires usage of GRH, update ibv_*_pinpong examples to accept
GIDs. GIDs are given as an index to the local port's table and are exchanged
between the client and the server through the socket connection.
Signed-off-by: Or Gerlitz
Signed-off-by: Eli Cohen
---
examples/devi
Modify libmlx4 to support IBoE, where the only user space piece
to handle is the creation of UD address handles - the L2 Ethernet
attributes have to be resolved from the DGID.
Derived from work by Eli Cohen
Signed-off-by: Or Gerlitz
---
src/mlx4.h |1 +
src/qp.c|1 +
src/verbs.c
Add VLANs support for the UD address handle creation flow,
where the vlan id is taken from the destination GID and the
and vlan priority from the IB SL specified by the application.
Signed-off-by: Or Gerlitz
---
src/mlx4.h |1 +
src/qp.c|1 +
src/verbs.c |9 +++--
src/wqe.h
align the list of ConnectX devices supported by the library to be
the same as the mlx4 driver from the upstream kernel
Signed-off-by: Or Gerlitz
---
Roland, these two simple awk/cut commands can be used to
actually validate the claim made by the change log...
$ grep MELLANOX libmlx4.git/src
Jason Gunthorpe wrote:
>> Modify libmlx4 to support IBoE, where the only user space piece
>> to handle is the creation of UD address handles - the L2 Ethernet
>> attributes have to be resolved from the DGID.
> Other than rejecting some GIDs it doesn't seem like there is any
> operational need to
Jason Gunthorpe wrote:
> What zeros the field in the current library?
the kernel uverbs code either zeros it (old kernels) or sets it (new kernels)
Or.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo
Jason Gunthorpe wrote:
> You also need to update ibv_event_type_str
sure, will do that.
> Has a kernel patch that reserves/uses the corresponding ID been accepted yet?
yes, its in Roland's for-next branch
> BTW, this has value for straight IB, hopefully it triggers on IB
> when the GUID table
Jason Gunthorpe wrote:
>> Since IBoE requires usage of GRH, update ibv_*_pinpong examples to accept
>> GIDs. GIDs are given as an index to the local port's table and are exchanged
>> between the client and the server through the socket connection.
> Since these examples are meant as programming e
Jason Gunthorpe wrote:
> Or Gerlitz wrote:
>> Jason, as the change log explains, under Ethernet link type the code
>> (see mlx4_resolve_grh_to_l2) goes and extracts the Ethernet L2 info
>> (mac and vlan, look on the last patch) from the GID.
>
> Looking at the cont
Jason Gunthorpe wrote:
> Or Gerlitz wrote:
>>> What zeros the field in the current library?
+ port_attr->link_layer = IBV_LINK_LAYER_UNSPECIFIED;
+ port_attr->pad = 0;
Or
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the
Jason Gunthorpe wrote:
> The commit comment states the goal is to have new apps see
> IBV_LINK_LAYER_UNSPECIFIED for all possible combinations of old stuff
> underneath.
> [...] The problem is with the userspace ABI out of libibverbs. Current
> libibverbs does not zero its ib_port_attr structur
Jason Gunthorpe wrote:
> Or Gerlitz wrote:
>> Jason Gunthorpe wrote:
>> +#define ibv_query_port(context, port_num, port_attr) \
>> + ___ibv_query_port(context, port_num, port_attr)
>> +
>> So we are okay now, correct?
> Looks like! Sorry for the confus
Jason Gunthorpe wrote:
> Yes, that was the only important thing I saw.
> I assume Roland was happy with the #define trick for compatibility
> last time it came up?
yes, that's correct.
> Looking at the diff, maybe re-shuffle the order of the #define trick
> so it is easier to follow, and call i
their IB code path.
The solution was suggested by: Roland Dreier
and Jason Gunthorpe
Signed-off-by: Or Gerlitz
Signed-off-by: Eli Cohen
---
changes from V0:
- applied feedback from Jason, made the ___ibv_query_port trick
clearer to header file readers
include/infiniband/verbs.h | 23
Add IB GID change event which is generated by the kernel
IBoE stack when the HW driver updates the GID table.
