http://bugzilla.mindrot.org/show_bug.cgi?id=52 ------- Comment #47 from tsi at ualberta.ca 2007-02-17 03:06 ------- (In reply to comment #46)> Yes, but that workaround doesn't need the patch. However, with latest > patch, there is no hang when I run test2.ksh and exit from the shell on > linux. But it is not the case with HP-UX. So Im wondering what the > patch is doing with respect to HP-UX.This is at odds with comment #26. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
http://bugzilla.mindrot.org/show_bug.cgi?id=52 ------- Comment #48 from senthilkumar_sen at hotpop.com 2007-02-19 18:25 ------- (In reply to comment #47)> (In reply to comment #46) > > Yes, but that workaround doesn't need the patch. However, with latest > > patch, there is no hang when I run test2.ksh and exit from the shell on > > linux. But it is not the case with HP-UX. So Im wondering what the > > patch is doing with respect to HP-UX. > > This is at odds with comment #26. >I checked the testcases described in <A href="http://bugzilla.mindrot.org/show_bug.cgi?id=52#c35">comment 35</a> and they worked as described there for both HP-UX and Linux. But HP-UX behaves differently only when I login and launch my script described in <A href="http://bugzilla.mindrot.org/show_bug.cgi?id=52#c44">comment 44</a>. So I think there is something special that goes in HP-UX at login which is not happening when I just run with -t option. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
http://bugzilla.mindrot.org/show_bug.cgi?id=52 ------- Comment #49 from tsi at ualberta.ca 2007-02-20 08:05 ------- (In reply to comment #48)> (In reply to comment #47) > > (In reply to comment #46) > > > Yes, but that workaround doesn't need the patch. However, with > > > latest patch, there is no hang when I run test2.ksh and exit from > > > the shell on linux. But it is not the case with HP-UX. So Im > > > wondering what the patch is doing with respect to HP-UX.> > This is at odds with comment #26.> I checked the testcases described in <A > href="http://bugzilla.mindrot.org/show_bug.cgi?id=52#c35">comment > 35</a> and they worked as described there for both HP-UX and Linux. > But HP-UX behaves differently only when I login and launch my script > described in <A > href="http://bugzilla.mindrot.org/show_bug.cgi?id=52#c44">comment > 44</a>. So I think there is something special that goes in HP-UX at > login which is not happening when I just run with -t option.OK. So, what happens if you do ... ./test2.ksh & exit ... when logged in? ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
http://bugzilla.mindrot.org/show_bug.cgi?id=52 ------- Comment #50 from senthilkumar_sen at hotpop.com 2007-02-20 15:25 ------- (In reply to comment #49)> (In reply to comment #48) > > (In reply to comment #47) > > > (In reply to comment #46) > > > > Yes, but that workaround doesn't need the patch. However, with > > > > latest patch, there is no hang when I run test2.ksh and exit from > > > > the shell on linux. But it is not the case with HP-UX. So Im > > > > wondering what the patch is doing with respect to HP-UX. > > > > This is at odds with comment #26. > > > I checked the testcases described in <A > > href="http://bugzilla.mindrot.org/show_bug.cgi?id=52#c35">comment > > 35</a> and they worked as described there for both HP-UX and Linux. > > But HP-UX behaves differently only when I login and launch my script > > described in <A > > href="http://bugzilla.mindrot.org/show_bug.cgi?id=52#c44">comment > > 44</a>. So I think there is something special that goes in HP-UX at > > login which is not happening when I just run with -t option. > > OK. So, what happens if you do ... > > ./test2.ksh & > exit > > ... when logged in? >Sorry for the embedded tags in my last comment. To answer the question, $ ./test2.ksh & [1] 2659 $ exit There are running jobs. [1] + Done ./test2.ksh & $ exit logout <It hangs after this> This happens irrespective of the patch in HP-UX. I'm using korn shell. Not sure is this something shell specific. I guess it can't be, because if I do the same steps after login through telnet to a HP-UX box, the connection get closed. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
http://bugzilla.mindrot.org/show_bug.cgi?id=52 ------- Comment #51 from djm at mindrot.org 2007-02-20 21:46 ------- sigh. five comments and no sign of a debug trace or a description of your configuration (anything non default?). ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
http://bugzilla.mindrot.org/show_bug.cgi?id=52 ------- Comment #52 from senthilkumar_sen at hotpop.com 2007-02-20 23:27 ------- Created an attachment (id=1242) --> (http://bugzilla.mindrot.org/attachment.cgi?id=1242&action=view) Debug output on sshd The debug output on sshd is attached here. I'm using OpenSSH 4.5 and the configure options are below, ./configure --prefix=<foo> --with-zlib=<foo> --with-tcp-wrappers=<foo> As far as client side is concerned, its normal debug output and as soon as I gave exit the following happens on client side, $ ./test2.ksh & [1] 18353 $ exit There are running jobs. [1] + Done ./test2.ksh & $ exit logout debug1: client_input_channel_req: channel 0 rtype exit-status reply 0 I will upload the ssh client side debug output if you feel it is needed. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
http://bugzilla.mindrot.org/show_bug.cgi?id=52 ------- Comment #53 from senthilkumar_sen at hotpop.com 2007-02-21 01:41 ------- (In reply to comment #51)> sigh. five comments and no sign of a debug trace or a description of > your configuration (anything non default?). >No, I just used the default sshd_config and ssh_config for this. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
http://bugzilla.mindrot.org/show_bug.cgi?id=52 ------- Comment #54 from tsi at ualberta.ca 2007-02-22 05:11 ------- (In reply to comment #52)> Created an attachment (id=1242)--> (http://bugzilla.mindrot.org/attachment.cgi?id=1242&action=view) [details]> Debug output on sshd> The debug output on sshd is attached here. I'm using OpenSSH 4.5 and > the configure options are below,> ./configure --prefix=<foo> --with-zlib=<foo> --with-tcp-wrappers=<foo>> As far as client side is concerned, its normal debug output and as soon > as I gave exit the following happens on client side,> $ ./test2.ksh & > [1] 18353 > $ exit > There are running jobs. > [1] + Done ./test2.ksh & > $ exit > logout > debug1: client_input_channel_req: channel 0 rtype exit-status reply 0> I will upload the ssh client side debug output if you feel it is > needed.Does it behave differently with protocol version 1, i.e. `ssh -1`? ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
http://bugzilla.mindrot.org/show_bug.cgi?id=52 ------- Comment #55 from senthilkumar_sen at hotpop.com 2007-02-22 15:17 ------- Created an attachment (id=1243) --> (http://bugzilla.mindrot.org/attachment.cgi?id=1243&action=view) SSHD debug output for protocol 1 No difference between protocol 1 and 2. They simply hang. The client side for protocol 1 is shown below, $ ./test2.ksh & [1] 3428 $ There are running jobs. [1] + Done ./test2.ksh & $ logout ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.