Re: problems compiling enbd-2.4.32pre and fr1-2.16 under kernel 2.6.8.1

In article <[email protected]> you wrote:
> hi

Hi.

> i have a few problems when compiling the
> enbd-2.4.32pre and fr1-2.16, 

On ...

> using the kernel source 2.6.8.1 .

Oh.

> 1. when compiling the enbd without patching the kernel
> source, i found
>    a typo @ line 309, in the
> ./nbd-2.4.32/kernel/linux-2.6.x/Makefile. 
>    this is easy to correct

This line is:

# Helpers built in scripts/

$(srctree)/scripts/docproc $(srctree)/scripts/split-include : $(srctree)/)scripts ;

(it's the line above). Ah. Extra paren? Well done. I wonder where that
came from!


>    
>    $(srctree)/scripts/docproc
> $(srctree)/scripts/split-include : $(srctree)/)scripts
> ;

Yes.


> 2. after i modified the Makefile and compiled it
> again, the make said she 
>    could not find
> ./nbd-2.4.32/kernel/linux-2.6.x/scripts/fixdep.c,
> which's 
>    needed to build fixdep. this time i don't know how
> to fix it. 

This is part of your kernel source. It comes from there. I don't think
you would normally need it, since it would be in the kernel dir.

>    so i patched the kernel and built it, it seems ok.

There's no need to do that. I'm more interested in why you don't have
fixdep.c.

  % locate fixdep.c
  /usr/local/src/linux-2.6.0-test2/scripts/fixdep.c
  /usr/local/src/linux-2.6.1/scripts/fixdep.c
  /usr/local/src/linux-2.6.3/scripts/fixdep.c
  /usr/local/src/linux-2.6.8.1-uml/scripts/basic/fixdep.c

etc. If you don't have it, you don't have kernel source, so how can you
compile?


> 3. when compiling the fr1 driver with raid1 robust
> read enabled, i found
>    a undeclared var 'rdev' in function map in raid1.c.

Oh, that's an old one. It was supposed to be *rdevp. It's been fixed in
FR1 patches for a while.

> i checked the mailing
>    list and found the solution. so it's ok.

Did you add the other change needed? The one that adds an assignment to
rdev before calling  map?

@@ -925,6 +1249,9 @@
                } else {
                        int disk;
                        bio = r1_bio->bios[r1_bio->read_disk];
+#ifdef CONFIG_MD_RAID1_ROBUST_READ
+                       rdev = conf->mirrors[r1_bio->read_disk].rdev;
+#endif /* CONFIG_MD_RAID1_ROBUST_READ */
                        if ((disk=map(mddev, &rdev)) == -1) {
                                printk(KERN_ALERT "raid1: %s: unrecoverable I/O"
                                       " read error for block %llu\n",

The latest FR1 patch should have both things in.



> 4. then i rebuilt the kernel. and this time, every
> thing seemed to be fine
>    except there is an undefined symbole '__udivdi3' in
> the md.c @ line 3692

This comes from your compiler not doing division adequately.

>    
>    realspeed = (j - mddev->resync_mark_cnt -
> atomic_read(&md_throttle[mddev->md_minor]))/2/((jiffies-mddev->resync_mark)/HZ
> +1) +1;
>    
>    how could i get rid of that?

Fix compiler! It probably can't divide by a variable value. But anyway,
why isn't it foxed by the line before?

                currspeed = (j-mddev->resync_mark_cnt)/2/((jiffies-mddev->resync_mark)/HZ +1) +1;
#ifdef MD_BITMAP_SUPPORT
                /*
                 * PTB some of the blocks are skipped, not synced, so
                 * should not count when limiting i/o. Let personality
                 * say.
                 */
                realspeed = (j - mddev->resync_mark_cnt - atomic_read(&md_thrott
le[mdidx(mddev)]))/2/((jiffies-mddev->resync_mark)/HZ +1) +1;
#endif /* MD_BITMAP_SUPPORT */

The only difference between the currspeed and realspeed calculations is
the "- atomic_read(&md_throttle[mdidx(mddev)]" in the numerator. If
your compiler does one without problems, the other should also be no
problem to it!

So split the calculation up a bit and see if you can help the compiler
out. What compiler is it?


Peter