SMP-ix86-threads-fork: Linux 2.4.x kernel problem identified [phantom read()]

Wolfram Gloger wg@malloc.de
Thu Sep 13 05:42:00 GMT 2001


> Note that I've checked in a slightly different patch.  It must be
> sufficient to use TEMP_FAILURE_RETRY since all the communication is
> with pipes.  If read/write would produce short reads or writes
> something would be wrong in the kernel.

I agree.  What you have checked in looks fine, and I also meant to
suggest to change the write() side in my previous mail.

Thanks,
Wolfram.



More information about the Libc-alpha mailing list