Re: ssh_handle_packets blocking issue

[Mikhail Kulinich <tysonite@xxxxxxxxx>]

Hello Aris,

Thanks for your feedback!

I will try to test a new libssh 0.5 against this issue. However, it can take
a time. As I understand, interfaces were changed very much, and,
unfortunately, I have code (already legacy) which needs to be checked and
changed.

The content is correct and it is a first call to the read_channel. There are
no other parallel sessions.
I can reproduce this issue 100% with libssh 0.4.4.

FYI, along time ago I changed this behavior with the patch for channel_read
function:
{{{
...
    rc = ssh_socket_poll(channel->session->socket, &w, &e);
*    if (rc == 0) {
       break;
    }*
   if (rc < 0) {
       /* system poll error */
       leave_function();
       return -1;
    }
...
somewhere call to packet_read...
}}}

Note: In the original code you check rc only for < 0.

FYI2, I do not rely on ssh_select for polling fds. I use my own
implementation for events handling and transfer only fd between my code and
libssh. May be this information can help...

2011/4/5 Aris Adamantiadis <aris@xxxxxxxxxxxx>

> Hi Mikhail,
>
> Does your problem still happen with 0.5 ? The network code was pretty
> much rewritten, and I think that bug should have disappeard together
> with the synchronous system.
>
> I'll check again, because it may be something completely different now
> that I read it. was this channel_read the *first* channel_read on the
> channel ? Do the content of the buffer you're dumping the correct
> content of your channel ? is it possible that it's some garbage or
> similar content of another session/channel ?
>
> If content is correct and buffer->used is null, there's indeed a problem.
>
> Thanks,
>
> Aris
> Le 5/04/11 21:46, Mikhail Kulinich a écrit :
> > Hello,
> >
> > It seems that guys meet an issue which I am observing for a year or more.
> > During migration to new bug tracker, that original issue (created by me)
> > was missed.
> > The latest info I sent to the mailing list concerning it:
> > {{{
> > #2  0x080e4138 in ssh_socket_unbuffered_read (s=0x8ca3eb0,
> > buffer=0x8c9e1c8, len=16)
> >     at /home/shamanin/libssh_lin/
> > libssh-0.4.4/libssh/socket.c:200
> > 200       rc = recv(s->fd,buffer, len, 0);
> > Current language:  auto; currently c
> > (gdb) up
> > #3  0x080e430b in ssh_socket_completeread (s=0x8ca3eb0,
> > buffer=0x8c9e1c8, len=16)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/socket.c:275
> > 275       while((r = ssh_socket_unbuffered_read(s, ((uint8_t*)buffer +
> > total), toread))) {
> >
> >
> >  (gdb) bt
> > #0  0xb776e424 in __kernel_vsyscall ()
> > #1  0xb770a428 in recv () from /lib/tls/i686/cmov/libpthread.so.0
> > #2  0x080e4138 in ssh_socket_unbuffered_read (s=0x8ca3eb0,
> > buffer=0x8c9e1c8, len=16)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/socket.c:200
> > #3  0x080e430b in ssh_socket_completeread (s=0x8ca3eb0,
> > buffer=0x8c9e1c8, len=16)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/socket.c:275
> > #4  0x080e4985 in ssh_socket_wait_for_data (s=0x8ca3eb0,
> > session=0x8ca4c58, len=16)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/socket.c:405
> > #5  0x080e1e54 in packet_read2 (session=0x8ca4c58)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/packet.c:94
> > #6  0x080e2429 in packet_read (session=0x8ca4c58)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/packet.c:401
> > #7  0x080cdd56 in channel_read (channel=0x8ca9720, dest=0x8cad460,
> > count=1, is_stderr=0)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/channels.c:1916
> > #8  0x0809958b in maux::SshDescriptor::read (this=0x8ca46d8,
> > cpBuf=0x8cad460 "", iBufLen=1,
> >     iFlag=0) at ../SshDescriptor.cc:415
> > #9  0x08053e57 in autotest::SshAccess::_receive (this=0x8ca46d8,
> > buffer=@0xb32b10a8)
> >     at ../SshAccess.cc:177
> > #10 0x080543ed in autotest::SshAccess::onActivity (this=0x8ca46d8,
> > pWaiter=0x8ca4298, iFlags=1)
