On Thu, Jan 07, 2016 at 09:28:08AM +0200, Erez Shitrit wrote:
>
> ipoib_mcast_restart_task calls ipoib_mcast_remove_list with the
> parameter mcast->dev. That mcast is a temporary (used as an iterator)
> variable that may be uninitialized.
> There is no need to send the variable dev to the functio
On Wed, Jan 06, 2016 at 02:25:52PM +0200, Haggai Eran wrote:
> On 06/01/2016 13:03, Yuval Shaia wrote:
> > On Thu, Dec 31, 2015 at 05:34:03PM +0200, Haggai Eran wrote:
> >>> + sock = sockfd_lookup(fd, &rc);
> >>> + if (IS_ERR_OR_NULL(sock))
> >>> +
On Thu, Dec 31, 2015 at 05:34:03PM +0200, Haggai Eran wrote:
> > + sock = sockfd_lookup(fd, &rc);
> > + if (IS_ERR_OR_NULL(sock))
> > + return -EINVAL;
> > +
> > + inetsock = inet_sk(sock->sk);
> > +
> > + neigh = neigh_lookup(&arp_tbl, &inetsock->inet_daddr, dev);
>
> Also, isn'
> > +
> > + gid = (union ib_gid *)(neigh->ha + 4);
> If you just take the GID from the hardware address why not do that from
> userspace?
Will do.
Thanks.
> Why do you need a new ioctl to do that for you?
>
>
> Regards,
> Haggai
>
> --
> To unsubscribe from this list: send the line "unsubscri
On Thu, Dec 31, 2015 at 05:31:44PM +0200, Or Gerlitz wrote:
>
> wait (2nd)
>
> what prevents you from getting this info in user space through
> netlink from the kernel rtnl routing/neighbour services?
Thanks.
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
>
To support security applications, that need to filter out connections based
on SGID, an ioctl command to retrieve SGID of a given socket is added.
Signed-off-by: Yuval Shaia
---
drivers/infiniband/ulp/ipoib/Makefile |3 +-
drivers/infiniband/ulp/ipoib/ipoib.h | 11 ++
drivers
On Wed, Nov 18, 2015 at 02:25:23PM +0100, Petr Mladek wrote:
> Kthreads are currently implemented as an infinite loop. Each
> has its own variant of checks for terminating, freezing,
> awakening. In many cases it is unclear to say in which state
> it is and sometimes it is done a wrong way.
>
> Th
On Thu, Jul 30, 2015 at 03:00:18PM -0600, Jason Gunthorpe wrote:
> On Thu, Jul 30, 2015 at 11:46:36PM +0300, Yuval Shaia wrote:
> > On Thu, Jul 30, 2015 at 11:15:38AM -0600, Jason Gunthorpe wrote:
> > > On Thu, Jul 30, 2015 at 11:51:12AM -0400, Doug Ledford wrote:
> > &
On Thu, Jul 30, 2015 at 03:58:13PM +0200, Yann Droneaud wrote:
> Hi,
>
> Le jeudi 30 juillet 2015 à 04:46 -0700, Yuval Shaia a écrit :
> > This enhancement suggest the usage of IB CRC instead of CSUM in IPoIB
> > CM. IPoIB CM uses RC (Reliable Connection) which guarantee
On Thu, Jul 30, 2015 at 07:03:22PM -0700, Bart Van Assche wrote:
> On 07/30/15 13:09, Yuval Shaia wrote:
> >On Thu, Jul 30, 2015 at 09:38:54AM -0700, Bart Van Assche wrote:
> >>On 07/30/2015 04:46 AM, Yuval Shaia wrote:
> >>> struct ipoib_cm_data {
> >>>
On Tue, Nov 17, 2015 at 10:11:17AM +0200, Or Gerlitz wrote:
> On 11/16/2015 8:54 PM, Yuval Shaia wrote:
> >-test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags))
> >+likely(test_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags)))
>
> this hunk has nothing to do
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
On Thu, Nov 05, 2015 at 03:10:35AM -0800, Christoph Hellwig wrote:
> On Thu, Nov 05, 2015 at 01:07:49PM +0200, Yuval Shaia wrote:
> > > What do you plan to use it for? We're aiming hard to get rid of FMRs
> > > from the whole kernel stack ASAP.
