On Wed, 28 Oct 2015 03:45:50 +0100
Arnd Bergmann wrote:
> As far as I can tell, there is a preexisting race condition
> regarding the cmd->use_events flag, which is not protected
> by any lock. When this flag is toggled while another command
> is being started, that command gets stuck until the m
On Mon, 16 Mar 2015 18:49:59 +0100 (CET)
Sebastian Ott wrote:
> From: Sebastian Ott
> To: linux-rdma@vger.kernel.org, linux-ker...@vger.kernel.org
> cc: Roland Dreier , Sean Hefty
> , Hal Rosenstock , Or
> Gerlitz , "David S. Miller"
> , Yishai Hadas , Ira Weiny
&g
On Wed, 31 Dec 2014 02:26:07 +0530
Bob Biloxi wrote:
> Hi,
>
> I was going through the mlx4 source code and had a few questions
> regarding the generation of interrupts upon execution of the NOP
> command from the VF driver.
>
> If i am running as a dedicated driver, then NOP seems to work fine
On Mon, 15 Dec 2014 15:07:58 +0530
Bob Biloxi wrote:
> am I correct in my understanding
> when i say that MAD_DEMUX feature is not required to be
> supported/implemented in Mellanox RoCE Drivers?
>
> It is required only for Infiniband drivers?
Actually, you will need to support MAD_DEMUX anyway
On Mon, 10 Nov 2014 19:58:46 +0530
Bob Biloxi wrote:
> Hi,
>
> Is FMR (Fast Memory Regions) supported in a multi-function mode?
In SRIOV, FMR is supported only for the PF, not for VFs (since this
feature requires writing directly to mapped ICM memory).
You can see this in file drivers/infiniba
On Mon, 10 Nov 2014 20:05:54 +0530
Bob Biloxi wrote:
> Hi,
>
> I was going through the mlx4 code and previous mailing lists when I
> came across the following thread:
>
>
> http://marc.info/?l=linux-rdma&m=134398354428293&w=2
>
>
> In that thread, it is mentioned as follows:
>
> Some Limita
On Tue, 23 Sep 2014 15:55:14 +
"Hefty, Sean" wrote:
> > --- a/drivers/infiniband/core/uverbs_main.c
> > +++ b/drivers/infiniband/core/uverbs_main.c
> > @@ -502,6 +502,10 @@ void ib_uverbs_qp_event_handler(struct ib_event
> > *event, void *context_ptr)
> > {
> > struct ib_uevent_object *u
On Wed, 9 Jul 2014 18:40:46 +0530
Bob Biloxi wrote:
> Hi,
>
> I was going through the mr.c file as part of understanding WRITE_MTT
> command in the mlx4 code.
>
> I could see that instead of issuing the WRITE_MTT HCR command, in case
> of SRIOV, we're directly accessing the ICM space for the MT
On Wed, 2 Jul 2014 18:51:17 -0400
Chuck Lever wrote:
> I built my own according to your description. The fix seems good.
>
Thanks, Chuck!
-Jack
--
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
On Tue, 1 Jul 2014 18:17:27 -0400
Chuck Lever wrote:
> Hi-
>
> This happens during boot, and means that lockdep and lock_stat are
> completely disabled during any testing I do.
>
> Does this splat have a known fix, or should I file an upstream
> bugzilla?
>
>
> Jul 1 15:50:22 manet kernel: =
.c:2875:3:
> warning: case value '7' not in enumerated type 'enum qp_transition'
> [-Wswitch]
>case MLX4_QP_ST_MLX:
>
> Reported-by: kbuild test robot
> Fixes: 99ec41d0a48c ("mlx4: Add infrastructure for selecting VFs
> to enable QP0
On Thu, 13 Feb 2014 00:18:22 +0530
Bob Biloxi wrote:
> The VFs need to allocate the memory for Send Queue Buffer, Receive
> Queue Buffer, Completion Queue Buffer, Event Queue Buffer.
>
> Is that right?
Yes.
