7.3-STABLE and Linux version of SIMetrix

Andre Albsmeier Andre.Albsmeier at siemens.com
Fri Jan 7 07:20:13 UTC 2011


On Thu, 06-Jan-2011 at 13:28:57 +0100, Alexander Leidinger wrote:
> Quoting Andre Albsmeier <Andre.Albsmeier at siemens.com> (from Thu, 6 Jan  
> 2011 11:40:57 +0100):
> 
> > On Thu, 06-Jan-2011 at 09:01:30 +0100, Alexander Leidinger wrote:
> >> Quoting Andre Albsmeier <Andre.Albsmeier at siemens.com> (from Wed, 5 Jan
> >> 2011 20:19:15 +0100):
> >>
> >> > Got it running... A short explanation:
> >> >
> >> > Linux' shm_open() fails because it wants to find some funky shmfs
> >> > to construct the full pathname. It starts to search at the default
> >> > mountpoint which is /dev/shm. If this fails it runs through fstab
> >> > and searches for shmfs and tmpfs. Whatever it finds will be
> >> > statfs()'ed to be checked for Linux' fs magic for shmfs (0x01021994).
> >>
> >> What does it expect as a filesystem type if it does not find shmfs in
> >> fstab but tmpfs? If it does not find tmpfs, will it try /tmp anyway
> >> (but check for some fstype magic)?
> >
> > It searches for every mount which is of type tmpfs or shm.
> > Whatever it finds must have the fs magic SHMFS_SUPER_MAGIC
> > (0x01021994). It's in sysdeps/unix/sysv/linux/shm_open.c:
> >
> > ------------------------- snip -------------------
> >
> >   /* Now read the entries.  */
> >   while ((mp = __getmntent_r (fp, &resmem, buf, sizeof buf)) != NULL)
> >     /* The original name is "shm" but this got changed in early Linux
> >        2.4.x to "tmpfs".  */
> 
> This looks like it is creating real files there, and the tmpfs thing  

Yes, that's true.

> shall make sure it is empty after a reboot. That's just an assumption...

Possibly. I have also seen statements in mailing lists where
they talk about that fs being some kind of memory based fs
to increase speed.

> 
> Do you have some temporary files in the location your /dev/shm points  
> to during running the linux application which needs it (I do not  
> expect that our implementation creates them, but as I haven't looked  
> at the code yet, this is something to be verified)?

Yes. In the case of my application (SIMetrix) files with a
size of approx. 1,3 MB are created and then removed.

	-Andre

> 
> Bye,
> Alexander.
> 
> -- 
> Is death legally binding?
> 
> http://www.Leidinger.net    Alexander @ Leidinger.net: PGP ID = B0063FE7
> http://www.FreeBSD.org       netchild @ FreeBSD.org  : PGP ID = 72077137
> 

-- 
Micro$oft: Which virus will you get today?


More information about the freebsd-emulation mailing list