> > We have two dr
On Thu, Nov 05, 2015 at 02:51:28AM -0800, Christoph Hellwig wrote:
> On Thu, Nov 05, 2015 at 12:48:45PM +0200, Yuval Shaia wrote:
> > On Thu, Oct 29, 2015 at 07:21:45PM +0200, Sagi Grimberg wrote:
> > > Hi Yuval,
> > >
> > > The title prefix should be IB/mlx4
On Thu, Oct 29, 2015 at 07:21:45PM +0200, Sagi Grimberg wrote:
> Hi Yuval,
>
> The title prefix should be IB/mlx4:
>
> >Expose max_fmr so it will be available to ULPs.
> >max_fmr is num_mpts minus reserved.
> >
> >Signed-off-by: Yuval Shaia
> >---
>
Expose max_fmr so it will be available to ULPs.
max_fmr is num_mpts minus reserved.
Signed-off-by: Yuval Shaia
---
drivers/infiniband/hw/mlx4/main.c |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/drivers/infiniband/hw/mlx4/main.c
b/drivers/infiniband/hw/mlx4/main.c
On Thu, Sep 03, 2015 at 04:29:45PM -0400, Doug Ledford wrote:
> On 07/30/2015 10:03 PM, Bart Van Assche wrote:
> > On 07/30/15 13:09, Yuval Shaia wrote:
> >> On Thu, Jul 30, 2015 at 09:38:54AM -0700, Bart Van Assche wrote:
> >>> On 07/30/2015 04:46 AM, Yuva
On Sun, Aug 16, 2015 at 07:05:30PM +0300, Eran Ben Elisha wrote:
> From: Maor Gottlieb
>
> Set the mcast loopback prevention bit in the QPC for ETH MLX QPs (not
> RSS QPs), when the firmware supports this feature. In addition, all rx
> ring QPs need to be updated in order not to enforce loopback
On Thu, Jul 30, 2015 at 11:15:38AM -0600, Jason Gunthorpe wrote:
> On Thu, Jul 30, 2015 at 11:51:12AM -0400, Doug Ledford wrote:
>
> > In its current state, I have my doubts about this patch. However, it
> > seems to me that this should be relatively easy to fix in such a way
> > that you get 90%
On Thu, Jul 30, 2015 at 11:51:12AM -0400, Doug Ledford wrote:
> On 07/30/2015 11:20 AM, Yuval Shaia wrote:
> > On Thu, Jul 30, 2015 at 03:58:13PM +0200, Yann Droneaud wrote:
> >> Hi,
> >>
> >> Le jeudi 30 juillet 2015 à 04:46 -0700, Yuval Shaia a écrit :
> &g
On Thu, Jul 30, 2015 at 09:38:54AM -0700, Bart Van Assche wrote:
> On 07/30/2015 04:46 AM, Yuval Shaia wrote:
> > struct ipoib_cm_data {
> > __be32 qpn; /* High byte MUST be ignored on receive */
> > __be32 mtu;
> >+__be16 sig; /* must be IPOIB_CM_PROTO
On Thu, Jul 30, 2015 at 03:58:13PM +0200, Yann Droneaud wrote:
> Hi,
>
> Le jeudi 30 juillet 2015 à 04:46 -0700, Yuval Shaia a écrit :
> > This enhancement suggest the usage of IB CRC instead of CSUM in IPoIB
> > CM. IPoIB CM uses RC (Reliable Connection) which guarantee
calculate
checksum before transmiting the packet in case the peer does not support
this feature.
With this enhancement throughput is increased by 60%.
Signed-off-by: Yuval Shaia
---
drivers/infiniband/ulp/ipoib/ipoib.h | 26
drivers/infiniband/ulp/ipoib/ipoib_cm.c | 46
etween CSUM and SG features.")
Signed-off-by: Yuval Shaia
Acked-by: Christian Marie
---
v5:
* Move ipoib_build_sge to ipoib.h
v4:
* Use 12 hex-digits commit identifier
* Add Acked-by: Christian Marie
* Merge ipoib_cm_dma_unmap_tx and ipoib_dma_unmap_tx
* Take out common code from ud&
etween CSUM and SG features.")
Signed-off-by: Yuval Shaia
Acked-by: Christian Marie
---
v4:
* Use 12 hex-digits commit identifier
* Add Acked-by: Christian Marie
* Merge ipoib_cm_dma_unmap_tx and ipoib_dma_unmap_tx
* Take out common code from ud's post_send and cm's post_send
v3:
On Wed, Jun 10, 2015 at 04:06:18PM +0300, Erez Shitrit wrote:
> > + if (unlikely(ipoib_dma_map_tx(priv->ca, tx_req))) {
> > + ++dev->stats.tx_errors;
> > + dev_kfree_skb_any(skb);
> > + return;
> > + }
>
ping
On Sun, Feb 22, 2015 at 08:50:27PM -0800, Yuval Shaia wrote:
> Current approach force one to implement all ops even when some functions can
> use the default implementation.
