diff mbox series

Request to cherry-pick f00432063db1 to 5.10

Message ID CAMdnWFC4+-mEubOVkzaoqC5jnJCwY5hpcQtDnkmgqJ-mY5_GYg@mail.gmail.com
State New
Headers show
Series Request to cherry-pick f00432063db1 to 5.10 | expand

Commit Message

Meena Shanmugam May 11, 2022, 2:33 a.m. UTC
Hi all,

The commit f00432063db1a0db484e85193eccc6845435b80e upstream (SUNRPC:
Ensure we flush any closed sockets before xs_xprt_free()) fixes
CVE-2022-28893, hence good candidate for stable trees.
The above commit depends on 3be232f(SUNRPC: Prevent immediate
close+reconnect)  and  89f4249(SUNRPC: Don't call connect() more than
once on a TCP socket). Commit 3be232f depends on commit
e26d9972720e(SUNRPC: Clean up scheduling of autoclose).

Commits e26d9972720e, 3be232f, f00432063db1 apply cleanly on 5.10
kernel. commit 89f4249 didn't apply cleanly. I have patch for 89f4249
below.

Thanks,
Meena

From: Trond Myklebust <trond.myklebust@hammerspace.com>
Date: Wed, 16 Mar 2022 19:10:43 -0400
Subject: [PATCH] SUNRPC: Don't call connect() more than once on a TCP socket

commit 89f42494f92f448747bd8a7ab1ae8b5d5520577d upstream.

Avoid socket state races due to repeated calls to ->connect() using the
same socket. If connect() returns 0 due to the connection having
completed, but we are in fact in a closing state, then we may leave the
XPRT_CONNECTING flag set on the transport.

Reported-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Fixes: 3be232f11a3c ("SUNRPC: Prevent immediate close+reconnect")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
[meenashanmugam: Backported to 5.10: Fixed merge conflict in
xs_tcp_setup_socket]
Signed-off-by: Meena Shanmugam <meena.shanmugam@google.com>
---
 include/linux/sunrpc/xprtsock.h |  1 +
 net/sunrpc/xprtsock.c           | 21 +++++++++++----------
 2 files changed, 12 insertions(+), 10 deletions(-)

  goto out;
@@ -2294,6 +2298,7 @@ static void xs_tcp_setup_socket(struct work_struct *work)
  break;
  case 0:
  case -EINPROGRESS:
+ set_bit(XPRT_SOCK_CONNECT_SENT, &transport->sock_state);
  case -EALREADY:
  xprt_unlock_connect(xprt, transport);
  return;
@@ -2345,13 +2350,9 @@ static void xs_connect(struct rpc_xprt *xprt,
struct rpc_task *task)

  WARN_ON_ONCE(!xprt_lock_connect(xprt, task, transport));

- if (transport->sock != NULL && !xprt_connecting(xprt)) {
+ if (transport->sock != NULL) {
  dprintk("RPC:       xs_connect delayed xprt %p for %lu "
- "seconds\n",
- xprt, xprt->reestablish_timeout / HZ);
-
- /* Start by resetting any existing state */
- xs_reset_transport(transport);
+ "seconds\n", xprt, xprt->reestablish_timeout / HZ);

  delay = xprt_reconnect_delay(xprt);
  xprt_reconnect_backoff(xprt, XS_TCP_INIT_REEST_TO);

