public inbox for linux-mtd@lists.infradead.org
 help / color / mirror / Atom feed
* can't rmmod cfi_cmdset_0002
@ 2002-05-10  8:16 Galen Seitz
  2002-05-10 21:35 ` Galen Seitz
  0 siblings, 1 reply; 3+ messages in thread
From: Galen Seitz @ 2002-05-10  8:16 UTC (permalink / raw)
  To: linux-mtd

In the course of hacking on pci.c, I've encountered an issue with removing
the cfi_cmdset_0002 module.  It seems that once I have successfully
loaded the pci module, I can't unload the cmdset module.  Is it likely
that I have created this problem?  How does one go about tracking
down a problem like this?

thanks,
galen

oz:~# cat /proc/version
Linux version 2.4.19-pre8 (galens@oz.seitzassoc.com) (gcc version 2.96 
20000731 (Red Hat Linux 7.1 2.96-98)) #1 Thu May 9 13:48:33 PDT 2002
oz:~# /sbin/modprobe -a mtdchar mtdblock cfi_cmdset_0002 map_rom cfi_probe
oz:~# /sbin/lsmod 
Module                  Size  Used by
cfi_probe               7424   0  (unused)
gen_probe               1712   0  [cfi_probe]
map_rom                  688   0  (unused)
chipreg                  704   0  [cfi_probe map_rom]
cfi_cmdset_0002        15616   0  (unused)
mtdblock                4608   0  (unused)
mtdchar                 4064   0  (unused)
mtdcore                 2176   0  [mtdblock mtdchar]
3c59x                  24928   1 
rtc                     6016   0  (autoclean)
oz:~# /sbin/rmmod mtdchar mtdblock cfi_cmdset_0002 map_rom cfi_probe
oz:~# /sbin/lsmod 
Module                  Size  Used by
gen_probe               1712   0 
chipreg                  704   0 
mtdcore                 2176   0 
3c59x                  24928   1 
rtc                     6016   0  (autoclean)
oz:~# /sbin/modprobe -a mtdchar mtdblock cfi_cmdset_0002 map_rom cfi_probe
oz:~# /sbin/modprobe -a pci                                               
oz:~# /sbin/lsmod 
Module                  Size  Used by
pci                     2880   0  (unused)
cfi_probe               7424   0 
map_rom                  688   0  (unused)
cfi_cmdset_0002        15616   2 
mtdblock                4608   0  (unused)
mtdchar                 4064   0  (unused)
gen_probe               1712   0  [cfi_probe]
chipreg                  704   0  [pci cfi_probe map_rom]
mtdcore                 2176   1  [pci mtdblock mtdchar]
3c59x                  24928   1 
rtc                     6016   0  (autoclean)
oz:~# /sbin/rmmod pci                                       
oz:~# /sbin/rmmod mtdchar mtdblock cfi_cmdset_0002 map_rom cfi_probe
cfi_cmdset_0002: Device or resource busy
oz:~# /sbin/lsmod 
Module                  Size  Used by
cfi_cmdset_0002        15616   1 
gen_probe               1712   0 
chipreg                  704   0 
mtdcore                 2176   0 
3c59x                  24928   1 
rtc                     6016   0  (autoclean)

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: can't rmmod cfi_cmdset_0002
  2002-05-10  8:16 can't rmmod cfi_cmdset_0002 Galen Seitz
@ 2002-05-10 21:35 ` Galen Seitz
  2002-05-11 22:15   ` David Woodhouse
  0 siblings, 1 reply; 3+ messages in thread
From: Galen Seitz @ 2002-05-10 21:35 UTC (permalink / raw)
  To: linux-mtd

> In the course of hacking on pci.c, I've encountered an issue with
> removing the cfi_cmdset_0002 module.  It seems that once I have
> successfully loaded the pci module, I can't unload the cmdset module.
> Is it likely that I have created this problem?  How does one go about
> tracking down a problem like this? 

I've done a little more experimenting and come up with another data point.
After installing physmap, the use count of cfi_cmdset_0002 goes from
0 to 2.  When I rmmod physmap, the use count of cfi_cmdset_0002 only drops
back to 1, which prevents the module from being unloaded.  Has anyone
else seen this problem?  Note that the only change I made to physmap.c
was to change ioremap to ioremap_nocache.  It appears from looking at
the code that a call to cfi_cmdset_0002 will result in the macro 
MOD_INC_USE_COUNT being executed twice, once in cfi_cmdset_0002, and again
in cfi_amdstd_setup.  cfi_cmdset_0001.c has a similar structure.
Is this intentional?

thanks,
galen

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: can't rmmod cfi_cmdset_0002
  2002-05-10 21:35 ` Galen Seitz
@ 2002-05-11 22:15   ` David Woodhouse
  0 siblings, 0 replies; 3+ messages in thread
From: David Woodhouse @ 2002-05-11 22:15 UTC (permalink / raw)
  To: Galen Seitz; +Cc: linux-mtd

galens@seitzassoc.com said:
>   It appears from looking at the code that a call to cfi_cmdset_0002
> will result in the macro  MOD_INC_USE_COUNT being executed twice, once
> in cfi_cmdset_0002, and again in cfi_amdstd_setup.  cfi_cmdset_0001.c
> has a similar structure. Is this intentional? 

Not really, no. The modular stuff is a bit messy and wants attention.

--
dwmw2

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2002-05-11 22:15 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2002-05-10  8:16 can't rmmod cfi_cmdset_0002 Galen Seitz
2002-05-10 21:35 ` Galen Seitz
2002-05-11 22:15   ` David Woodhouse

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox