On 11/8/2011 3:09 AM, Jason Gunthorpe wrote:
Roland Dreier wrote:
Let's make sure we learn from our mistakes. Let's say we create a
new ext_counters directory. What should the format of those files
be? Should they be assumed to be 64-bit quantities? Do we want to
allow some way of
Jason Gunthorpe jguntho...@obsidianresearch.com wrote:
The netdev counters are all the same size and there is some other way
to discover what the size is. I'd like to see that for IB counters
too, but it is probably infeasible. So if we have counters that are
not the same size as netdev
Jason Gunthorpe jguntho...@obsidianresearch.com wrote:
We're talking now only on the IB extended counters who are all 64 bits
netdev counters are 32 bit or 64 bit, depending on how the kernel was
compiled. I think indicating the size explicitly, or always being 64
bit (and extending all the
Sebastian Riemer sebastian.rie...@profitbricks.com wrote:
2011/12/20 Or Gerlitz ogerl...@mellanox.com:
Beep(2), so your system has distro which is based on kernel 2.6.32 and iscsi
initiator tools version 2.0.871 and per your needs, you've booted it with
kernel 3.0 .
At this point should you
On 12/21/2011 10:01 AM, Sebastian Riemer wrote:
could you provide quick pointers / 1-2 examples for such API/ABI changes
which aren't deployed in the upstream iser code?
you wrote long emails, I'm asking for one concrete example for that enum
crunching of adding entries
not at the end,
I tested the upstream kernel iser against the upstream iscsi tools from
git://github.com/mikechristie/open-iscsi
(commit 4323e342d2c9fb8ed7233ce855001c189ec55b23), it works
Or.
--
To unsubscribe from this list: send the line unsubscribe linux-rdma in
the body of a message to
On 12/22/2011 3:25 AM, Roland Dreier wrote:
[...] The git shortlog since libibverbs 1.1.5 is: [...]
Roland Dreier (5):
Package description shouldn't start with an article
Don't ship .la files in Debian libibverbs-dev
Update Debian Standards-Version to 3.9.2
Debian:
On 11/8/2011 2:54 AM, Roland Dreier wrote:
Let's make sure we learn from our mistakes. Let's say we create a new
ext_counters directory. What should the format of those files be?
Should they be assumed to be 64-bit quantities? Do we want to allow
some way of indicating the number of bits (ie
or call an existing function with some new enum value
5. what happens if we just want to enhance an -- existing -- function
- suppose we want to enhance ibv_post_send / ibv_poll_cq to support
features like LSO, checksum offload, masked atomic operations, fast
memory remote invalidate,
On 1/2/2012 7:34 PM, Hefty, Sean wrote:
this field is a -- union -- how would that work if more than one
extension is to be applied for a structure?
The fields at the end of the structure should only be accessed if the structure
is of the correct type. In this case, ext.xrc_recv is only
On Sun, Dec 11, 2011 at 4:38 PM, Or Gerlitz ogerl...@mellanox.com wrote:
This series deals with fixes related to Ethernet 8021q vlan
priority-bits used by mlx4 code under IBoE. It is made of
one kernel/mlx4_ib patch and three user-space/libmlx4 patches.
Hi Roland,
Just a reminder... did you
On Tue, Dec 20, 2011 Jason Gunthorpe jguntho...@obsidianresearch.com wrote:
On Tue, Dec 20, 2011 Or Gerlitz wrote:
Jason Gunthorpe jguntho...@obsidianresearch.com wrote:
The netdev counters are all the same size and there is some other way
to discover what the size is. I'd like to see
Roland Dreier rol...@purestorage.com wrote:
It is possible with ConnectX (cf MLX QPs in the kernel driver). However I
don't
know what documentation is available and some hacking would be needed to use
this for
something more general than sending MADs on special QPs.
Greg, RAW Ethernet QPs
On 1/4/2012 11:23 PM, Greg I Kerr wrote:
Yes I want to do this with IB. I wasn't aware that I could use a UD QP
to write my own packets to the wire. Is this what you're suggesting I
can use a UD QP for?