> As a result, for new DMA ops (e.x new arch) many functions just wrap the
> default function.
> The
On Mon, Mar 23, 2015 at 11:17:49AM -0600, Jason Gunthorpe wrote:
> On Sun, Mar 22, 2015 at 11:21:50AM +0200, Yuval Shaia wrote:
> > On Sun, Mar 15, 2015 at 05:16:16PM +0200, Yuval Shaia wrote:
> > > Hi,
> > > I didn't got any further comments on this one.
> >
On Wed, Apr 01, 2015 at 01:17:19PM -0400, ira.weiny wrote:
> On Mon, Mar 23, 2015 at 11:17:49AM -0600, Jason Gunthorpe wrote:
> > On Sun, Mar 22, 2015 at 11:21:50AM +0200, Yuval Shaia wrote:
> > > On Sun, Mar 15, 2015 at 05:16:16PM +0200, Yuval Shaia wrote:
> > > >
On Sun, Mar 15, 2015 at 05:16:16PM +0200, Yuval Shaia wrote:
> Hi,
> I didn't got any further comments on this one.
> Any idea why SG in CM is un-welcome?
By mistake I sent a private mail only.
Cc: Roland Dreier
Cc: Sean Hefty
Cc: Hal Rosenstock
Your advice would be very appre
be override.
Signed-off-by: Yuval Shaia
---
include/rdma/ib_verbs.h | 22 +++---
1 files changed, 11 insertions(+), 11 deletions(-)
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 0d74f1d..166c01a 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma
On Tue, Feb 17, 2015 at 02:47:03PM +0200, Matan Barak wrote:
>
>
> On 2/16/2015 7:32 PM, Yuval Shaia wrote:
> >ping
> >
> >On Mon, Jan 26, 2015 at 06:30:23AM -0800, Yuval Shaia wrote:
> >>1. Add descriptions of all features.
> >>2. Add indication
ping
On Mon, Jan 26, 2015 at 06:30:23AM -0800, Yuval Shaia wrote:
> 1. Add descriptions of all features.
> 2. Add indication whether feature is supported or not.
> With this enhancement all device features will be printed along with flags
> that shows if feature is supported by F
On Wed, Jan 28, 2015 at 12:36:27PM +0100, Yann Droneaud wrote:
> Hi,
>
> Le mardi 27 janvier 2015 à 03:21 -0800, Yuval Shaia a écrit :
> > With this fix Scatter-Gather will be supported also in connected mode
> >
>
> Please explain the issue with current kernel
On Mon, Dec 15, 2014 at 06:14:31PM +0200, Or Gerlitz wrote:
> >>
> >> NO, as Matan wrote you "except for Or's comments" - we do want to dump
> > Sure, was not ignoring this one, it was just a warn :)
> >> what is supported by both the device (firmware) and the driver, and
> >> not more. A subset of
With this fix Scatter-Gather will be supported also in connected mode
Signed-off-by: Yuval Shaia
---
drivers/infiniband/ulp/ipoib/ipoib.h |8 +--
drivers/infiniband/ulp/ipoib/ipoib_cm.c | 107 +++--
drivers/infiniband/ulp/ipoib/ipoib_ib.c |3 +-
drivers
1. Add descriptions of all features.
2. Add indication whether feature is supported or not.
With this enhancement all device features will be printed along with flags that
shows if feature is supported by FW, Driver or both.
Signed-off-by: Yuval Shaia
---
drivers/net/ethernet/mellanox/mlx4
1. Add descriptions of all features.
2. Add indication whether feature is supported or not.
With this enhancement all device features will be printed along with flags that
shows if feature is supported by FW, Driver or both.
Signed-off-by: Yuval Shaia
---
drivers/net/ethernet/mellanox/mlx4
> >> >- mlx4_dbg(dev, "%s\n", fname[i]);
> >> >+ if (fname[i])
> >> >+ mlx4_dbg(dev, "(%c) %s\n", (flags & (1LL <<
> >> >i)) ?