>
> Also, as the QPs, CQs etc are created by the HCA when ALLOC_RES
> command is issue
On Wed, 29 Jan 2014 15:52:09 +0530
Bob Biloxi wrote:
> These paths are taken based on the return value of mlx4_is_func(dev).
> This is true for MASTER or SLAVE which I believe is Physical Function
> Driver/Virtual Function Driver. So for SRIOV, it covers all cases.
>
> The MAP_ICM portion which
On Sun, 29 Dec 2013 23:47:20 +0100
Julia Lawall wrote:
> diff --git a/drivers/infiniband/hw/mlx4/sysfs.c
> b/drivers/infiniband/hw/mlx4/sysfs.c index 97516eb..db2ea31 100644
> --- a/drivers/infiniband/hw/mlx4/sysfs.c
> +++ b/drivers/infiniband/hw/mlx4/sysfs.c
> @@ -582,8 +582,10 @@ static int add
On Mon, 21 Oct 2013 00:12:54 -0400
Doug Ledford wrote:
> I think I like my suggestion better: go back to having a full table,
> but use a bitmap to indicate valid entries and then use the bitmap to
> limit our comparisons in the find_cached* functions, and put the
> get_* funtions back to being O
ACK. Looks good!
Very nice catch, Doug!
-Jack
On Tue, 24 Sep 2013 17:16:27 -0400
Doug Ledford wrote:
> The cma_acquire_dev function was changed by commit 3c86aa70bf67
> to use find_gid_port because multiport devices might have
> either IB or IBoE formatted gids. The old function assumed that
On Tue, 24 Sep 2013 17:16:29 -0400
Doug Ledford wrote:
> @@ -85,13 +91,26 @@ int ib_get_cached_gid(struct ib_device *device,
>
> cache = device->cache.gid_cache[port_num -
> start_port(device)];
> - if (index < 0 || index >= cache->table_len)
> + if (index < 0 || index >= cache->
On Sun, 6 Oct 2013 14:53:18 +0400
Vasiliy Tolstov wrote:
> 2013/10/6 Or Gerlitz :
> > # lspci | grep Mell
> >
> > 06:00.0 InfiniBand: Mellanox Technologies MT26428 [ConnectX VPI
> > PCIe 2.0 5GT/s - IB QDR / 10GigE] (rev b0)
> > 07:00.0 Network controller: Mellanox Technologies MT27520 Family
> >
On Wed, 2 Oct 2013 12:49:07 +0200
Alexander Gordeev wrote:
> Subject: [PATCH RFC 51/77] mthca: Update MSI/MSI-X interrupts
> enablement code Date: Wed, 2 Oct 2013 12:49:07 +0200
> Sender: linux-rdma-ow...@vger.kernel.org
> X-Mailer: git-send-email 1.7.7.6
>
> As result of recent re-design of t
On Wed, 2 Oct 2013 12:49:02 +0200
Alexander Gordeev wrote:
> As result of recent re-design of the MSI/MSI-X interrupts enabling
> pattern this driver has to be updated to use the new technique to
> obtain a optimal number of MSI/MSI-X interrupts required.
>
> Signed-off-by: Alexander Gordeev
On Wed, 2 Oct 2013 12:49:02 +0200
Alexander Gordeev wrote:
UPDATING THIS REPLY.
Your change log confused me. The change below is not from a "recent
re-design", it is required due to an earlier patch in this patch set.
>From the log, I assumed that the change you are talking about is already
upst
On Wed, 2 Oct 2013 12:49:02 +0200
Alexander Gordeev wrote:
NACK. This change does not do anything logically as far as I can tell.
pci_enable_msix in the current upstream kernel itself calls
pci_msix_table_size. The current code yields the same results
as the code suggested below. (i.e., the su
Hi Roland,
This is a re-posting (and rewording) of a question I sent you on July 6,
2009.
I've been looking at the write-combining support in the kernel,
and it looks good. The caller simply invokes pgprot_writecombine() and
if write combining is available, the region is mapped for it (if wc is
n
On Wed, 28 Aug 2013 13:45:17 -0400
do...@rdmamojo.com wrote:
> Hi Yann.