You'll have to further clarify the my own packets part of your
sentence... IB have some
, patches 3 4 need you to pull Linus tree, but this should
be done now anyway for any future patches to the IB stack through out the 3.3
cycle
and elsewhere
Or Gerlitz (6):
IB: use central enum for speed instead of hard-coded values
IB/core: fix wrong display of rate in sysfs
IB/mlx4: fix wrong
, to advertize their port link rate.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
drivers/infiniband/core/sysfs.c | 15 +--
drivers/infiniband/hw/amso1100/c2_provider.c |2 +-
drivers/infiniband/hw/cxgb3/iwch_provider.c |2 +-
drivers/infiniband/hw/cxgb4/provider.c
commit 71eeba16 IB: Add new InfiniBand link speeds introduced a bug
under which the rate for IB SDR/4X links was displayed as 8.5Gbs
instead of 10Gbs, fix that.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
drivers/infiniband/core/sysfs.c | 12 +---
1 files changed, 9 insertions
.
Signed-off-by: Dotan Barak dot...@mellanox.com
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
drivers/infiniband/hw/mlx4/main.c | 97 -
1 files changed, 53 insertions(+), 44 deletions(-)
diff --git a/drivers/infiniband/hw/mlx4/main.c
b/drivers/infiniband
to change the mtu is exposed only when running in
NON SRIOV mode. To allow changing the mtu for the master in SRIOV
mode, PF initiated FLR (Function Level Reset) has to be implemented.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
Roland, on v1 - http://marc.info/?l=linux-rdmam=130636143927387w=2
approach with
future libibverbs checksum offload reporting towards applications, where
adding a bit flag doesn't break the library ABI. This patch was during
conversation with Liran Liss lir...@mellanox.com
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
drivers/infiniband/hw/mlx4/cq.c
interfaces are.
A major use case for clone childs is for virtualization purposes, e.g under
schemes where a per VM NIC / HW queue is desired at the hypervisor level.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
for the case of non-legacy and same pkey childs I wanted to use
a notation
On Wed, Jan 11, 2012 at 9:30 PM, Hefty, Sean sean.he...@intel.com wrote:
This function ends up like this, with my comments inline:
rate = (25 * attr.active_speed) / 10;
We set rate here, but [...] The above switch statement
changes the value for 3 of the 5 cases, with 1 case not
Shawn Bohrer sboh...@rgmadvisors.com wrote:
Is there any estimate on when we might see something like this upstream?
Could you elaborate a little on your use case for multicast IBoE
traffic? e.g how the setup looks like and how are your Ethernet
switches act to route that traffic.
Or.
--
To
Sebastian Riemer sebastian.rie...@profitbricks.com wrote:
[...] we've also tested a 3.0.15 mainline kernel with in-tree IB modules
together with the OFED-1.5.4 user-space and this has much better IPoIB
performance than the kernel stuff from OFED. So, we want to use them
instead, but there is
On 1/12/2012 11:23 AM, Sebastian Riemer wrote:
We are running iSER directly on the host. KVM is compiled in but there
aren't any VMs on our iSER test server. It is a diskless SuperMicro
server with NFS root. On productive servers we have a live-image and
KVM uses the iSER driven block devices
On 1/12/2012 5:18 PM, Sebastian Riemer wrote:
How do I find out if ib1 is on mlx4_1 or mlx4_0
you do ip addr show and compare with
/sys/class/infiniband/mlx4_*/ports/1/gid/0
you didn't send the kernel logs from the failure after opening the iser
(debug_level=2) and libiscsi
On 1/11/2012 10:09 PM, Or Gerlitz wrote:
[...] I'll give 3.0.15 a try tomorrow, however, the error you're
getting iser_drain_tx_cq:tx id 88402391f898 status 4 vend_err 57
means that iser got local protection error (=4) on the first buffer we
used with IB (the connection handshake buffers
, to advertize their port link rate.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
changes from v0:
fixed typo in the enum type name (was ib_port_seed instead of ib_port_speed)
drivers/infiniband/core/sysfs.c | 15 +++-
drivers/infiniband/core/uverbs_cmd.c |3
On Thu, Jan 12, 2012 at 9:30 PM, Roland Dreier rol...@kernel.org wrote:
Seems to have the raw packet QP stuff mixed in now?
sorry, my bad, will fix and resend
Or.