Signed-off-by: Or Gerlitz
Signed-off-by: Eli Cohen
---
changes from V0:
- applied feedback from Jason, added entry to ibv_event_type_str
examples/asyncwatch.c |2
Or Gerlitz wrote:
> the enclosed patch series enhances the user space IB stack to support IBoE.
> [PATCH 1/7] libibverbs: Add link layer field port attribute
> [PATCH 2/7] libibverbs: change kernel API to accept link layer
> [PATCH 3/7] libibverbs: add GID change event
> [PATCH
This resolves a crash in iser_sg_to_page_vec on this system.
Signed-off-by: Or Gerlitz
---
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h
b/drivers/infiniband/ulp/iser/iscsi_iser.h
index 342cbc1..db6f3ce 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp
RFC3270 mandates that iSCSI PDUs are padded to the closest integer
number of four byte words. Enhance the iser code to support that
on both the TX/RX flows.
Signed-off-by: Or Gerlitz
---
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c
b/drivers/infiniband/ulp/iser/iscsi_iser.c
index
On Wed, Jul 20, 2011 at 10:43 PM, Or Gerlitz wrote:
>> the enclosed patch series enhances the user space IB stack to support IBoE.
>> [PATCH 1/7] libibverbs: Add link layer field port attribute
>> [PATCH 2/7] libibverbs: change kernel API to accept link layer
>> [PATCH
> All the IBoE userspace patches should be applied and pushed out.
> Let me know if I missed anything.
Yes, thanks, I see them now, I just didn't notice an explicit ack from
you over the list which made me think there's not there yet, sure,
I'll test with clones of your trees and let you know.
Or
On 9/15/2011 1:01 AM, Roland Dreier wrote:
Sorry, haven't been very productive, partially due to patchwork being
down (along with the rest of kernel.org). So not much to report...
Roland, Sean was saying (SB) that he has them on the ofa git, so in case
you have
a local clone, maybe you can st
Hal Rosenstock wrote:
> Hefty, Sean wrote:
>> That's was an issue with the wire protocol format, correct?
> Right.
>> Why carry that same concept into a software interface? It seems complex.
> So as not to break existing verbs apps by inventing new values to
> existing fields which wouldn't
On 9/20/2011 10:53 AM, Roland Dreier wrote:
Fix a leak of entry if radix_tree_insert() fails.
Also, reduce the indentation and make the flow easier to read
Looks fine to me
Acked-by: Or Gerlitz
Yevgeny?
Or.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma&qu
On Wed, Oct 5, 2011 at 8:40 PM, Hefty, Sean wrote:
>> The following extended speeds are introduced:
>> FDR-10 - is a proprietary link speed which is 10.3125 Gbps at 64/66
>> encoding rather
>> FDR - represents the IBA extended speed: 14.0625 Gbps.
>> EDR - represents the IBA extended spe
On Sat, Oct 8, 2011 at 10:40 PM, Or Gerlitz wrote:
> I was under the misconception that the top (Linus) mainline kernel is
> under hold as of the kernel.org being down, upto two days ago when I
> noted that the first github hit for Linux is Linus tree which is @ -rc9,
> anyway,
Roland,
This is a batch of IBoE related fixes plus addition of extended PMA counters.
Or.
Or Gerlitz (5):
ib/mlx4: enable 4K MTU for IBoE
ib/mlx4: remove setting of vlan in IBoE WQEs control segment
net/mlx4_core: remove module param controlling the vlan table size
ib
The IBoE port mtu is derived from the corresponding EN netdevice
mtu, which can go upto supporting Jumbo-Frames of 9KB and hence
surely supports the max IB mtu of 4K
Signed-off-by: Or Gerlitz
---
drivers/infiniband/hw/mlx4/main.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff
There's no need to set the vlan related fields in an IBoE send WQE
control segment. This b/c the vlan to be used by a UD QP is specified
through the the datagram segment, and with GSI (CM) QP, all the headers
down to the 8021q and mac ones are built by the software anyway.
Signed-off-b
Enable the maximum size (128) supported by the device for the shadow
vlans table, without a module param to override it. This table is only
used by the IBoE control plane for setting a vlan index into an RC/UC
QP context or UD Address Handle.