Comments

Greg KH May 13, 2022, 8:25 a.m. UTC | #1
On Thu, May 12, 2022 at 10:38:04AM -0700, Meena Shanmugam wrote:
> On Thu, May 12, 2022 at 9:23 AM Greg KH <gregkh@linuxfoundation.org> wrote:
> >
> > On Tue, May 10, 2022 at 07:33:23PM -0700, Meena Shanmugam wrote:
> > > Hi all,
> > >
> > > The commit f00432063db1a0db484e85193eccc6845435b80e upstream (SUNRPC:
> > > Ensure we flush any closed sockets before xs_xprt_free()) fixes
> > > CVE-2022-28893, hence good candidate for stable trees.
> > > The above commit depends on 3be232f(SUNRPC: Prevent immediate
> > > close+reconnect)  and  89f4249(SUNRPC: Don't call connect() more than
> > > once on a TCP socket). Commit 3be232f depends on commit
> > > e26d9972720e(SUNRPC: Clean up scheduling of autoclose).
> > >
> > > Commits e26d9972720e, 3be232f, f00432063db1 apply cleanly on 5.10
> > > kernel. commit 89f4249 didn't apply cleanly. I have patch for 89f4249
> > > below.
> >
> > We also need this for 5.15.y first, before we can apply it to 5.10.y.
> > Can you provide a working backport for that tree as well?
> >
> > And as others pointed out, your patch is totally corrupted and can not
> > be used, please fix your email client.
> >
> > thanks,
> >
> > greg k-h
> 
> For 5.15.y commit f00432063db1a0db484e85193eccc6845435b80e((SUNRPC:
> Ensure we flush any closed sockets before xs_xprt_free())) applies
> cleanly. The depend patch
> 3be232f(SUNRPC: Prevent immediate close+reconnect) also applies
> cleanly. Patch  89f4249
> (SUNRPC: Don't call connect() more than once on a TCP socket) is
> already present in 5.15.34 onwards.
> 
> Sorry about the patch corruption, I will fix it.

Sorry, but this did not work out at all, I get build errors when
attempting it for 5.10.y:

  CC [M]  net/sunrpc/xprtsock.o
net/sunrpc/xprtsock.c: In function ‘xs_tcp_setup_socket’:
net/sunrpc/xprtsock.c:2276:13: error: too few arguments to function ‘test_and_clear_bit’
 2276 |         if (test_and_clear_bit(XPRT_SOCK_CONNECT_SENT),
      |             ^~~~~~~~~~~~~~~~~~
In file included from ./arch/x86/include/asm/bitops.h:391,
                 from ./include/linux/bitops.h:29,
                 from ./include/linux/kernel.h:12,
                 from ./include/asm-generic/bug.h:20,
                 from ./arch/x86/include/asm/bug.h:93,
                 from ./include/linux/bug.h:5,
                 from ./include/linux/mmdebug.h:5,
                 from ./include/linux/gfp.h:5,
                 from ./include/linux/slab.h:15,
                 from net/sunrpc/xprtsock.c:24:
./include/asm-generic/bitops/instrumented-atomic.h:81:20: note: declared here
   81 | static inline bool test_and_clear_bit(long nr, volatile unsigned long *addr)
      |                    ^~~~~~~~~~~~~~~~~~
net/sunrpc/xprtsock.c:2276:55: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 2276 |         if (test_and_clear_bit(XPRT_SOCK_CONNECT_SENT),
      |                                                       ^
net/sunrpc/xprtsock.c:2312:17: warning: this statement may fall through [-Wimplicit-fallthrough=]
 2312 |                 set_bit(XPRT_SOCK_CONNECT_SENT, &transport->sock_state);
      |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
net/sunrpc/xprtsock.c:2313:9: note: here
 2313 |         case -EALREADY:
      |         ^~~~
make[2]: *** [scripts/Makefile.build:280: net/sunrpc/xprtsock.o] Error 1
make[1]: *** [scripts/Makefile.build:497: net/sunrpc] Error 2


And I am not quite sure what order you want me to apply things for 5.15.y.

So please, send me a properly backported series of patches for this for 5.15.y
and 5.10.y and I will be glad to pick them up.  Right now I'm just confused as
this was obviously not tested at all :(

thanks,

greg k-h
Meena Shanmugam May 13, 2022, 6:10 p.m. UTC | #2
I tested my original patch(which was corrupted by email client).
When the patch is manually backported to fix white space, the patch
was fixed wrongly :(
I sent my original patch again for 5.10.y without any corruption.
Sorry for the inconvenience caused.