>
>
> >
> > I'd like the behavior of ibv_create_qp() regarding
> > cap.max_inline_data be clearer:
> >
> > 1) Should ibv_create_qp() failed if max_inline_data requested being
> > greater than supported ?
> Since the max_inli
On Wednesday 03 July 2013 22:26, Roland Dreier wrote:
> Look at the actual timer code. del_timer_sync() won't work if
> something unrelated re-adds the timer, but it will work if the timer
> itself is what re-adds itself.
>
> Documentation/DocBook/kernel-locking.tmpl says:
>
> Another comm
On Thursday 04 July 2013 17:17, Joe Perches wrote:
> Not the vertical spacing, the newline inside the quotes.
> ie: "unknown status code\n" should be "unknown status code"
>
You're right! I misunderstood, and did not notice the newline char.
Good catch!
While I was at it, though, I eliminated th
On Thursday 04 July 2013 16:15, Jack Morgenstein wrote:
> > > + *inlen = sizeof **cqb + sizeof *(*cqb)->pas * ncont;
> >
> > sizeof always uses parentheses
>
I'll fix this, too.
-Jack
--
To unsubscribe from this list: send the line "unsubscribe linux-rdm
On Thursday 04 July 2013 00:10, Joe Perches wrote:
> On Wed, 2013-07-03 at 20:13 +0300, Or Gerlitz wrote:
> > From: Eli Cohen
>
> More trivia:
>
> > diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h
> > b/drivers/infiniband/hw/mlx5/mlx5_ib.h
> []
> > +#define mlx5_ib_dbg(dev, format, arg...)
On Wednesday 03 July 2013 23:29, Joe Perches wrote:
> On Wed, 2013-07-03 at 20:13 +0300, Or Gerlitz wrote:
> > From: Eli Cohen
>
> trivial comments:
>
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
> > b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
> []
> > +static const char *
On Wednesday 03 July 2013 23:59, Joe Perches wrote:
> On Wed, 2013-07-03 at 20:13 +0300, Or Gerlitz wrote:
> > From: Eli Cohen
> > diff --git a/drivers/infiniband/hw/mlx5/ah.c
> > b/drivers/infiniband/hw/mlx5/ah.c
> []
> > +struct ib_ah *create_ib_ah(struct ib_ah_attr *ah_attr,
> > +
On Monday 01 July 2013 22:18, David Miller wrote:
> Actually, you should adjust both decrements to read "i--".
>
OK, will do it.
-Jack
--
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
On Saturday 29 June 2013 07:10, David Miller wrote:
> From: Or Gerlitz
> Date: Wed, 26 Jun 2013 17:22:12 +0300
>
> > + for (--i; i >= 0; --i) {
>
> Please, "i--" is more canonical in for() loops.
>
> > + for (--i; i >= 0; --i) {
>
> Likewise.
> --
> To unsubscribe from this list: send the
On Tuesday 21 May 2013 13:43, Or Gerlitz wrote:
> On 21/05/2013 13:42, Bart Van Assche wrote:
> > On 05/21/13 11:40, Or Gerlitz wrote:
> >> 2. is possible in the Linux kernel for one hard irq callback to flash on
> >> CPU X while another hard irq callback is running on the same CPU?
> >
> > I think
00f5cd8859 RSP: f3eafa28 RFLAGS: 00200202
RAX: RBX: f3eafaf8 RCX: 0b6d4f00
RDX: 001c RSI: 001c RDI:
RBP: ffff810bf1a0a120 R8: 0000 R9: 0000
R10: R11: 00
On Friday 12 April 2013 02:13, Or Gerlitz wrote:
> Klaus Wacker wrote:
> > Hi all, we are doing a project on IBM x-blades using Mellanox/RoCE. I have
> > some
> > difficulties to get RoCE based VLANs to run.
> > We are currently on Linux 3.8. No OFED material involved.