Signed-off-by: Or Gerlitz
---
drivers/net/mlx4
issue a PMA classportinfo query towards the HW driver, and if extended
counters are supported, expose new sysfs entries which allow to read them.
Signed-off-by: Vu Pham
Signed-off-by: Or Gerlitz
---
---
drivers/infiniband/core/sysfs.c | 100 ++-
include
With the HW counters being 64bit ones, use the existing IBoE
PMA counters to support also extended counters.
Signed-off-by: Vu Pham
Signed-off-by: Or Gerlitz
---
drivers/infiniband/hw/mlx4/mad.c | 49 -
1 files changed, 42 insertions(+), 7 deletions
On 10/10/2011 7:00 PM, Roland Dreier wrote:
Can we really do this? Doesn't this break anyone who has this option
configured in modprobe.conf?
This is one reason why I hate adding module parameters -- you can't
just delete them later like this
Yes, I now understand even better why you hate ad
On Tue, Oct 11, 2011 at 7:28 PM, Roland Dreier wrote:
> No, I think we need to switch to ignoring the parameter and printing a
> warning if someone sets it, and then we can remove it in a year or so.
no problem, I don't think I will be able to send such patch before
Sunday, so in spite of the mer
Enable the maximum size (128) supported by the device for the shadow
vlans table, igonring the module param which could override it.
This table is only used by the IBoE control plane for setting a vlan
index into an RC/UC QP context or UD Address Handle.
Signed-off-by: Or Gerlitz
---
changes
On 10/11/2011 7:28 PM, Roland Dreier wrote:
No, I think we need to switch to ignoring the parameter and printing a
warning if someone sets it, and then we can remove it in a year or so
done, I sent you V1 of patch 3/5 which fixes that, any other comment
for the rest of the series?
Or.
--
T
On 10/14/2011 10:22 PM, Hefty, Sean wrote:
Just an update: the issues that I was seeing were caused by missing
patches in my libraries (one in libmlx4 and the other in the ofed
compatibility layer). The XRC patches in for-next are testing out fine
for me, though it would be good if someone from
Hefty, Sean wrote:
> The rdma_xclient / rdma_xserver tests are a place for those. They support RC
>and
> XRC, so they're in the master branch for the RC support. I have extensions
> for them in
> a private branch which aren't quite ready yet.
again, and just to make sure I got it - for basic
On Fri, Oct 14, 2011 at 4:02 PM, Christoph Lameter wrote:
> Running ibv_ud_pingpong and ibc_uc_pingpong between two hosts. One with
> OFED 1.5.3.1 (Ubuntu LTS 10.04) and another on linux 3.1.0-rc9 (Same
> ubuntu version uderlying) with the upstream libraries.
> ibv_ud_pingpong
> OFED:
> # ibv_ud_p
> On Tue, Oct 18, 2011 at 5:14 PM, Christoph Lameter wrote:
> -s 100 fixes the issue.
Good to know! I'll take a look into whether/how we could improve this
little but annoying incompatibility...
Or.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a messa
On Wed, Oct 19, 2011 at 12:52 AM, Roland Dreier wrote:
> Do you understand why -s 100 fixes things? If so can you explain it to me? ;)
Oh sorry if I was a bit cryptic here (BTW 99 and 101 and 256,512,2048 would
work as well...), I'm not next to the code and writing from backoffice my
head cache
Roland,
I just recalled that once the 3.2 merge window opens Dave is going to
push that drivers/net separation change done by Jeff K. which moves
drivers/net/mlx4 to drivers/net/ethernet/mellanox/mlx4 see
https://github.com/davem330/net-next/tree/master/drivers/net/ethernet/mellanox/mlx4
- is this
Roland Dreier wrote:
> I'm sure we can work it out.
Thanks for the clarification...are you looking on my IBoE fixes for 3.2?
Or.
--
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.kern
Roland Dreier wrote:
> Is there any reason mlx4_get_ext_port_caps() couldn't be more like
> mlx4_check_ext_port_caps(), and have that function set
> MLX4_FLAG_EXTENDED_PORT_INFO? Is it possible that
> only some ports on the same HCA handle the extended port info query?
> If so is there some reaso
On 10/19/2011 1:05 AM, Or Gerlitz wrote:
in few hours I will be next to the code...