For 5.15.y, this is the cherry-pick order:

3be232f11a3cc9b0ef0795e39fa11bdb8e422a06(SUNRPC: Prevent immediate
close+reconnect)
f00432063db1a0db484e85193eccc6845435b80e(SUNRPC: Ensure we flush any
closed sockets before xs_xprt_free())

Thanks,
Meena

On Fri, May 13, 2022 at 1:25 AM Greg KH <gregkh@linuxfoundation.org> wrote:
>
> On Thu, May 12, 2022 at 10:38:04AM -0700, Meena Shanmugam wrote:
> > On Thu, May 12, 2022 at 9:23 AM Greg KH <gregkh@linuxfoundation.org> wrote:
> > >
> > > On Tue, May 10, 2022 at 07:33:23PM -0700, Meena Shanmugam wrote:
> > > > Hi all,
> > > >
> > > > The commit f00432063db1a0db484e85193eccc6845435b80e upstream (SUNRPC:
> > > > Ensure we flush any closed sockets before xs_xprt_free()) fixes
> > > > CVE-2022-28893, hence good candidate for stable trees.
> > > > The above commit depends on 3be232f(SUNRPC: Prevent immediate
> > > > close+reconnect)  and  89f4249(SUNRPC: Don't call connect() more than
> > > > once on a TCP socket). Commit 3be232f depends on commit
> > > > e26d9972720e(SUNRPC: Clean up scheduling of autoclose).
> > > >
> > > > Commits e26d9972720e, 3be232f, f00432063db1 apply cleanly on 5.10
> > > > kernel. commit 89f4249 didn't apply cleanly. I have patch for 89f4249
> > > > below.
> > >
> > > We also need this for 5.15.y first, before we can apply it to 5.10.y.
> > > Can you provide a working backport for that tree as well?
> > >
> > > And as others pointed out, your patch is totally corrupted and can not
> > > be used, please fix your email client.
> > >
> > > thanks,
> > >
> > > greg k-h
> >
> > For 5.15.y commit f00432063db1a0db484e85193eccc6845435b80e((SUNRPC:
> > Ensure we flush any closed sockets before xs_xprt_free())) applies
> > cleanly. The depend patch
> > 3be232f(SUNRPC: Prevent immediate close+reconnect) also applies
> > cleanly. Patch  89f4249
> > (SUNRPC: Don't call connect() more than once on a TCP socket) is
> > already present in 5.15.34 onwards.
> >
> > Sorry about the patch corruption, I will fix it.
>
> Sorry, but this did not work out at all, I get build errors when
> attempting it for 5.10.y:
>
>   CC [M]  net/sunrpc/xprtsock.o
> net/sunrpc/xprtsock.c: In function ‘xs_tcp_setup_socket’:
> net/sunrpc/xprtsock.c:2276:13: error: too few arguments to function ‘test_and_clear_bit’
>  2276 |         if (test_and_clear_bit(XPRT_SOCK_CONNECT_SENT),
>       |             ^~~~~~~~~~~~~~~~~~
> In file included from ./arch/x86/include/asm/bitops.h:391,
>                  from ./include/linux/bitops.h:29,
>                  from ./include/linux/kernel.h:12,
>                  from ./include/asm-generic/bug.h:20,
>                  from ./arch/x86/include/asm/bug.h:93,
>                  from ./include/linux/bug.h:5,
>                  from ./include/linux/mmdebug.h:5,
>                  from ./include/linux/gfp.h:5,
>                  from ./include/linux/slab.h:15,
>                  from net/sunrpc/xprtsock.c:24:
> ./include/asm-generic/bitops/instrumented-atomic.h:81:20: note: declared here
>    81 | static inline bool test_and_clear_bit(long nr, volatile unsigned long *addr)
>       |                    ^~~~~~~~~~~~~~~~~~
> net/sunrpc/xprtsock.c:2276:55: warning: left-hand operand of comma expression has no effect [-Wunused-value]
>  2276 |         if (test_and_clear_bit(XPRT_SOCK_CONNECT_SENT),
>       |                                                       ^
> net/sunrpc/xprtsock.c:2312:17: warning: this statement may fall through [-Wimplicit-fallthrough=]
>  2312 |                 set_bit(XPRT_SOCK_CONNECT_SENT, &transport->sock_state);
>       |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> net/sunrpc/xprtsock.c:2313:9: note: here
>  2313 |         case -EALREADY:
>       |         ^~~~
> make[2]: *** [scripts/Makefile.build:280: net/sunrpc/xprtsock.o] Error 1
> make[1]: *** [scripts/Makefile.build:497: net/sunrpc] Error 2
>
>
> And I am not quite sure what order you want me to apply things for 5.15.y.
>
> So please, send me a properly backported series of patches for this for 5.15.y
> and 5.10.y and I will be glad to pick them up.  Right now I'm just confused as