> >> >+'*' : ' ', fname[i]);
> >> > }
> >>
> >> I think that the strings
On Mon, Dec 15, 2014 at 06:14:31PM +0200, Or Gerlitz wrote:
> >> >> >+ [52] =RSS on fragmented IP datagram support",
> >> >>
> >> >> Hi,
> >> >>
> >> >> Except for Or's comments, [55] exists twice. Furthermore, please
> >> >> move [56]-[57] to be after [55] =Port link type sensing sup
On Mon, Dec 15, 2014 at 11:35:52AM +0200, Or Gerlitz wrote:
> On Mon, Dec 15, 2014 at 11:21 AM, Yuval Shaia wrote:
> > On Wed, Dec 10, 2014 at 12:10:07PM +0200, Matan Barak wrote:
> >>
> >>
> >> On 12/8/2014 1:48 AM, Yuval Shaia wrote:
> >> >1. Ad
1. Add indication whether feature is supported or not.
2. Add descriptions of all features.
Without this fix there is no way to tell if feature is not supported or that
description is not exist.
Signed-off-by: Yuval Shaia
---
drivers/net/ethernet/mellanox/mlx4/fw.c | 36
On Wed, Dec 10, 2014 at 12:10:07PM +0200, Matan Barak wrote:
>
>
> On 12/8/2014 1:48 AM, Yuval Shaia wrote:
> >1. Add indication whether feature is supported or not.
> >2. Add descriptions of all features.
> >Without this fix there is no way to tell if featu
On Wed, Dec 10, 2014 at 12:13:51PM +0200, Matan Barak wrote:
>
>
> On 12/7/2014 7:05 PM, Yuval Shaia wrote:
> >This patch is merely makes code more nice and readable.
> >Instead of checking for DPDP on every loop cycle the check moves out of the
> >loop.
> &
The check for DPDP moves out of the loop to make the code more correct and
readable
Signed-off-by: Yuval Shaia
---
drivers/net/ethernet/mellanox/mlx4/main.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c
b/drivers/net
On Tue, Dec 09, 2014 at 07:47:27PM +0200, Or Gerlitz wrote:
> On Sun, Dec 7, 2014 at 7:05 PM, Yuval Shaia wrote:
> > This patch is merely makes code more nice and readable.
> > Instead of checking for DPDP on every loop cycle the check moves out of the
> > loop.
>
>
On Tue, Dec 09, 2014 at 07:47:27PM +0200, Or Gerlitz wrote:
> On Sun, Dec 7, 2014 at 7:05 PM, Yuval Shaia wrote:
> > This patch is merely makes code more nice and readable.
> > Instead of checking for DPDP on every loop cycle the check moves out of the
> > loop.
>
>
On Tue, Dec 09, 2014 at 07:55:07PM +0200, Or Gerlitz wrote:
> On Mon, Dec 8, 2014 at 1:48 AM, Yuval Shaia wrote:
> > 1. Add indication whether feature is supported or not.
> > 2. Add descriptions of all features.
> > Without this fix there is no way to tell if featur
This value is used to calculate max_qp_dest_rdma.
Default value of 4 brings us to 16 while HW supports 128 (max_requester_per_qp).
Although this value can be changed by module param it is best that default be
optimized.
Acked-by: Wengang Wang
Signed-off-by: Yuval Shaia
---
drivers/net
1. Add indication whether feature is supported or not.
2. Add descriptions of all features.
Without this fix there is no way to tell if feature is not supported or that
description is not exist.
Signed-off-by: Yuval Shaia
---
drivers/net/ethernet/mellanox/mlx4/fw.c | 37
This patch is merely makes code more nice and readable.
Instead of checking for DPDP on every loop cycle the check moves out of the
loop.
Signed-off-by: Yuval Shaia
---
drivers/net/ethernet/mellanox/mlx4/main.c |7 +++
1 files changed, 3 insertions(+), 4 deletions(-)
diff --git a
On Sun, Dec 07, 2014 at 09:24:26PM -0800, Roland Dreier wrote:
> On Sat, Dec 6, 2014 at 10:46 AM, Yuval Shaia wrote:
> >
> > - for (i = 0; i < dev->caps.num_ports - 1; i++) {
> > - if (port_type[i] != port_type[i + 1]) {
> > -
Signed-off-by: Yuval Shaia
---
drivers/net/ethernet/mellanox/mlx4/main.c |7 +++
1 files changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c
b/drivers/net/ethernet/mellanox/mlx4/main.c
index 7920f5f..e3fd50a 100644
--- a/drivers/net/ethernet
On Wed, Oct 08, 2014 at 02:34:07PM -0400, Chuck Lever wrote:
> Hi Yuval-
>
> On Oct 4, 2014, at 7:45 AM, Yuval Shaia wrote:
>
> > This enhancement suggest the usage of IB CRC instead of CSUM in IPoIB CM.