> >     at ../SshAccess.cc:300
> > #11 0x08092fd3 in maux::Waiter::onActivity (this=0x8ca4298,
> > pxDescr=0x8ca46d8, iFlags=1)
> >     at ../Waiter.cc:915
> > #12 0x08093b7d in maux::Waiter::dispatchOneEvent (this=0x8ca4298) at
> > ../Waiter.cc:692
> > #13 0x08094992 in maux::Waiter::doMainLoop (this=0x8ca4298) at
> > ../Waiter.cc:506
> > #14 0x08096f04 in maux::OwnThread<maux::Waiter>::main (this=0x8ca42f4)
> > at ../Thread.hh:183
> > #15 0x08091afc in maux::mainThreadFunction (pxThreadStartup=0xb63f972c)
> > at ../Thread.cc:40
> > #16 0xb770350f in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
> > #17 0xb74f07ee in clone () from /lib/tls/i686/cmov/libc.so.6
> > (gdb) up
> > #4  0x080e4985 in ssh_socket_wait_for_data (s=0x8ca3eb0,
> > session=0x8ca4c58, len=16)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/socket.c:405
> > 405         r = ssh_socket_completeread(session->socket,buf,to_read);
> > (gdb) up
> > #5  0x080e1e54 in packet_read2 (session=0x8ca4c58)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/packet.c:94
> > 94            rc = ssh_socket_wait_for_data(session->socket, session,
> > blocksize);
> > (gdb) up
> > #6  0x080e2429 in packet_read (session=0x8ca4c58)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/packet.c:401
> > 401       return packet_read2(session);
> > (gdb) up
> > #7  0x080cdd56 in channel_read (channel=0x8ca9720, dest=0x8cad460,
> > count=1, is_stderr=0)
> >     at /home/shamanin/libssh_lin/libssh-0.4.4/libssh/channels.c:1916
> > 1916        if ((packet_read(session)) != SSH_OK ||
> > (gdb) print channel->stdout_buffer
> > $1 = (ssh_buffer) 0x8ca8640
> > (gdb) print channel->stdout_buffer->used
> > $2 = 0
> > (gdb) print *channel->stdout_buffer
> > $3 = {
> >   data = 0x8cacc58 "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<hello
> >
> xmlns=\"urn:ietf:params:xml:ns:netconf:base:1.0\">\n<capabilities>\n<capability>urn:ietf:params:netconf:base:1.0</capability>\n<capability>urn:ietf:params:netc"...,
> > used = 0, allocated = 2048, pos = 0}
> >
> > }}}
> >
> > If you need more details I'll try to get them.
> >
> > 2011/4/5 Aris Adamantiadis <aris@xxxxxxxxxxxx <mailto:aris@xxxxxxxxxxxx
> >>
> >
> >     Jeetu,
> >
> >     I read your email in details and still am not sure it's correlated
> with
> >     Sharon's problem.
> >     Could you provide me a backtrace (with full symbols if possible) of a
> >     stuck process (with the -1 timeout) so I can have an idea of what
> >     condition it's waiting for and why it doesn't arrive.
> >
> >     A verbose log (with ssh_options_set) of the same session would
> greatly
> >     help, along with the code snippet you're using to handle ssh
> connexions.
> >
> >     Thanks,
> >
> >     Aris
> >
> >     Le 5/04/11 14:20, jeetu.golani@xxxxxxxxx
> >     <mailto:jeetu.golani@xxxxxxxxx> a écrit :
> >     > Hi,
> >     >
> >     >> Hello Jeetu, Sharon,
> >     >
> >     >> I'll look at that issue asap. It's a blocking problem for release
> >     0.5.
> >     >
> >     >> Aris
> >     >
> >     > Aris, Thanks so much :) Please let me know if you need me to try
> >     something
> >     > or any further data to help you debug this.
> >     >
> >     >> If this is on Windows, which Windows version are you running? We
> >     have a
> >     >> poll(2)-emulation for older Windows versions. This version
> >     doesn't set
> >     >> POLLERR, maybe that's causing the problem.
> >     >>
> >     >
> >     > Andreas, I'm seeing this behaviour on Linux (debian testing).
> >     Please let
> >     > me know if you need any further info to debug this or if I can
> >     help in any way.
> >     >
> >     > Thanks so much yet again for a very nice piece of work :)
> >     >
> >     > Bye for now
> >     > Jeetu
> >     >
> >     > ebrain.in <http://ebrain.in> | Beehive Computing
> >     > Discover and run software from any device around you.
> >     >
> >
> >
> >
> >
> > --
> > Mikhail
> >
>
>


-- 
Mikhail