> this was obviously not tested at all :(
>
> thanks,
>
> greg k-h
Meena Shanmugam May 13, 2022, 6:14 p.m. UTC | #3
On Fri, May 13, 2022 at 1:25 AM Greg KH <gregkh@linuxfoundation.org> wrote:
>
> On Thu, May 12, 2022 at 10:38:04AM -0700, Meena Shanmugam wrote:
> > On Thu, May 12, 2022 at 9:23 AM Greg KH <gregkh@linuxfoundation.org> wrote:
> > >
> > > On Tue, May 10, 2022 at 07:33:23PM -0700, Meena Shanmugam wrote:
> > > > Hi all,
> > > >
> > > > The commit f00432063db1a0db484e85193eccc6845435b80e upstream (SUNRPC:
> > > > Ensure we flush any closed sockets before xs_xprt_free()) fixes
> > > > CVE-2022-28893, hence good candidate for stable trees.
> > > > The above commit depends on 3be232f(SUNRPC: Prevent immediate
> > > > close+reconnect)  and  89f4249(SUNRPC: Don't call connect() more than
> > > > once on a TCP socket). Commit 3be232f depends on commit
> > > > e26d9972720e(SUNRPC: Clean up scheduling of autoclose).
> > > >
> > > > Commits e26d9972720e, 3be232f, f00432063db1 apply cleanly on 5.10
> > > > kernel. commit 89f4249 didn't apply cleanly. I have patch for 89f4249
> > > > below.
> > >
> > > We also need this for 5.15.y first, before we can apply it to 5.10.y.
> > > Can you provide a working backport for that tree as well?
> > >
> > > And as others pointed out, your patch is totally corrupted and can not
> > > be used, please fix your email client.
> > >
> > > thanks,
> > >
> > > greg k-h
> >
> > For 5.15.y commit f00432063db1a0db484e85193eccc6845435b80e((SUNRPC:
> > Ensure we flush any closed sockets before xs_xprt_free())) applies
> > cleanly. The depend patch
> > 3be232f(SUNRPC: Prevent immediate close+reconnect) also applies
> > cleanly. Patch  89f4249
> > (SUNRPC: Don't call connect() more than once on a TCP socket) is
> > already present in 5.15.34 onwards.
> >
> > Sorry about the patch corruption, I will fix it.
>
> Sorry, but this did not work out at all, I get build errors when
> attempting it for 5.10.y:
>
>   CC [M]  net/sunrpc/xprtsock.o
> net/sunrpc/xprtsock.c: In function ‘xs_tcp_setup_socket’:
> net/sunrpc/xprtsock.c:2276:13: error: too few arguments to function ‘test_and_clear_bit’
>  2276 |         if (test_and_clear_bit(XPRT_SOCK_CONNECT_SENT),
>       |             ^~~~~~~~~~~~~~~~~~
> In file included from ./arch/x86/include/asm/bitops.h:391,
>                  from ./include/linux/bitops.h:29,
>                  from ./include/linux/kernel.h:12,
>                  from ./include/asm-generic/bug.h:20,
>                  from ./arch/x86/include/asm/bug.h:93,
>                  from ./include/linux/bug.h:5,
>                  from ./include/linux/mmdebug.h:5,
>                  from ./include/linux/gfp.h:5,
>                  from ./include/linux/slab.h:15,
>                  from net/sunrpc/xprtsock.c:24:
> ./include/asm-generic/bitops/instrumented-atomic.h:81:20: note: declared here
>    81 | static inline bool test_and_clear_bit(long nr, volatile unsigned long *addr)
>       |                    ^~~~~~~~~~~~~~~~~~
> net/sunrpc/xprtsock.c:2276:55: warning: left-hand operand of comma expression has no effect [-Wunused-value]
>  2276 |         if (test_and_clear_bit(XPRT_SOCK_CONNECT_SENT),
>       |                                                       ^
> net/sunrpc/xprtsock.c:2312:17: warning: this statement may fall through [-Wimplicit-fallthrough=]
>  2312 |                 set_bit(XPRT_SOCK_CONNECT_SENT, &transport->sock_state);
>       |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> net/sunrpc/xprtsock.c:2313:9: note: here
>  2313 |         case -EALREADY:
>       |         ^~~~
> make[2]: *** [scripts/Makefile.build:280: net/sunrpc/xprtsock.o] Error 1
> make[1]: *** [scripts/Makefile.build:497: net/sunrpc] Error 2
>
>
> And I am not quite sure what order you want me to apply things for 5.15.y.
>
> So please, send me a properly backported series of patches for this for 5.15.y
> and 5.10.y and I will be glad to pick them up.  Right now I'm just confused as
> this was obviously not tested at all :(
>
> thanks,
>
> greg k-h