> > IPoIB Connected Mode driver uses RC (Reliable Connection) w
On Sun, Oct 05, 2014 at 04:35:00PM -0600, Jason Gunthorpe wrote:
> On Sat, Oct 04, 2014 at 04:45:47AM -0700, Yuval Shaia wrote:
>
> > + if (cm_ibcrc_as_csum)
> > + skb->ip_summed = CHECKSUM_UNNECESSARY;
> > +
>
> As I said before, I think this scheme
Hi all,
Sorry but this patch will not go smoothly to latest kernel.
I'm currently adjusting it to fit to v3.17-rc4 and will send new one soon.
Yuval
On Sat, Oct 04, 2014 at 04:45:47AM -0700, Yuval Shaia wrote:
> This enhancement suggest the usage of IB CRC instead of CSUM in IPoIB CM.
will be able to calculate checksum before transmiting
the packet in case the other end does not support this feature.
A support for fragmented skb is added to transmit path.
Signed-off-by: Yuval Shaia
---
drivers/infiniband/ulp/ipoib/ipoib.h | 25 ++
drivers/infiniband/ulp/ipoib
On Wed, Oct 01, 2014 at 03:13:33PM +0300, Or Gerlitz wrote:
>
> On 10/1/2014 2:55 PM, Yuval Shaia wrote:
> >On transmit checksum calculation is done while copying buffer from
> >user-space so not much benefit
> but this copying can go away, right?
I assume this is exactly
On Tue, Sep 16, 2014 at 09:47:43AM +0300, Or Gerlitz wrote:
>
> On the post you wrote "[...] resolve the issue by removing the physically
> contiguous memory requirement using Scatter/Gather feature that exists in
> Linux".
>
> I assume you refer to NETIF_F_SG, right? so your claim is that Linu
On Sun, Sep 14, 2014 at 09:46:22PM +0300, Yuval Shaia wrote:
> Hi,
> Lately i was working on fixing an issue with IPoIB driver and i'd like to
> share the
> details with you.
>
> By default, IPoIB-CM driver uses 64k MTU. Larger MTU gives better performance.
> This
On Mon, Sep 22, 2014 at 10:28:02PM +0300, Yuval Shaia wrote:
> On Tue, Sep 16, 2014 at 09:47:43AM +0300, Or Gerlitz wrote:
> >
> > On 9/15/2014 9:55 PM, Yuval Shaia wrote:
> > >On Mon, Sep 15, 2014 at 05:47:19PM +0300, Or Gerlitz wrote:
> > >>>On Su
On Tue, Sep 16, 2014 at 09:47:43AM +0300, Or Gerlitz wrote:
>
> On 9/15/2014 9:55 PM, Yuval Shaia wrote:
> >On Mon, Sep 15, 2014 at 05:47:19PM +0300, Or Gerlitz wrote:
> >>>On Sun, Sep 14, 2014 at 9:46 PM, Yuval Shaia
> >>>wrote:
> >>>>>By d
On Mon, Sep 15, 2014 at 10:58:20AM -0600, Jason Gunthorpe wrote:
> On Mon, Sep 15, 2014 at 05:47:19PM +0300, Or Gerlitz wrote:
>
> > > [...] The proposal is to tell to network stack that IPoIB-CM supports IP
> > > Checksum offload. This enables Linux IPoIB-CM driver to use
> > > Scatter/Gather fe
On Mon, Sep 15, 2014 at 05:47:19PM +0300, Or Gerlitz wrote:
> On Sun, Sep 14, 2014 at 9:46 PM, Yuval Shaia wrote:
> > By default, IPoIB-CM driver uses 64k MTU. Larger MTU gives better
> > performance.
> > This MTU plus overhead puts the memory allocation for IP based packet
Hi,
Lately i was working on fixing an issue with IPoIB driver and i'd like to share
the
details with you.
By default, IPoIB-CM driver uses 64k MTU. Larger MTU gives better performance.
This MTU plus overhead puts the memory allocation for IP based packets at 32 4k
pages
(order 5), which have to
65 matches
Mail list logo