--
To unsubscribe from this list: send the line unsubscribe linux-rdma in
the body of a message to majord...@vger.kernel.org
More
, to advertize their port link rate.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
changes from v0:
fixed typo in the enum type name (was ib_port_seed instead of ib_port_speed)
changes from v1:
removed raw qp code which went in by mistake
drivers/infiniband/core/sysfs.c | 15
do nested flushing only if the device isn't a child
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
setting CONFIG_DEBUG_MUTEXES I see the below warning, however,
for some reason, I didn't manage to trigger it without my other
patch that adds the clones, I don't see how that patch could
On 1/12/2012 9:30 PM, Roland Dreier wrote:
Seems to have the raw packet QP stuff mixed in now?
Hi Roland,
Will you be looking on this series for 3.3?
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
Sebastian Riemer sebastian.rie...@profitbricks.com wrote:
On 12/01/12 17:14, Or Gerlitz wrote:
you didn't send the kernel logs from the failure after opening the iser
(debug_level=2) and libiscsi (debug_libiscsi_session=1
debug_libiscsi_conn=1) debug prints
OK, I've also set mlx4_core
Flavio Baronti f.baro...@list-group.com wrote:
what is then the proper way to stop a thread blocked on
rdma_get_cm_event() (or on ibv_get_cq_event())?
Oh, this is simple - don't get there... have your code
1. get the FDs associated with IB - which you can retrieve from
librdmacm and from
-by: Or Gerlitz ogerl...@mellanox.com
---
changes from previous posting (called V3, December 2010):
pointers follow
0/4 http://marc.info/?l=linux-rdmam=129162076818154w=2
1/4 http://marc.info/?l=linux-rdmam=129162091418289w=2
2/4 http://marc.info/?l=linux-rdmam=129162097418418w=2
. Only processes with the NET_RAW capability
may open such qp. The name raw packet was selected to resemble the
similarity to AF_PACKET / SOL_RAW sockets.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
drivers/infiniband/core/uverbs_cmd.c |3 +++
drivers/infiniband/core/verbs.c |1
Implementation of the RAW Packet QP for the mlx4_ib driver,
use the MLX transport as done by the mlx4_en Ethernet driver.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
drivers/infiniband/hw/mlx4/qp.c | 29 ++---
1 files changed, 26 insertions(+), 3 deletions
to AF_PACKET / SOL_RAW sockets.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
include/infiniband/verbs.h |3 ++-
man/ibv_create_qp.3|2 +-
man/ibv_modify_qp.3| 10 ++
3 files changed, 13 insertions(+), 2 deletions(-)
diff --git a/include/infiniband/verbs.h b
Basic support for RAW Packet QP.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
src/qp.c |4
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/src/qp.c b/src/qp.c
index 40a6689..90c4e80 100644
--- a/src/qp.c
+++ b/src/qp.c
@@ -286,6 +286,10 @@ int mlx4_post_send(struct
On 1/17/2012 12:26 PM, Sebastian Riemer wrote:
In the log from working iSER there is the RDMA mapping debug message
at the position of the error in the other log
I didn't understand that comment, please clarify, anyway, could you try
and patch your 3.0.15 kernel with commit
On 1/17/2012 5:08 PM, Steve Wise wrote:
I think this series should add some new send flags for HW that does
checksum offload [...] also, on ingress, most hardware can do INET
checksum validation, and a way to indicate the results to the
application is needed. Perhaps flags in the CQE? [...]
On 1/18/2012 10:43 AM, Swapna Thete wrote:
@@ -1963,6 +1963,16 @@ local:
+ } else {
+ memcpy(response, recv, sizeof(*response));
+ response-header.recv_wc.wc =response-header.wc;
+ response-header.recv_wc.recv_buf.mad =response-mad.mad;
+
On Wed, Jan 18, 2012 at 8:47 PM, Hefty, Sean sean.he...@intel.com wrote:
+ MLX4_RAW_QP_MTU = 7,
Is there any special meaning for MTU 7 that should be standardized in
ib_verbs.h?
In other places, the code uses IB_MTU_4096 (5):
nope, its the ConnectX firmware that dictates this
On 1/19/2012 1:29 PM, Sebastian Riemer wrote:
just cherry-picking that commit didn't do the job. Therefore, I've
backported the whole ib_iser code from 3.2.1 to 3.0.15. Now it works fine
I would suspect one of these
# git shortlog v3.0..v3.2 drivers/infiniband/ulp/iser
[...]
