Re: git clone over smart-http hanging for just one repo.

["Shawn O. Pearce" <spearce@spearce.org>]

Brady Catherman <brady@catherman.org> wrote:
> I have a git repo that fails to clone or fetch over smart-http, but  
> works great over dav. I am wondering if somebody can help me debug the  
> issue since I am at a loss why this is happening.

Yea, I'm at a loss too. :-(

> The interesting parts of a strace of git-http-backend following a git  
> clone follow:
...
> 12037 close(1)                          = 0

Why did the CGI process just close stdout?  I'm guessing this is
part of the exec of the upload-pack child in the background.  Oh,
right, we closed it because we passed the descriptor to the child
and now the parent CGI doesn't want it anymore.

> 12037 write(1, "Status: 500 Internal Server Error\r\n", 35) = -1 EBADF  
> (Bad file descriptor)

This smells like the backend upload-pack process got into trouble and
exited early, so now the CGI is trying to change the status to 500
since the backend exited with a non-zero status.  Only its too late,
as the filedescriptor was already closed after the successful fork().

We're stuck in a loop because we're failing during the die routine.

Because the file descriptor is closed, safe_write() which was what
originated that write(1, ...) above, tries to call die().  But that
die() call invokes die_webcgi() which in turn tries to write that
500 error message again to 1.  So this goes on a for a while...

> 12037 --- SIGSEGV (Segmentation fault) @ 0 (0) ---
> 12037 +++ killed by SIGSEGV +++

and then we run out of stack space, due to too many recursions,
and the process is aborted by a SIGSEGV.

> Anybody have any thoughts why this would happen or what can be done to  
> fix it?

A gdb trace or something of the upload-pack process would help.
That appears to have also died and we don't know why.  Its death
is what contributed to the CGI crashing above.

I'll try to send a patch for this recursive crashing problem in
the CGI.

-- 
Shawn.