[Xitami] SMP Problem (WAS: Xitami 2.4d9/Xitami 2.5b5)

[Dr. Martin R. Hadam]

On Mon, 25 Jun 2001 22:03:06 +1200, Ewen McNeill wrote:

>However as I said, Xitami tries to buck this trend and be single
>threaded, single process -- which has its advantages and disadvantages.

 since I'm running cgi programs written in rexx, smp is nice as all
spawned processes are distributed evenly on both processors.

>All that said, I'm puzzled what could be causing your problem on a SMP
>machine, which is fixed by running it on a single processor, if it isn't
>a threading issue.  

 not being a programmer (other than rexx <g>) I can't help you much.
But what transpires to me as feedback on my (other) bug reports is that
thread timing seems critical on smp. My last bug report just the other
week was fixed within 24 hours - the programmers's immediate response
was "my gut feeling is it's this one..."

>Can you tell me what build flags you used to build Xitami?  And which
>version of the compiler and build environment?

 I think there is misunderstanding: I did not build Xitami myself, I
only used the patched version that was made by Bernd Hohmann. HOWEVER,
the same applies also EACH version after v2.4d6 that was available on
the Xitami website. So it's probably independent on the tools used to
build the binary.

 When I first noticed (without doing the execmode test), I though that
it might relate to my EMX version used, which is the newer than the one
recommended. My current version is:
EMX : revision = 61
EMXIO : revision = 60
EMXLIBC : revision = 63
EMXLIBCM : revision = 64
EMXLIBCS : revision = 64
EMXWRAP : revision = 60
 I should add, however, that this does seem unlikely in view of my
later observations using execmode. Also, prior revisions (still above
the recommended level) showed the same problem.

 Hope this will get you any further.

 Martin

 Martin R. Hadam
 Kinderklinik - Medizinische Hochschule
 D-30623 Hannover
 Germany
 Email: Hadam.Martin@MH-Hannover.de