I tested my original patch(which was corrupted by email client).
When the patch is manually backported to fix white space, the patch
was fixed wrongly :(
I sent my original patch again for 5.10.y without any corruption.
Sorry for the inconvenience caused.

For 5.15.y, this is the cherry-pick order:

3be232f11a3cc9b0ef0795e39fa11bdb8e422a06(SUNRPC: Prevent immediate
close+reconnect)
f00432063db1a0db484e85193eccc6845435b80e(SUNRPC: Ensure we flush any
closed sockets before xs_xprt_free())

Thanks,
Meena
Greg KH May 16, 2022, 12:42 p.m. UTC | #4
On Fri, May 13, 2022 at 11:14:51AM -0700, Meena Shanmugam wrote:
> On Fri, May 13, 2022 at 1:25 AM Greg KH <gregkh@linuxfoundation.org> wrote:
> >
> > On Thu, May 12, 2022 at 10:38:04AM -0700, Meena Shanmugam wrote:
> > > On Thu, May 12, 2022 at 9:23 AM Greg KH <gregkh@linuxfoundation.org> wrote:
> > > >
> > > > On Tue, May 10, 2022 at 07:33:23PM -0700, Meena Shanmugam wrote:
> > > > > Hi all,
> > > > >
> > > > > The commit f00432063db1a0db484e85193eccc6845435b80e upstream (SUNRPC:
> > > > > Ensure we flush any closed sockets before xs_xprt_free()) fixes
> > > > > CVE-2022-28893, hence good candidate for stable trees.
> > > > > The above commit depends on 3be232f(SUNRPC: Prevent immediate
> > > > > close+reconnect)  and  89f4249(SUNRPC: Don't call connect() more than
> > > > > once on a TCP socket). Commit 3be232f depends on commit
> > > > > e26d9972720e(SUNRPC: Clean up scheduling of autoclose).
> > > > >
> > > > > Commits e26d9972720e, 3be232f, f00432063db1 apply cleanly on 5.10
> > > > > kernel. commit 89f4249 didn't apply cleanly. I have patch for 89f4249
> > > > > below.
> > > >
> > > > We also need this for 5.15.y first, before we can apply it to 5.10.y.
> > > > Can you provide a working backport for that tree as well?
> > > >
> > > > And as others pointed out, your patch is totally corrupted and can not
> > > > be used, please fix your email client.
> > > >
> > > > thanks,
> > > >
> > > > greg k-h
> > >
> > > For 5.15.y commit f00432063db1a0db484e85193eccc6845435b80e((SUNRPC:
> > > Ensure we flush any closed sockets before xs_xprt_free())) applies
> > > cleanly. The depend patch
> > > 3be232f(SUNRPC: Prevent immediate close+reconnect) also applies
> > > cleanly. Patch  89f4249
> > > (SUNRPC: Don't call connect() more than once on a TCP socket) is
> > > already present in 5.15.34 onwards.
> > >
> > > Sorry about the patch corruption, I will fix it.
> >
> > Sorry, but this did not work out at all, I get build errors when
> > attempting it for 5.10.y:
> >
> >   CC [M]  net/sunrpc/xprtsock.o
> > net/sunrpc/xprtsock.c: In function ‘xs_tcp_setup_socket’:
> > net/sunrpc/xprtsock.c:2276:13: error: too few arguments to function ‘test_and_clear_bit’
> >  2276 |         if (test_and_clear_bit(XPRT_SOCK_CONNECT_SENT),
> >       |             ^~~~~~~~~~~~~~~~~~
> > In file included from ./arch/x86/include/asm/bitops.h:391,
> >                  from ./include/linux/bitops.h:29,
> >                  from ./include/linux/kernel.h:12,
> >                  from ./include/asm-generic/bug.h:20,
> >                  from ./arch/x86/include/asm/bug.h:93,
> >                  from ./include/linux/bug.h:5,
> >                  from ./include/linux/mmdebug.h:5,
> >                  from ./include/linux/gfp.h:5,
> >                  from ./include/linux/slab.h:15,
> >                  from net/sunrpc/xprtsock.c:24:
> > ./include/asm-generic/bitops/instrumented-atomic.h:81:20: note: declared here
> >    81 | static inline bool test_and_clear_bit(long nr, volatile unsigned long *addr)
> >       |                    ^~~~~~~~~~~~~~~~~~
> > net/sunrpc/xprtsock.c:2276:55: warning: left-hand operand of comma expression has no effect [-Wunused-value]
> >  2276 |         if (test_and_clear_bit(XPRT_SOCK_CONNECT_SENT),
> >       |                                                       ^
> > net/sunrpc/xprtsock.c:2312:17: warning: this statement may fall through [-Wimplicit-fallthrough=]
> >  2312 |                 set_bit(XPRT_SOCK_CONNECT_SENT, &transport->sock_state);
> >       |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > net/sunrpc/xprtsock.c:2313:9: note: here
> >  2313 |         case -EALREADY:
> >       |         ^~~~
> > make[2]: *** [scripts/Makefile.build:280: net/sunrpc/xprtsock.o] Error 1
> > make[1]: *** [scripts/Makefile.build:497: net/sunrpc] Error 2
> >
> >
> > And I am not quite sure what order you want me to apply things for 5.15.y.
> >
> > So please, send me a properly backported series of patches for this for 5.15.y
> > and 5.10.y and I will be glad to pick them up.  Right now I'm just confused as
> > this was obviously not tested at all :(
> >
> > thanks,
> >
> > greg k-h
> 
> I tested my original patch(which was corrupted by email client).
> When the patch is manually backported to fix white space, the patch
> was fixed wrongly :(
> I sent my original patch again for 5.10.y without any corruption.
> Sorry for the inconvenience caused.
> 
> For 5.15.y, this is the cherry-pick order:
> 
> 3be232f11a3cc9b0ef0795e39fa11bdb8e422a06(SUNRPC: Prevent immediate
> close+reconnect)

