Re: OpenSSH 5.1: call for testing - openssh

This is a discussion on Re: OpenSSH 5.1: call for testing - openssh ; On Tue, 8 Jul 2008, Cristian Ionescu-Idbohrn wrote: Thanks for the detailed report! > Warnings: > > xmalloc.h:26: warning: `__nonnull__' attribute directive ignored We'll have to ignore this, it is harmless though. > bindresvport.c: In function `bindresvport_sa': > bindresvport.c:57: warning: ...

+ Reply to Thread
Results 1 to 9 of 9

Thread: Re: OpenSSH 5.1: call for testing

  1. Re: OpenSSH 5.1: call for testing

    On Tue, 8 Jul 2008, Cristian Ionescu-Idbohrn wrote:

    Thanks for the detailed report!

    > Warnings:
    >
    > xmalloc.h:26: warning: `__nonnull__' attribute directive ignored


    We'll have to ignore this, it is harmless though.

    > bindresvport.c: In function `bindresvport_sa':
    > bindresvport.c:57: warning: declaration of `sin' shadows a global declaration
    > :0: warning: shadowed declaration is here


    Fixed.

    > rresvport.c: In function `rresvport_af':
    > rresvport.c:92: warning: implicit declaration of function `close'


    Fixed.

    > umac.c:187: warning: declaration of `buffer_ptr' shadows a global declaration
    > buffer.h:31: warning: shadowed declaration is here


    Fixed.

    > auth.c: In function `auth_openkeyfile':
    > auth.c:489: warning: implicit declaration of function `close'


    Fixed.

    > I also get this warning from the linker:
    >
    > /usr/local/cris/r64/lib/gcc-lib/cris-axis-linux-gnu/3.2.1/../../../../cris-axis-linux-gnu/bin/ld: warning: type and size of dynamic symbol `AES_encrypt' are not defined
    >
    > which goes away if I build with:
    >
    > CFLAGS += -DUSE_BUILTIN_RIJNDAEL


    That shouldn't be necessary unless you are using quite an old OpenSSL.
    Is this the case?

    -d
    _______________________________________________
    openssh-unix-dev mailing list
    openssh-unix-dev@mindrot.org
    https://lists.mindrot.org/mailman/li...enssh-unix-dev


  2. Re: OpenSSH 5.1: call for testing

    Thanks for those fixes.

    On Mon, 14 Jul 2008, Damien Miller wrote:

    > > I also get this warning from the linker:
    > >
    > > /usr/local/cris/r64/lib/gcc-lib/cris-axis-linux-gnu/3.2.1/../../../../cris-axis-linux-gnu/bin/ld: warning: type and size of dynamic symbol `AES_encrypt' are not defined
    > >
    > > which goes away if I build with:
    > >
    > > CFLAGS += -DUSE_BUILTIN_RIJNDAEL

    >
    > That shouldn't be necessary unless you are using quite an old OpenSSL.
    > Is this the case?


    Is 0.9.7m too old?


    Cheers,

    --
    Cristian
    _______________________________________________
    openssh-unix-dev mailing list
    openssh-unix-dev@mindrot.org
    https://lists.mindrot.org/mailman/li...enssh-unix-dev


  3. Re: OpenSSH 5.1: call for testing

    On Mon, 14 Jul 2008, Cristian Ionescu-Idbohrn wrote:

    > > > CFLAGS += -DUSE_BUILTIN_RIJNDAEL

    > >
    > > That shouldn't be necessary unless you are using quite an old OpenSSL.
    > > Is this the case?

    >
    > Is 0.9.7m too old?


    Not at all - I think AES has been in OpenSSL since 0.9.5 or 0.9.6,
    unless it has been disabled at build time.

    Does the linked ssh/sshd link OK and pass the regress tests?

    -d
    _______________________________________________
    openssh-unix-dev mailing list
    openssh-unix-dev@mindrot.org
    https://lists.mindrot.org/mailman/li...enssh-unix-dev


  4. Re: OpenSSH 5.1: call for testing

    On Mon, 14 Jul 2008, Damien Miller wrote:

    > On Mon, 14 Jul 2008, Cristian Ionescu-Idbohrn wrote:
    >
    > > > > CFLAGS += -DUSE_BUILTIN_RIJNDAEL
    > > >
    > > > That shouldn't be necessary unless you are using quite an old OpenSSL.
    > > > Is this the case?

    > >
    > > Is 0.9.7m too old?

    >
    > Not at all - I think AES has been in OpenSSL since 0.9.5 or 0.9.6,
    > unless it has been disabled at build time.


    Ok.

    > Does the linked ssh/sshd link OK and pass the regress tests?


    Any easy way to do that?

    I'm cross building.
    I build just some of the build targets.
    Target system lacks development tools.


    Cheers,

    --
    Cristian
    _______________________________________________
    openssh-unix-dev mailing list
    openssh-unix-dev@mindrot.org
    https://lists.mindrot.org/mailman/li...enssh-unix-dev


  5. Re: OpenSSH 5.1: call for testing

    On Mon, 14 Jul 2008, Cristian Ionescu-Idbohrn wrote:

    > > Does the linked ssh/sshd link OK and pass the regress tests?

    >
    > Any easy way to do that?
    >
    > I'm cross building.
    > I build just some of the build targets.
    > Target system lacks development tools.


    You would need at least "make" on the target, but you should be able
    to copy over the build directory and run "make tests".