> > While non-VLAN devices
On Thursday 04 April 2013 16:01, Kleber Sacilotto de Souza wrote:
> On 04/02/2013 02:00 PM, Roland Dreier wrote:
> >> diff --git a/drivers/infiniband/hw/mlx4/qp.c
> >> b/drivers/infiniband/hw/mlx4/qp.c
> >> index 35cced2..0fa4f72 100644
> >> --- a/drivers/infiniband/hw/mlx4/qp.c
> >> +++ b/drivers
On Tuesday 02 April 2013 12:15, Or Gerlitz wrote:
> On 28/03/2013 20:45, Kleber Sacilotto de Souza wrote:
> > When the PCI adapter is going through error recovery, a call to
> > mlx4_ib_post_send() will return success without the command actually
> > arriving to the hardware. Adding a call to pci_c
On Monday 25 February 2013 19:23, Roland Dreier wrote:
> On Mon, Feb 25, 2013 at 8:54 AM, Roland Dreier wrote:
> > I'm finally noticing that this is in the build_mlx_header() function,
> > which is pretty much a slow path. Certainly another compare isn't
> > going to change performance given all
On Thursday 21 February 2013 11:02, Paul Bolle wrote:
> diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
> index 19e0637..512fde3 100644
> --- a/drivers/infiniband/hw/mlx4/qp.c
> +++ b/drivers/infiniband/hw/mlx4/qp.c
> @@ -1778,8 +1778,8 @@ static int build_mlx_header
00.0: mlx4_handle_slave_flr
> [ 1438.969776] mlx4_core :82:00.0: mlx4_handle_slave_flr: clean slave: 1
> [ 1439.070931] assign device 0:82:0.1
> [ 1439.073469] mlx4_core :82:00.0: FLR event for slave: 1
> [ 1439.073526] mlx4_core 0000:82:00.0: mlx4_handle_slave_flr
> [ 1439.073529] ml
On Tuesday 29 January 2013 15:15, Mathis GAVILLON wrote:
> In fact, one boxe wasn't check in IOMMU submenu (Support for Interrupt
> Remapping (EXPERIMENTAL)). I've recompiled the kernel of the host and
> guest (is it necessary for this last one ?). But the error message is
> already present. Just b
On Tuesday 29 January 2013 15:15, Mathis GAVILLON wrote:
> In fact, one boxe wasn't check in IOMMU submenu (Support for Interrupt
> Remapping (EXPERIMENTAL)). I've recompiled the kernel of the host and
> guest (is it necessary for this last one ?)
Yes, it is necessary. Without that checked, I not
x4_core :00:08.0: IPoIB checksum offload
> mlx4_core :00:08.0: P_Key violation counter
> mlx4_core :00:08.0: Q_Key violation counter
> mlx4_core :00:08.0: DPDP
> mlx4_core :00:08.0: Big LSO headers
> mlx4_core :00:08.0: APM support
> mlx4_c
You could use:
u16 uninitialized_var(vlan);
instead.
Although this in the special QP data flow, I still prefer to avoid adding extra
code (even setting
initial values at procedure entry). The line above will also do the job.
"uninitialized_var"
is used elsewhere in the driver. See, for e
ConnectX3 when flow steering is enabled.
Currently, the IB case in mlx4_QP_FLOW_STEERING_ATTACH_wrapper()
is missing, so the procedure returns -EINVAL and IPoIB on VFs fails to operate.
Signed-off-by: Jack Morgenstein
Signed-off-by: Or Gerlitz
---
.../net/ethernet/mellanox/mlx4/resource_tracker.c
ur for-next and
> also the initial
> patch which in a way is more lengthy and heavy. I wonder whether
> wouldn't it be fare to allow for Jack to review it before pushing?
> Jack is back by Tuesday.
>
Acked-by: Jack Morgenstein for patches 1-3
Nice going, Roland!
-
Acked-by: Jack Morgenstein
Thanks, Roland! Good catches and good fixes!
Regarding the mutex replacing the semaphore, at one time we toyed
with the idea of multiple comm channel commands "in the air", but
we did not pursue the idea.
