Patch: linux-2.5.2-pre8/drivers/md partial compilation fixes

Patch: linux-2.5.2-pre8/drivers/md partial compilation fixes

Post by Adam J. Richte » Sun, 06 Jan 2002 18:20:07

        The following patches fix the kdev-related compilation
errors in linux-2.5.2/drivers/md, but they do not fix the bio-related
compilation errors, which is a pretty separate task.

        Applying this patch should make life simpler for anyone
attempting to update driver/md to the new kernel.  So, if nobody
complains, I would ask that it be applied to the stock kernel.

        By the way, if possible, please ignore my previously posted
patch for linux-2.5.2-pre8/drivers/md/raid1.c, in favor of the one
included here.  The only difference is that I replaced an instance of
"dev = mk_dev(0,0)" with the more correct "dev = NODEV".  If that
is not convenenient, then feel free to ignore the raid1.c patch
in this set and I will resubmit the one different after pre9 is

Adam J. Richter     __     ______________   4880 Stevens Creek Blvd, Suite 104

+1 408 261-6630         | g g d r a s i l   United States of America
fax +1 408 261-6631      "Free Software For The Rest Of Us."

30K Download

1. Patch?: linux-2.5.2-pre8/drivers/video kdev_t compilation fixes

        Most (all?) of the frame buffer drivers in
linux-2.5.2-pre8/drivers/video have a kdev_t compilation error,
because they try to set their fb_info->node field to -1 (it is
a kdev_t).  Shortly thereafter, regsiter_framebuffer sets it to
a more useful value.  So far, I have been unable to spot any
code that relies on fb_info->node being initialized to a
particular value prior to the call to register_framebuffer,
although I have not looked very hard.  What I did see was
some "???" comments beside some of the initializations to -1.
I suspsect that this is just useless initialization.

        Anyhow, deleting all of those initializations to -1
allows all of the framebuffer drivers to compile, which is
that this patch does.  I have also included a tiny patch to
a commented out line in include/linux/fb.h, updating it to
use minor() instead of MINOR(), in case it is every uncommented.

        I am still a long way from getting pre8 to run, so I
have not tested change.  Can anyone point out a place in
drivers/framebuffer that relies on fb_info->node being set to -1,
or minor(fb_info->node) being -1?

        By the way, don't worry if a few of the line offsets in this
patch are off.  That's because of some other changes I have in my
drivers/framebuffer (which I have previous posted, specifically
PCI ID tables, fb font modularization with demand loading, and the
fb core as a loadable module).

Adam J. Richter     __     ______________   4880 Stevens Creek Blvd, Suite 104

+1 408 261-6630         | g g d r a s i l   United States of America
fax +1 408 261-6631      "Free Software For The Rest Of Us."

21K Download

2. SOCK_PACKET: Why not reading outgoing packets ?

3. Patch: linux-2.5.2-pre8/drivers/ieee1394 kdev_t compilation fixes

4. automake: different CFLAGS for different objects

5. Patch: linux-2.5.2-pre8/drivers/mtd compilation fixes

6. Why "blinking cursor" and "paren highlighting" are grayed in xemacs-devel-21.4.8.?

7. Patch: linux-2.5.2-pre8/drivers/sound compilation fixes: MINOR-->minor

8. MkLinux dinner proposal (Golden Master Candidate)

9. linux-2.5.2-pre8/drivers/telephony kdev_t compilation fixes

10. PATCH: linux-2.5.2-pre7/drivers/block/rd.c kdev_t fix

11. Patch?: linux-2.5.6-pre1/drivers/scsi/pci2000.c compilation fixes + pci device ID table

12. small patch to fix compilation error in drivers/char/rocket.c

13. patch for cs46xx driver on 2.5.47 to fix compilation error