[Freeipmi-users] libfreeipmi select() may exceed the max number of file descriptors in ipmi-openipmi-driver.c

[Ole Holm Nielsen]

Dear FreeIPMI developers,

In freeipmi-1.6.11 the driver libfreeipmi/driver/ipmi-openipmi-driver.c in 
line 532 calls the select() function:

    if ((n = select (ctx->device_fd + 1,
      ...

The Slurm Resource Manager batch queue system for Linux clusters has been 
hit by crashes due to a bug which has been attributed to the 
ipmi-openipmi-driver.c using select().

The issue is described in this bug report: 
https://bugs.schedmd.com/show_bug.cgi?id=17639#c30

On a very busy Linux cluster it seems that ipmi-openipmi-driver.c may 
sometimes exceed the number of file descriptors available (1024).  The 
select(2) man-page states:

> BUGS
>        POSIX allows an implementation to define an upper limit, advertised via
>        the  constant  FD_SETSIZE, on the range of file descriptors that can be
>        specified in a file descriptor set.  The Linux kernel imposes no  fixed
>        limit,  but  the  glibc  implementation makes fd_set a fixed-size type,
>        with FD_SETSIZE defined  as  1024,  and  the  FD_*()  macros  operating
>        according  to  that  limit.   To  monitor file descriptors greater than
>        1023, use poll(2) instead.

Question: Would it be possible for you to replace the select() by poll() 
in the FreeIPMI driver/ipmi-openipmi-driver.c code?

Thanks a lot,
Ole

--
Ole Holm Nielsen
PhD, Senior HPC Officer
Department of Physics, Technical University of Denmark