I agree with changing slave_sem to a mutex. If we
Thanks, Roland!
-Jack
On Monday 24 September 2012 21:34, Roland Dreier wrote:
> So I applied this whole series, with the plan to merge this for 3.7.
>
> Please send any changes as patches on top of what's already merged.
>
> Thanks,
> Roland
>
--
To unsubscribe from this list: send the line
Hi Roland,
I am on vacation until next Tuesday -- I'll look at this then.
-Jack
On Monday 24 September 2012 21:36, Roland Dreier wrote:
> On Mon, Sep 24, 2012 at 7:02 AM, Stephen Rothwell
> wrote:
> > After merging the akpm tree, today's linux-next build (powerpc
> > ppc64_defconfig) failed li
On Friday 21 September 2012 02:15, Or Gerlitz wrote:
> On Tue, Sep 11, 2012 at 8:10 PM, Doug Ledford wrote:
> > On 8/3/2012 4:40 AM, Jack Morgenstein wrote:
> > > struct mlx4_ib_sriov{} is created by the master only.
> > > It is a container for the following:
> >
On Wednesday 12 September 2012 19:48, Doug Ledford wrote:
> > On the Hypervisor, however, we assume that if both versions of the pkey are
> > in its pkey table,
> > then for its own infiniband operation (as opposed to performing its pkey
> > virtualizing function),
> > it should operate with the
On Tuesday 11 September 2012 19:52, Doug Ledford wrote:
> On 8/3/2012 4:40 AM, Jack Morgenstein wrote:
> > Enhance the cached and non-cached pkey table lookups to enable limited and
> > full
> > members of the same pkey to co-exist in the pkey table.
> >
> > Thi
On Monday 10 September 2012 16:27, Or Gerlitz wrote:
> I took a look on the practice/wrapping used over the mm subsystem for
> radix_tree_lookup calls, whose maintainer,
> Andrew Morton is signed on the patch Roland pointed to, its just
> rcu_read_lock/unlock, seems this is what to do as well.
>
On Monday 10 September 2012 13:35, Or Gerlitz wrote:
> Jack, Max
>
> Actually, can't we do well with rcu_read_lock() in mlx4_cq_completion()
> as that commit documentation suggests?
>
I don't know. I do notice (in file include/linux/rcupdate.h) that
rcu_read_lock/unlock
is meant to be used in
On Sunday 09 September 2012 18:10, Roland Dreier wrote:
>
> Please look at commit 7cf9c2c76c1a ("[PATCH] radix-tree: RCU lockless
> readside")
>
Roland,
What about the following note (from the commit diff mentioned above):
+/**
+ * Radix-tree synchronization
+ *
+ * The radix-tree API requires
Thanks, Doug!
-Jack
On Wednesday 05 September 2012 17:55, Doug Ledford wrote:
> On 8/3/2012 4:40 AM, Jack Morgenstein wrote:
> > Reserve bits 26-31 for internal use by low-level drivers. Two
> > such bits are used in the mlx4 driver SRIOV IB implementation.
> >
>
On Tuesday 14 August 2012 19:20, Roland Dreier wrote:
> On Tue, Aug 14, 2012 at 7:56 AM, Jack Morgenstein
> wrote:
> > I just checked patchwork, and do not see the bug-fix patch I submitted on
> > Aug 3:
> > [PATCH] IB/mlx4: fix possible deadlock with sm_lo
On Tuesday 10 July 2012 20:08, Roland Dreier wrote:
> Not sure if everyone saw this, but thanks to the kernel.org crew,
> https://patchwork.kernel.org/project/linux-rdma/list/ is back online.
>
> I've been delegating opensm patches to Alex Netes and diags
> patches to Ira Weiny. Can someone refre
Hi Bart,
I submitted a patch to Roland on August 3 (along with SRIOV-IB V2) to fix this:
[PATCH] IB/mlx4: fix possible deadlock with sm_lock spinlock
I notice that you tested out the fix and it worked.