Already in 5.15.34.

> f00432063db1a0db484e85193eccc6845435b80e(SUNRPC: Ensure we flush any
> closed sockets before xs_xprt_free())

Now queued up, thanks.

greg k-h
diff mbox series

Patch

diff --git a/include/linux/sunrpc/xprtsock.h b/include/linux/sunrpc/xprtsock.h
index 8c2a712cb242..689062afdd61 100644
--- a/include/linux/sunrpc/xprtsock.h
+++ b/include/linux/sunrpc/xprtsock.h
@@ -89,5 +89,6 @@  struct sock_xprt {
 #define XPRT_SOCK_WAKE_WRITE (5)
 #define XPRT_SOCK_WAKE_PENDING (6)
 #define XPRT_SOCK_WAKE_DISCONNECT (7)
+#define XPRT_SOCK_CONNECT_SENT (8)

 #endif /* _LINUX_SUNRPC_XPRTSOCK_H */
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index 60c58eb9a456..33a81f9703b1 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -2260,10 +2260,14 @@  static void xs_tcp_setup_socket(struct
work_struct *work)
  struct rpc_xprt *xprt = &transport->xprt;
  int status = -EIO;

- if (!sock) {
- sock = xs_create_sock(xprt, transport,
- xs_addr(xprt)->sa_family, SOCK_STREAM,
- IPPROTO_TCP, true);
+ if (xprt_connected(xprt))
+ goto out;
+ if (test_and_clear_bit(XPRT_SOCK_CONNECT_SENT,
+        &transport->sock_state) ||
+     !sock) {
+ xs_reset_transport(transport);
+ sock = xs_create_sock(xprt, transport, xs_addr(xprt)->sa_family,
+       SOCK_STREAM, IPPROTO_TCP, true);
  if (IS_ERR(sock)) {
  status = PTR_ERR(sock);