[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190711143302.GH25821@mellanox.com>
Date: Thu, 11 Jul 2019 14:33:07 +0000
From: Jason Gunthorpe <jgg@...lanox.com>
To: Bernard Metzler <BMT@...ich.ibm.com>
CC: Leon Romanovsky <leon@...nel.org>,
Stephen Rothwell <sfr@...b.auug.org.au>,
Doug Ledford <dledford@...hat.com>,
David Miller <davem@...emloft.net>,
Networking <netdev@...r.kernel.org>,
Linux Next Mailing List <linux-next@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: Re: Re: linux-next: build failure after merge of the net-next
tree
On Thu, Jul 11, 2019 at 12:29:21PM +0000, Bernard Metzler wrote:
>
> >To: "Bernard Metzler" <BMT@...ich.ibm.com>
> >From: "Jason Gunthorpe" <jgg@...lanox.com>
> >Date: 07/11/2019 01:53PM
> >Cc: "Leon Romanovsky" <leon@...nel.org>, "Stephen Rothwell"
> ><sfr@...b.auug.org.au>, "Doug Ledford" <dledford@...hat.com>, "David
> >Miller" <davem@...emloft.net>, "Networking" <netdev@...r.kernel.org>,
> >"Linux Next Mailing List" <linux-next@...r.kernel.org>, "Linux Kernel
> >Mailing List" <linux-kernel@...r.kernel.org>
> >Subject: [EXTERNAL] Re: Re: linux-next: build failure after merge of
> >the net-next tree
> >
> >On Thu, Jul 11, 2019 at 08:00:49AM +0000, Bernard Metzler wrote:
> >
> >> That listen will not sleep. The socket is just marked
> >> listening.
> >
> >Eh? siw_listen_address() calls siw_cep_alloc() which does:
> >
> > struct siw_cep *cep = kzalloc(sizeof(*cep), GFP_KERNEL);
> >
> >Which is sleeping. Many other cases too.
> >
> >Jason
> >
> >
> Ah, true! I was after really deep sleeps like user level
> socket accept() calls ;) So you are correct of course.
I've added this patch to the rdma tree to fix the missing locking.
The merge resolution will be simply swapping
for_ifa to in_dev_for_each_ifa_rtnl.
Jason
>From c421651fa2295d1219c36674c7eb8c574542ceea Mon Sep 17 00:00:00 2001
From: Jason Gunthorpe <jgg@...lanox.com>
Date: Thu, 11 Jul 2019 11:29:42 -0300
Subject: [PATCH] RDMA/siw: Add missing rtnl_lock around access to ifa
ifa is protected by rcu or rtnl, add the missing locking. In this case we
have to use rtnl since siw_listen_address() is sleeping.
Fixes: 6c52fdc244b5 ("rdma/siw: connection management")
Reviewed-by: Bernard Metzler <bmt@...ich.ibm.com>
Signed-off-by: Jason Gunthorpe <jgg@...lanox.com>
---
drivers/infiniband/sw/siw/siw_cm.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/infiniband/sw/siw/siw_cm.c b/drivers/infiniband/sw/siw/siw_cm.c
index 8e618cb7261f62..c25be723c15b64 100644
--- a/drivers/infiniband/sw/siw/siw_cm.c
+++ b/drivers/infiniband/sw/siw/siw_cm.c
@@ -1975,6 +1975,7 @@ int siw_create_listen(struct iw_cm_id *id, int backlog)
id, &s_laddr.sin_addr, ntohs(s_laddr.sin_port),
&s_raddr->sin_addr, ntohs(s_raddr->sin_port));
+ rtnl_lock();
for_ifa(in_dev)
{
if (ipv4_is_zeronet(s_laddr.sin_addr.s_addr) ||
@@ -1989,6 +1990,7 @@ int siw_create_listen(struct iw_cm_id *id, int backlog)
}
}
endfor_ifa(in_dev);
+ rtnl_unlock();
in_dev_put(in_dev);
} else if (id->local_addr.ss_family == AF_INET6) {
struct inet6_dev *in6_dev = in6_dev_get(dev);
--
2.21.0
Powered by blists - more mailing lists