Or Gerlitz (4
This short series fixes two kind of badly hurting bugs that we
stepped on lately. We would be happy to have them treated quickly
so they can be pushed further to -stable, distros, etc.
The 1st bug is in the mlx4_ib driver where vendor mads are just silently
dropped, which means that vendor
. For unsupported attributes,
the firmware returns a GET_RESPONSE MAD containing an error status.
Signed-off-by: Jack Morgenstein ja...@mellanox.com
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
For example, this FDR10 node with LID 11
# ibstat mlx4_0 1
CA: 'mlx4_0'
Port 1:
State: Active
-by: Shlomo Pongratz shlo...@mellanox.com
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
drivers/infiniband/ulp/ipoib/ipoib_ib.c | 15 +--
1 files changed, 13 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
On 1/26/2012 4:39 PM, Or Gerlitz wrote:
This short series fixes two kind of badly hurting bugs that we
stepped on lately. We would be happy to have them treated quickly
so they can be pushed further to -stable, distros, etc.
Roland, when (hopefully soon) accepting these, could you please add
On 1/30/2012 6:36 AM, Roland Dreier wrote:
On Thu, Jan 26, 2012 at 6:43 AM, Or Gerlitzogerl...@mellanox.com wrote:
The GRO flow makes a check in every layer to ensure the packets
are actually merged only if they match at all layers.
The first GRO check, at L2 always fails for IPoIB, since it
On 1/30/2012 10:04 AM, Eric Dumazet wrote:
Hmm, do we really need to compare ether header, thats the question.
IMHO, GRO could avoid this check, as legal trafic could be never
merged (eg multipath)
Yep, basically I tend to agree that GRO could do very well with L3 and
L4 checks, the
On Mon, Jan 30, 2012 at 8:11 PM, Roland Dreier rol...@kernel.org wrote:
But is this general fix so difficult that we can't just do the right thing?
no one said its so difficult, its just not trivial, b/c of the many
use cases over which the GRO logic plays... lets see if the network
folks come
On 2/1/2012 4:07 PM, Eric Dumazet wrote:
If we just turn it into a memcmp with a variable length would
that work for you?
I think yes, FWIW we will compare the IPoIB header, Roland is that okay for you?
A memcmp(xxx, yyy, variable_len) will be out of line and slow, its a bit sad
... Are
On 2/2/2012 4:38 PM, Eric Dumazet wrote:
Maybe we can keep a fastpath for ethernet case... (the if (hlen == ETH_HLEN)
being always predicted) Maybe need to introduce gro_hard_header_len as well)
today, IPoIB advertizes hard_header_len which is bigger than the IPoIB
header len, this is done
On Thu, 26 Jan 2012, Or Gerlitz wrote:
From: Shlomo Pongratz shlo...@mellanox.com
The GRO flow makes a check in every layer to ensure the packets
are actually merged only if they match at all layers.
The first GRO check, at L2 always fails for IPoIB, since it assumes
that all packets have 14
On 2/3/2012 10:24 PM, Hefty, Sean wrote:
Hmm... a tcpdump would help to understand what is going on (for example if GRO
kicks in)
I should stop trying to apply patches so late. I must have done something
wrong with my quick test. Re-applying the patch on 3.2, the ipoib UD
performance
On 2/6/2012 5:21 PM, Eric Dumazet wrote:
Sure, I suggest two patches then. One from me (the one I sent
earlier), introducing the core infrastructure, and one from you adding
the dev-gro_mac_header_len = IPOIB_ENCAP_LEN; in
drivers/infiniband/ulp/ipoib/ipoib_main.c If you agree, I'll resend my
On 2/6/2012 6:31 PM, David Miller wrote:
We really need an explanation, probably both in the commit message and
the comments next to this new struct member, explaining why in the
world we can't use -hard_header_len for this.
Dave,
As I wrote earlier on this thread, the reason is that in
On 2/6/2012 6:58 PM, David Miller wrote:
So if IPoIB path resolution was properly integrated into the neighbour
cache state machine, instead of being implemented awkwardly in the
device transmit path, this crap wouldn't be necessary right?
Dave, say we have integrated the path resolution
On 2/6/2012 7:07 PM, Eric Dumazet wrote:
OK, I'll resend my first patch then, using hard_header_len
so, we will be back to square one... as the hard_header_len which
advertized now by IPoIB will fail
the GRO L2 check... lets see where this discussion evolves.