    -d
    _______________________________________________
    openssh-unix-dev mailing list
    openssh-unix-dev@mindrot.org
    https://lists.mindrot.org/mailman/li...enssh-unix-dev


  6. Re: OpenSSH 5.1: call for testing

    On Tue, 15 Jul 2008, Damien Miller wrote:

    > On Mon, 14 Jul 2008, Cristian Ionescu-Idbohrn wrote:
    >
    > > > Does the linked ssh/sshd link OK and pass the regress tests?

    > >
    > > Any easy way to do that?
    > >
    > > I'm cross building.
    > > I build just some of the build targets.
    > > Target system lacks development tools.

    >
    > You would need at least "make" on the target,


    I was afraid you would say that

    > but you should be able to copy over the build directory and run
    > "make tests".


    I'll see if I can build and squeeze make on the target system.

    Just found another warning (this is 5.0p1) you might want to look
    into:

    ssh.c: In function `control_client':
    ssh.c:1507: warning: long unsigned int format, unsigned int arg (arg 4)



    Cheers,

    --
    Cristian
    _______________________________________________
    openssh-unix-dev mailing list
    openssh-unix-dev@mindrot.org
    https://lists.mindrot.org/mailman/li...enssh-unix-dev


  7. Re: OpenSSH 5.1: call for testing

    On Wed, 16 Jul 2008, Cristian Ionescu-Idbohrn wrote:

    > On Tue, 15 Jul 2008, Damien Miller wrote:
    >
    > > On Mon, 14 Jul 2008, Cristian Ionescu-Idbohrn wrote:
    > >
    > > > > Does the linked ssh/sshd link OK and pass the regress tests?
    > > >
    > > > Any easy way to do that?
    > > >
    > > > I'm cross building.
    > > > I build just some of the build targets.
    > > > Target system lacks development tools.

    > >
    > > You would need at least "make" on the target,

    >
    > I was afraid you would say that
    >
    > > but you should be able to copy over the build directory and run
    > > "make tests".

    >
    > I'll see if I can build and squeeze make on the target system.


    Failing that, just running (i.e. connecting to/from) sshd and ssh on
    the target should show whether the linker warning is harmless or not.

    > Just found another warning (this is 5.0p1) you might want to look
    > into:
    >
    > ssh.c: In function `control_client':
    > ssh.c:1507: warning: long unsigned int format, unsigned int arg (arg 4)


    I think that code was moved entirely after 5.0p1

    -d
    _______________________________________________
    openssh-unix-dev mailing list
    openssh-unix-dev@mindrot.org
    https://lists.mindrot.org/mailman/li...enssh-unix-dev


  8. Re: OpenSSH 5.1: call for testing

    On Wed, 16 Jul 2008, Damien Miller wrote:

    > On Wed, 16 Jul 2008, Cristian Ionescu-Idbohrn wrote:
    >
    > > I'll see if I can build and squeeze make on the target system.

    >
    > Failing that, just running (i.e. connecting to/from) sshd and ssh on
    > the target should show whether the linker warning is harmless or not.


    Alright. Done that now. Quick test and no problems.
    But I'd still want to know what that linker warning means.

    > > Just found another warning (this is 5.0p1) you might want to look
    > > into:
    > >
    > > ssh.c: In function `control_client':
    > > ssh.c:1507: warning: long unsigned int format, unsigned int arg (arg 4)

    >
    > I think that code was moved entirely after 5.0p1


    Right. And that gave me the opportunity to check latest snapshot
    Today's snapshot shows no such warning.


    Cheers,

    --
    Cristian
    _______________________________________________
    openssh-unix-dev mailing list
    openssh-unix-dev@mindrot.org
    https://lists.mindrot.org/mailman/li...enssh-unix-dev


  9. Re: OpenSSH 5.1: call for testing

    I think I owe an explanation.

    On Mon, 14 Jul 2008, Damien Miller wrote:

    > On Tue, 8 Jul 2008, Cristian Ionescu-Idbohrn wrote:
    >
    > > I also get this warning from the linker:
    > >
    > > /usr/local/cris/r64/lib/gcc-lib/cris-axis-linux-gnu/3.2.1/../../../../cris-axis-linux-gnu/bin/ld: warning: type and size of dynamic symbol `AES_encrypt' are not defined
    > >
    > > which goes away if I build with:
    > >
    > > CFLAGS += -DUSE_BUILTIN_RIJNDAEL

    >
    > That shouldn't be necessary unless you are using quite an old OpenSSL.
    > Is this the case?


    Showed up the above warning was caused by ourselfs
    We misbuilded openssl. Fixed now.

    Crossbuilt (for cris v10) latest snapshot. Builds and runs fine; no
    (arch) patches required

    I'm seeing these warnings when I build for host:

    sftp.c: In function 'parse_dispatch_command':
    sftp.c:1238: warning: 'n_arg' may be used uninitialized in this function
    sftp.c:1237: warning: 'hflag' may be used uninitialized in this function
    sftp.c:1237: warning: 'iflag' may be used uninitialized in this function
    sftp.c:1237: warning: 'lflag' may be used uninitialized in this function
    sftp.c:1237: warning: 'pflag' may be used uninitialized in this function

    Shouldn't cost much to give them a sensible initial value.

    All tests passed on a debian lenny box.


    Cheers,

    --
    Cristian
    _______________________________________________
    openssh-unix-dev mailing list
    openssh-unix-dev@mindrot.org
    https://lists.mindrot.org/mailman/li...enssh-unix-dev


+ Reply to Thread