Testing Call

[Leigh Brown]

Yesterday, Ben Lindstrom wrote:
> 2. AIX (some version) stall on large data output.  Current CVS patch is
> wrong and will be pulled unless an agreement can be made.

The (apparent) bug in AIX applies to AIX 4.3.3ML03 and above.  Having
at last found a solution that works I'd be really grateful if we can
have a workaround for the bug before the release.

I have attached a patch that moves the workaround to the correct place,
above the write, and changes "c->isatty" to "isatty(c->wfd)", because
c->isatty is not true if nonblock is false (I don't know why it is
setup like this).

It may not be the most correct or efficient solution, but it only
applies to AIX, and without it openssh locks up solid very easily
when ssh'ing from a box that you telnet'd or rlogin'd to.

Please apply!

BTW I'm running last night's snapshot with this patch applied and
privilege separation enabled, and it all seems to be working fine.
I will report if I encounter any problems.

Cheers,

Leigh.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: aix.fix.patch
Type: application/octet-stream
Size: 596 bytes
Desc: not available
Url : http://lists.mindrot.org/pipermail/openssh-unix-dev/attachments/20020718/18844d62/attachment.obj