Roland, please take the patch and submit to Linus. This fixes a bug in
the upstream 3.6-RC1
2 flag parameters in mlx4_MAD_IFC already:
ignore_bkey and ignore_mkey. These two parameters are replaced
by a single "mad_ifc_flags" parameter, with different bits set
for each flag. A third flag is added: "network-view/host-view".
Signed-off-by: Jack Morgenstein
---
drivers
use of the 2 most-significant bits in enum
mlx4_ib_qp_flags (based upon enum ib_qp_create_flags (ib_verbs.h))
We need these bits in the low-level driver for internal purposes.
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/cq.c | 31 ++-
drivers/infiniband/hw/mlx4/mlx4_ib.h
ndex 10 to virtual index 1, that guest, whenever
it uses its pkey index 1 will actually be using the real pkey index 10.
Based on patch from Erez Shitrit
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/Makefile |2 +-
drivers/infiniband/hw/mlx4/alias_GUID.c |6 +
(paravirtualized) PF code.
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c | 12 +-
drivers/infiniband/hw/mlx4/qp.c | 104 +++--
drivers/net/ethernet/mellanox/mlx4/fw.c | 175
drivers/net/ethernet/mellano
receive.
When the real QP0 goes down, we should indicate that
the port is not active.
Signed-off-by: Jack Morgenstein
---
drivers/net/ethernet/mellanox/mlx4/fw.c | 71 +--
drivers/net/ethernet/mellanox/mlx4/qp.c | 38 +++--
2 files changed, 92 insertions
When have VFs and PFs on same host, the VFs are activated within the
mlx4_core module before the mlx4_ib kernel module is loaded.
When the mlx4_ib module initializes the PF (master), it now creates
mad paravirtualization contexts for any VFs already active.
Signed-off-by: Jack Morgenstein
Remove the error returns for IB ports from mlx4_ib_add,
mlx4_INIT_PORT_wrapper, and mlx4_CLOSE_PORT_wrapper.
Currently, SRIOV is supported only for devices for which the
link-layer is IB on all ports; RoCE support will be implemented at later time.
Signed-off-by: Jack Morgenstein
---
drivers
This is necessary in order to support > 1 VF/PF in a VM for software that
uses the node guid as a discriminator, such as librdmacm.
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c | 14 ++
drivers/infiniband/hw/mlx4/mlx4_ib.h |3 +++
driv
1. Slave may not set the IS_SM capability for the port.
2. No DEV_MGR in multifunc mode.
Signed-off-by: Jack Morgenstein
---
drivers/net/ethernet/mellanox/mlx4/port.c | 10 ++
include/linux/mlx4/device.h |5 +
2 files changed, 15 insertions(+), 0 deletions
process context (due to 6 above, and also if smp_snoop
is used).
This fix was found and implemented by Saeed Mahameed
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c |6 +-
drivers/infiniband/hw/mlx4/main.c | 26
generate.
Signed-off-by: Oren Duer
Signed-off-by: Eli Cohen
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/Makefile |2 +-
drivers/infiniband/hw/mlx4/mad.c | 60 ++-
drivers/infiniband/hw/mlx4/main.c| 18 +-
drivers/infiniband/hw/mlx4/mcg.c | 1187
guests (and so these
IDs are not guaranteed to be globally unique).
The guest's comm ID is stored, and is returned to the response MAD when it
arrives.
Signed-off-by: Amir Vadai
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/Makefile |2 +-
drivers/infiniband/hw
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c| 162 +--
drivers/net/ethernet/mellanox/mlx4/fw.c |6 +
2 files changed, 161 insertions(+), 7 deletions(-)
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/m
ation ID in other CM messages
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c | 567 +-
1 files changed, 565 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c
index 2413a08..
changed before propagating an event to the slave).
To support this caching, added capability to __mlx4_ib_query_gid()
to obtain the network view (i.e., physical view) gid at index X,
not just the host (paravirtualized) view.
Based on a patch from Erez Shitrit
Signed-off-by: Jack Morgenstein
e for propagating port-management
events (client-reregister, etc) to slaves.