Or.
--
To unsubscribe from
David Miller da...@davemloft.net wrote:
Actually there is a way to make this work.
Define your ipoib_skb_cb something like:
struct ipoib_skb_cb {
struct qdisc_skb_cb qdisc_cb;
... ipoib stuff goes here ...
};
That way you can use the SKB cb area for your ipoib info
On Wed, Feb 8, 2012 at 10:50 PM, David Miller da...@davemloft.net wrote:
From: Eric Dumazet eric.duma...@gmail.com
Shlomo Pongratz reported GRO L2 header check was suited for Ethernet
only, and failed on IB/ipoib traffic.
He provided a patch faking a zeroed header to let GRO aggregates
On Thu, Feb 2, 2012 at 7:08 PM, Roland Dreier rol...@kernel.org wrote:
On Thu, Feb 2, 2012 at 3:12 AM, Eric Dumazet eric.duma...@gmail.com wrote:
Any news on this patch ?
Sorry, just dropped it in the shuffle. I'll get it into 3.4, thanks.
Roland,
I noted that you typically use the
On Wed, Feb 8, 2012 at 11:31 PM, David Miller da...@davemloft.net wrote:
From: Or Gerlitz or.gerl...@gmail.com
Hi Dave, for correct operation / future bisection, you should 1st
apply Roland's patch which reduces the hard header len advertized by
ipoib to be only the size of the ipoib header
On Thu, Feb 9, 2012 at 12:41 AM, Mike Heinz michael.he...@qlogic.com wrote:
This behavior has been demonstrated on 1.5.4 and 1.5.4.1, and on RHEL6.0, 6.1
and 5.7
tried kernel.org? this list deals with kernels from that fabric.
Or.
--
To unsubscribe from this list: send the line unsubscribe
On Thu, Feb 9, 2012 at 1:09 AM, David Miller da...@davemloft.net wrote:
I don't think this is an appropriate bug fix at all.
I'm not sure to understand
Apparently this problem has existed since day one and the world has
kept on spinning meanwhile.
Dave, the bug was introduced on 2.6.38 when
On 2/9/2012 1:26 AM, David Miller wrote:
the bug was introduced on 2.6.38 when LRO was removed from IPoIB
and GRO added, so users who run distributions containing pre 2.6.38
kernels don't hit the bug.
This I didn't know, Ok I'll apply this stuff to the net tree and add it
to my -stable queue as
On Thu, Feb 9, 2012 at 8:52 PM, David Miller da...@davemloft.net wrote:
I've fix this now, thanks.
Yep, I see it there, still, for correctness/biscetion, you should
rebase such that the order of commits when pushed to Linus is
3. 5ca3b72c5da47d95b83857b768def6172fbc080a gro: more generic L2
On Thu, Feb 9, 2012 at 10:29 PM, David Miller da...@davemloft.net wrote:
I cannot rebase my tree, too many people use it and I would break
their stuff. We simply have to live with this error.
I see, so you will add CC to stable, such that they will be picked
into -stable once you push them to
On Wed, Jan 11, 2012 at 6:52 PM, Or Gerlitz ogerl...@mellanox.com wrote:
The series is made of four elements
Hi Roland,
Can you provide feedback / comment on these patches? I've got Sean's
okay on patches #1 saying I need one fix, and also I addressed your
comment and posted V2 @
http
-space in Ethernet environments, e.g by applications
that do TCP/IP offloading. Only processes with the NET_RAW capability
may open such qp. The name raw packet was selected to resemble the
similarity to AF_PACKET / SOL_RAW sockets.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
Reviewed
On 1/11/2012 9:45 PM, Or Gerlitz wrote:
Yep, this function is surely not programming state of the art..
however, except for SDR, the above line doing rate = rate *
ib_width_enum_to_int(attr.active_width) would hold for all the speeds
and all the widths (1X, 4X, 12X), if we just set rate
On 2/27/2012 7:23 PM, Roland Dreier wrote:
It doesn't really work if everything needs me to micromanage every
step... the original patch obviously made some ugly code uglier, Sean
pointed out an obvious way to make things cleaner. If no one has a
problem, I'll just go ahead with the
link speeds) introduced a bug
where eg the rate for IB 4X SDR links iss displayed as 8.5 Gb/sec
instead of 10 Gb/sec as it used to be. Fix that.