The theory failed, looking on commit
577818b50990853bf112ba2a2c336020dd278622 "Update examples for IBoE"
it doesn't change the default message size for the ud pingpong, so we
have to dig de
Kelly Burkhart wrote:
>> I'm having a problem connecting using RDMA_CM over ethernet
> Actually not. My network admin found some sort of misconfiguration on
> the switch that caused the problem Everything works now.
Interesting, could you share the nature of problem?
Or.
--
To unsubscrib
On 10/10/2011 10:51 AM, Or Gerlitz wrote:
This is a batch of IBoE related fixes plus addition of extended PMA counters.
Hi Roland,
With the merge window opening around yesterday, just checking if we're
on track
for merging this patch set into 3.2 (I sent V1 for patch #3 as you asked).
Hi Sean,
Looking on that case, I noted that the CM code (SB) checks that the GID
in the incoming
REQ is present in at least of one the ports of the relevant device, but
not specifically on
the port this request arrived to, is that following IBTA? I thought it
could be problematic
e.g in the ca
On 10/27/2011 6:55 PM, Jason Gunthorpe wrote:
You can use the CM to establish a connection on another port, eg send
CM GMPs to port 1, specify the primary data path is port 2 and specify
the alternate data path is port 1. Jason
makes sense, still, maybe the code here could/should be made more
On Mon, Oct 31, 2011 at 9:08 AM, Vlad Weinbaum
wrote:
> [...] I found detail that I cannot explain. I query the QP after connect and
> get timeout value 16,
> that must be 4 us * 2^16 = 256 ms, but I get about 800 ms.
As Sean indicated, the timeout is **based** on the packet_life_time,
in case
On Mon, Oct 31, 2011 at 9:38 PM, Roland Dreier wrote:
> Sorry for the late review here
Oh yes... BTW this is patch 4/5, I don't see patches 1,2,3 on your for-next
tree/branch @ kernel.org, have you accepted them?
> Sorry for the late review here, but does it seem like the best
> approach to ha
Jason Gunthorpe wrote:
> Is there any reason to expose the 32 and 64 bit version of the same
> counter? That seems needless. Emit the largest version available and
> prepend 0's to fill out to the available width so that userspace can
> know the counter size.
Basically, the approach you suggest s
Jason Gunthorpe wrote:
Guys (Roland, Jason), I'm open to any comments, any time, for any
patch, but for a patch which was posted weeks ago it's pretty unfair
to have your comments coming only eight days after the merge window
has been opened, lets try to come quickly to decision so I can fix
this
On Tue, Nov 1, 2011 at 11:42 PM, Roland Dreier wrote:
> The least bad way forward does seem like it is probably
> the separate new directory thing.
I agree
Or.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More maj
Jason Gunthorpe wrote:
> I don't mean the 32 bit counters are useless, I mean exposing PMA
> counters that saturate and can be randomly reset by external agents
> through sysfs is useless. You can't make any kind of data collection
> based on such a system.
> Ideally the sysfs counters are all n
On Tue, Nov 1, 2011 at 11:49 PM, Roland Dreier wrote:
> There's no obligation to merge something just because you posted it before
> the merge window, and in fact Linus's complaint at the kernel summit is
> always that sub-maintainers don't say no enough.
>
> And let's be honest in this specific c
Jason Gunthorpe wrote:
> Why have a sysfs counter at all when you can just ask the PMA and get exactly
> the
> same data?
The HW/FW PMA agent isn't supported for IBoE only for IB, the
counters are for both
Or.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body
On 11/2/2011 12:46 AM, Ira Weiny wrote:
What do you mean "for both"? Ira
The sysfs PMA counters are functional for both IB and IBoE, the latter uses
a HW counter allocated per device/port for which all the QPs created on
that
port are reporting their rx/tx bytes/packets, see commits
cfcde11c3
On 11/1/2011 11:58 PM, Jason Gunthorpe wrote:
maybe you should patch to un-export them until things can be fixed
sanely...
Jason,
I don't think we want to work this way, bug are either opened or
attempted to be fixed or fixed.
You can't just come out of the the blue and remove functionality
801 - 900 of 1832 matches
Mail list logo