Signed-off-by: Jack Morgenstein
---
drivers/net/ethernet/mellanox/mlx4/eq.c | 237 +++--
drivers/net/ethernet/mellanox/mlx4/mlx4.h |1 +
include/linux/mlx4/device.h | 28
3
the other pkey indices mapped to the
reserved (invalid) pkey at index 127.
Finally, Add logic in smp_snoop for maintaining the phys_pkey_cache.
and generating events on the master only if a pkey actually changed.
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c
starts up. In addition, the master also
creates the real special QPs. (The ib_core layer on the master causes
creation of proxy special qp's, since the master is also
paravirtualized at the ib_core layer).
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c |
Reserve bits 26-31 for internal use by low-level drivers. Two
such bits are used in the mlx4 driver SRIOV IB implementation.
These enum additions guarantee that the core layer will never use
these bits, so that low level drivers may safely make use of them.
Signed-off-by: Jack Morgenstein
In addition, pass the proxy and tunnel QP numbers to slaves so
the driver can perform sqp paravirtualization.
Signed-off-by: Jack Morgenstein
---
drivers/net/ethernet/mellanox/mlx4/fw.c| 14 +
drivers/net/ethernet/mellanox/mlx4/fw.h|3 ++
drivers/net
membership bit).
Signed-off-by: Jack Morgenstein
Signed-off-by: Or Gerlitz
---
drivers/infiniband/core/cache.c | 32
include/rdma/ib_cache.h | 16
2 files changed, 48 insertions(+), 0 deletions(-)
diff --git a/drivers/infiniband/core/cache.c
pkey exists earlier in the table).
B. If the full member version is not in the table,
but the limited-member version is in the table,
return the index of the limited pkey.
Signed-off-by: Liran Liss
Signed-off-by: Jack Morgenstein
Signed-off-by: Or Gerlitz
---
drivers/infiniband/core/ca
is now rolled into V1 patch 22. This allowed us to eliminate
function
mlx4_ib_indexed_gid from patch 22 (replaced by using __mlx4_ib_query_gid()
from V0 patch 26).
Amir Vadai (1):
IB/mlx4: Add CM paravirtualization
Erez Shitrit (1):
IB/sa: Add GuidInfoRecord query support.
Jack Morgenstein (26):
net/mlx4
+0xad/0x100 [mlx4_ib]
[] ? is_module_address+0x23/0x30
[] handle_port_mgmt_change_event+0xeb/0x150 [mlx4_ib]
[] mlx4_ib_event+0x117/0x160 [mlx4_ib]
[] ? _raw_spin_lock_irqsave+0x61/0x70
[] mlx4_dispatch_event+0x6c/0x90 [mlx4_core]
[] mlx4_eq_int+0x500/0x950 [mlx4_core]
Reported by: Or Gerlitz
Signed-off-
Hello Roland,
I am back from vacation. Hopefully, you are progressing with your review of
the huge patch set I submitted -- please let me know if there is anything
I can do to help.
I noticed that the KVM group submitted a fix for the SRIOV IRQ guest problem
(Point #3 of the "Some Limitations" s
On Thursday 12 July 2012 05:13, Stephen Rothwell wrote:
> Hi all,
>
> Today's linux-next merge of the net-next tree got a conflict in
> include/linux/mlx4/device.h between commit 396f2feb05d7 ("mlx4_core:
> Implement mechanism for reserved Q_Keys") from the infiniband tree and
> commit 0ff1fb654be
On Thursday 12 July 2012 05:09, Stephen Rothwell wrote:
> Hi all,
>
> Today's linux-next merge of the net-next tree got a conflict in
> drivers/net/ethernet/mellanox/mlx4/main.c between commit 6634961c14d3
> ("mlx4: Put physical GID and P_Key table sizes in mlx4_phys_caps struct
> and paravirtuali
pkey exists earlier in the table).
B. If the full member version is not in the table,
but the limited-member version is in the table,
return the index of the limited pkey.
