PCI Hotplug and Linux 2.5.4-pre2
I can't find the definition of pcihpfs_fs_type used in pci_hotplug_core.c in register_filesystem(), unregister_filesystem(), and kern_mount(). This is causing a compilation error.
PS - I don't use this in my kernel, so I can disable it from my config without penalty. But seeing as I don't see any posts about this on LKML, should I try to build kernels with everything enabled in the future (just to get these small errors stomped out quickly)?
Quoted 3 lines · John Weber
I can't find the definition of
pcihpfs_fs_typeused inpci_hotplug_core.cinregister_filesystem(),unregister_filesystem(), andkern_mount(). This is causing a compilation error.
Here's the fix, I'll send it on.
thanks,
greg k-h
drivers/hotplug/pci_hotplug_core.c +3 -3
@@ -414,7 +414,7 @@ }spin_unlock (&mount_lock); - mnt =kern_mount (&pcihpfs_fs_type); + mnt =kern_mount (&pcihpfs_type);if (IS_ERR (mnt)) {err ("could not mount the fs...erroring out!\n" );return -ENODEV ;
@@ -1139,7 +1139,7 @@spin_lock_init (&list_lock);dbg ("registering filesystem.\n" ); - result =register_filesystem (&pcihpfs_fs_type); + result =register_filesystem (&pcihpfs_type);if (result) {err ("register_filesystem failed with %d\n" , result); goto exit;
@@ -1153,7 +1153,7 @@ static void __exit pci_hotplug_exit (void) { -unregister_filesystem (&pcihpfs_fs_type); +unregister_filesystem (&pcihpfs_type); }module_init (pci_hotplug_init);
Quoted 8 lines · John Weber
I can't find the definition of
pcihpfs_fs_typeused inpci_hotplug_core.cinregister_filesystem(),unregister_filesystem(), andkern_mount(). This is causing a compilation error.PS - I don't use this in my kernel, so I can disable it from my config without penalty. But seeing as I don't see any posts about this on LKML, should I try to build kernels with everything enabled in the future (just to get these small errors stomped out quickly)?
I vote yes; I always turn on as many modularizable chunks as possible, to increase "compile coverage". And we all know if it compiles, it works, right?
Quoted 3 lines · Brian Strand
I vote yes; I always turn on as many modularizable chunks as possible, to increase "compile coverage". And we all know if it compiles, it works, right?
I hope that was sarcasm 8-) Some things such as the NCR5380 'fixes' in Linus tree have put the driver into a state where it compiles and maybe even runs, but it's almost guaranteed to corrupt. There's a more complete set of fixes for some of these problems in my tree, but alas, doesn't compile..
Numerous other bio related problems have been reported and had 'quick fixes' offered for them which turn out to be worse than a non-compiling driver.
Signature · 2 lines
| Dave Jones. http://www.codemonkey.org.uk | SuSE Labs