Reported-by: Or Gerlitz ogerl...@mellanox.com
Signed-off-by: Roland Dreier rol...@purestorage.com
Index: linux-2.6/drivers/infiniband
, to advertize their port link rate.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
changes from v0:
fixed typo in the enum type name (was ib_port_seed instead of
ib_port_speed)
changes from v1:
removed raw qp code which went in by mistake
changes from v2:
changed the patch such that it applies
On 2/25/2012 9:36 PM, Doug Ledford wrote:
We allocate the login dma buffers in iser_verbs.c as part of
alloc_ib_conn_resources(), however we are freeing them in
iser_initiator.c as part of iser_free_rx_descriptors(). This is
needlessly confusing. We have an alloc_rx_descriptors() and it
processes with the NET_RAW capability may open
such qp. The name raw packet was selected to resemble the similarity
to AF_PACKET / SOL_RAW sockets.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
changes from V0:
removed spaces at ib_verbs.h that went in by mistake
drivers/infiniband/core
On 2/25/2012 9:36 PM, Doug Ledford wrote:
@@ -274,6 +274,18 @@ static int iser_free_ib_conn_res(struct iser_conn
*ib_conn, int can_destroy_id)
ib_conn-cma_id = NULL;
kfree(ib_conn-page_vec);
+ if (ib_conn-login_buf) {
+ if (ib_conn-login_req_dma)
+
alloc something that the free_rx_descriptors() frees, and we have an
alloc_ib_conn_resources() that allocs something not freed by
free_ib_conn_resources(). Clean that up.
Signed-off-by: Doug Ledford dledf...@redhat.com
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
---
changes from V0:
fixed
.
To solve that, rely on the flags bits in the login request to have FFP (0x3)
in the lower nibble, as a marker for the final login request, and post an
initial chunk of receive buffers before sending that login request instead
of after getting the login response.
Signed-off-by: Or Gerlitz ogerl
error event
Or Gerlitz (1):
net/mlx4: remove redundant maintainance of extended port capabilities
drivers/infiniband/hw/mlx4/cq.c |3 +-
drivers/infiniband/hw/mlx4/main.c | 27 --
drivers/net/ethernet/mellanox/mlx4/eq.c | 32
,
since there's no real reaon to maintain the extended capabilities of the port,
as they can be queried on demand (e.g the FDR10 one). This patch reverts
commit 97285b7817 and removes the check for extended caps from the mlx4_ib
driver port query flow.
Signed-off-by: Or Gerlitz ogerl...@mellanox.com
From: Jack Morgenstein ja...@dev.mellanox.co.il
Print an error message when an async event of thermal error is
being reported by the HW.
Signed-off-by: Jack Morgenstein ja...@dev.mellanox.co.il
Signed-off-by: Dotan Barak dot...@mellanox.com
---
drivers/net/ethernet/mellanox/mlx4/eq.c | 32
From: Eli Cohen e...@mellanox.com
If an erroneous CQE was polled in the first iteration (e.g. npolled == 0), we
would not update the consumer index hence the hardware could get a wrong notion
of how many CQEs software polled. Fix this by unconditionally updating the
doorbell record. We could put
On 3/6/2012 5:44 AM, Bob Pearson wrote:
From the RTS state issuing a modify qp with qp_attr_mask only set to state to
the SQD state
AFAIK SQD isn't supported by connectx
Or.
--
To unsubscribe from this list: send the line unsubscribe linux-rdma in
the body of a message to
On 1/11/2012 6:52 PM, Or Gerlitz wrote:
This patch comes to address the long pending patch for 4k mtu
along the lines of what was requested over the last review
[PATCH 4/6] net/mlx4: allow for dynamic mtu configuration for IB ports
This patch is simple and comes to unify the approach
On 2/16/2012 5:27 PM, Christoph Lameter wrote:
On Thu, 16 Feb 2012, Or Gerlitz wrote:
Also here, Sean provided his reviewed-by signature, people (CCed Christoph and
others) keep asking me about this patch set and I didn't get any feedback from
you.
I sure would like to see this merged
On Mon, Mar 5, 2012 at 8:09 PM, Roland Dreier rol...@kernel.org wrote:
Handle this by having the driver allocate at least enough MTT entries to
cover 2 * totalram pages.
just curious, why we want to cover totalram? also the commit title
has mxl4 instead of mlx4
Or.
--
To unsubscribe from this
On 3/13/2012 1:29 AM, Roland Dreier wrote:
It's just a heuristic, but I figured some app might want to register
essentially all of memory, and then we want some more to cover other
users. The amount of memory used for unused MTT space is pretty small,
I think.
If an app registered
On 3/20/2012 7:18 PM, parav.pan...@emulex.com wrote:
From: Parav Panditparav.pan...@emulex.com
Emulex One Connect Adapter is RDMA (RoCE) capable multi-function PCI Express
device. This driver patch enables RoCE support on such adapter. This ocrdma
driver depends on be2net NIC driver.
This
On Mon, Mar 19, 2012 at 7:11 PM, Roland Dreier rol...@kernel.org wrote:
InfiniBand/RDMA changes for the 3.4 merge window. Nothing big really stands
out; by
patch count lots of fixes to the mlx4 driver plus some cleanups and fixes to
the core
and other drivers.
Hi Roland,
So again, any
On Wed, Mar 21, 2012 at 11:03 PM, Christoph Lameter c...@linux.com wrote:
On Wed, 21 Mar 2012, Or Gerlitz wrote:
So again, any reason not to merge the RAW QP patches for 3.4? they
have been posted few months ago, its two kernel patches and we have
Sean's reviewd-by Signature for the core
On 4/2/2012 2:16 PM, Bart Van Assche wrote:
On 04/02/12 10:33, Or Gerlitz wrote:
As far as I can see the link layer value is fine:
$ cat /sys/class/infiniband/mlx4_0/ports/1/link_layer
InfiniBand
$ cat /sys/class/infiniband/mlx4_0/ports/2/link_layer
InfiniBand
So the two ports are actually
On 4/2/2012 2:48 PM, Bart Van Assche wrote:
The two ports are connected back-to-back to another mlx4 HCA. I
noticed this behavior change since opensm stopped working after
rebooting into 3.4-rc1.
can you add these prints and send me the output after attempting to cat
the rate file?
Or.
On 4/2/2012 3:51 PM, Or Gerlitz wrote:
can you add these prints and send me the output after attempting to
cat the rate file?
okay, on a system which has IB on port 1 and Ethernet on port 2, using
this patch
I get these prints:
ib_link_query_port active_speed 4
rate_show ret 0
On 4/2/2012 4:35 PM, Bart Van Assche wrote:
Some additional info:
- This issue only occurs if the back-to-back connected system is down,
not if it is running.
- The output I get with the other system down is:
# cat /sys/class/infiniband/mlx4_0/ports/1/link_layer
InfiniBand
# dmesg
When the IB port is down, the active_speed value returned by the MAD_IFC
command equals seven (7) which isn't among the IB speeds defined by the
ib_port_speed enum. This results in invalid speed value seen by higher
layers or applications who do port query. Fix that by setting the speed
to be SDR
Or Gerlitz (2):
IB/mlx4: fix the case of invalid speed value returned when the port is down
IB/core: add missing string for the display of SDR rates in sysfs
drivers/infiniband/core/sysfs.c |1 +
drivers/infiniband/hw/mlx4/main.c |4
2 files changed, 5 insertions(+), 0
commits 2e96691c IB: Use central enum for speed instead of hard-coded values
and e9319b0cb IB/core: Fix SDR rates in sysfs still didn't fill in the SDR
string in the SDR switch case, fix that.
---
drivers/infiniband/core/sysfs.c |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff
On 4/2/2012 4:25 PM, Hal Rosenstock wrote:
I think there are 3 main issues here:
1. EINVAL can be returned from rate_show and hence Invalid argument
rate string should be handled in libibumad. I think this was Bart's original
point.
2. Why is rate_show returning EINVAL ? I think that's what
On 4/2/2012 7:35 PM, Hal Rosenstock wrote:
Rather than always overwriting active_speed in this case, wouldn't it
be better to only do that for invalid values?
Yes, I have thought about that, however, spotting invalid values would
make the code a bit ugly, so I took this approach, Roland?
401 - 500 of 2157 matches
Mail list logo