Signed-off-by: Liran Liss
Signed-off-by: Jack Morgenstein
Signed-off-by: Or Gerlitz
---
drivers/infiniband/core/ca
use of the 2 most-significant bits in enum
mlx4_ib_qp_flags (based upon enum ib_qp_create_flags (ib_verbs.h))
We need these bits in the low-level driver for internal purposes.
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/cq.c | 31 ++-
drivers/infiniband/hw/mlx4/mlx4_ib.h
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c| 162 +--
drivers/net/ethernet/mellanox/mlx4/fw.c |6 +
2 files changed, 161 insertions(+), 7 deletions(-)
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/m
generate.
Signed-off-by: Oren Duer
Signed-off-by: Eli Cohen
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/Makefile |2 +-
drivers/infiniband/hw/mlx4/mad.c | 60 ++-
drivers/infiniband/hw/mlx4/main.c| 18 +-
drivers/infiniband/hw/mlx4/mcg.c | 1187
ndex 10 to virtual index 1, that guest, whenever
it uses its pkey index 1 will actually be using the real pkey index 10.
Based on a patch from Erez Shitrit
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/Makefile |2 +-
drivers/infiniband/hw/mlx4/alias_GUID.c |6 +
needs to go through wrapper on master.
6. in mlx4_ib_event, PORT_MGMT_EVENT needs be handled in a work queue
on the master, since it propagates events to slaves using GEN_EQE
7. Do not support FMR on slaves.
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c |6
These macros will be reused by the mlx4 SRIOV-IB CM
paravirtualization code, and there is no reason to have
them declared both in the IB core in the mlx4 IB driver.
Signed-off-by: Jack Morgenstein
Signed-off-by: Or Gerlitz
---
drivers/infiniband/core/cm_msgs.h | 12
include/rdma
1. Slave may not set the IS_SM capability for the port.
2. No DEV_MGR in multifunc mode.
Signed-off-by: Jack Morgenstein
---
drivers/net/ethernet/mellanox/mlx4/port.c | 10 ++
include/linux/mlx4/device.h |5 +
2 files changed, 15 insertions(+), 0 deletions
Remove the error returns for IB ports from mlx4_ib_add,
mlx4_INIT_PORT_wrapper, and mlx4_CLOSE_PORT_wrapper.
Currently, SRIOV is supported only for devices for which the
link-layer is IB on all ports; RoCE support will be implemented at later time.
Signed-off-by: Jack Morgenstein
---
drivers
This is necessary in order to support more than one VF/PF on a VM for software
that uses the node guid as a discriminator, such as librdmacm.
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/mad.c | 14 ++
drivers/infiniband/hw/mlx4/mlx4_ib.h |3
receive.
When the real QP0 goes down, we indicate that the port is not active.
Signed-off-by: Jack Morgenstein
---
drivers/net/ethernet/mellanox/mlx4/fw.c | 71 +--
drivers/net/ethernet/mellanox/mlx4/qp.c | 38 +++--
2 files changed, 92 insertions(+), 17
changed before propagating an event to the slave).
To support this caching, added capability to __mlx4_ib_query_gid()
to obtain the network view (i.e., physical view) gid at index X,
not just the host (paravirtualized) view.
Based on a patch from Erez Shitrit
Signed-off-by: Jack Morgenstein
e for propagating port-management
events (client-reregister, etc) to slaves.
Signed-off-by: Jack Morgenstein
---
drivers/net/ethernet/mellanox/mlx4/eq.c | 237 +++--
drivers/net/ethernet/mellanox/mlx4/mlx4.h |1 +
include/linux/mlx4/device.h | 28
3
guests (and so these
IDs are not guaranteed to be globally unique).
The guest's comm ID is stored, and is returned to the response MAD when it
arrives.
Signed-off-by: Amir Vadai
Signed-off-by: Jack Morgenstein
---
drivers/infiniband/hw/mlx4/Makefile |2 +-
drivers/infiniband/hw
1 - 100 of 222 matches
Mail list logo