FreeBSD 10.1 Memory Exhaustion
Christopher Forgeron
csforgeron at gmail.com
Mon Jul 13 18:26:32 UTC 2015
I found the same thing - The older patch does apply on 10.1 with one line
needing to be added manually. Did that, but I'm running into this when I
try a make buildworld:
/usr/src/cddl/lib/libzpool/../../../sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c:2726:30:
error: use of undeclared identifier 'freemem'
if (zfs_arc_cache_target && freemem < zfs_arc_cache_target +
btop(size)) {
^
/usr/src/cddl/lib/libzpool/../../../sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c:2726:63:
warning: implicit declaration of function 'btop' is invalid in C99
[-Wimplicit-function-declaration]
if (zfs_arc_cache_target && freemem < zfs_arc_cache_target +
btop(size)) {
^
/usr/src/cddl/lib/libzpool/../../../sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c:2750:3:
warning: implicit declaration of function 'kmem_cache_reap' is invalid in
C99 [-Wimplicit-function-declaration]
kmem_cache_reap(cache, zfs_arc_cache_free_max);
^
/usr/src/cddl/lib/libzpool/../../../sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c:2812:14:
warning: implicit declaration of function 'kmem_cache_free_size' is invalid
in C99 [-Wimplicit-function-declaration]
cachefree = kmem_cache_free_size(buf_cache) +
I see freemem defined in sys/cddl/compat/opensolaris/sys/kmem.h, and I've
done a make clean; make clean; make cleanworld; make depend before running
make buildworld.
What am I forgetting to make here?
On Mon, Jul 13, 2015 at 11:15 AM, Matthew Seaman <
m.seaman at infracaninophile.co.uk> wrote:
> On 2015/07/13 14:55, Karl Denninger wrote:
> > The previous version of the patch MAY apply against RELENG/10.1.
>
> Close....
>
> backup-4:/usr/src:# svn patch ~matthew/arc_cache_reclaim_uma_stable10.patch
> U sys/cddl/compat/opensolaris/kern/opensolaris_kmem.c
> U sys/cddl/compat/opensolaris/sys/kmem.h
> U sys/cddl/compat/opensolaris/sys/param.h
> U sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
> > applied hunk @@ -592,6 +687,13 @@ with offset 2
> > applied hunk @@ -2421,6 +2523,7 @@ with offset 28
> > applied hunk @@ -2429,6 +2532,9 @@ with offset 28
> > applied hunk @@ -2439,12 +2545,19 @@ with offset 28
> > applied hunk @@ -2454,16 +2567,26 @@ with offset 28
> > applied hunk @@ -2491,8 +2614,19 @@ with offset 28
> > applied hunk @@ -2503,26 +2637,43 @@ with offset 28
> > applied hunk @@ -2529,6 +2680,7 @@ with offset 28
> > applied hunk @@ -2537,7 +2689,7 @@ with offset 28
> > applied hunk @@ -2552,20 +2704,131 @@ with offset 28
> > applied hunk @@ -2577,10 +2840,36 @@ with offset 28
> > applied hunk @@ -2588,6 +2877,7 @@ with offset 28
> > applied hunk @@ -2602,6 +2892,7 @@ with offset 28
> > applied hunk @@ -2618,6 +2909,7 @@ with offset 28
> > applied hunk @@ -2692,6 +2984,7 @@ with offset 28
> > applied hunk @@ -2713,20 +3006,6 @@ with offset 28
> > applied hunk @@ -2763,6 +3042,9 @@ with offset 28
> > applied hunk @@ -3885,20 +4167,16 @@ with offset 29
> > applied hunk @@ -3911,7 +4189,7 @@ with offset 29
> > applied hunk @@ -3939,8 +4217,10 @@ with offset 29
> > applied hunk @@ -3994,6 +4274,7 @@ with offset 29
> U sys/cddl/contrib/opensolaris/uts/common/fs/zfs/range_tree.c
> U sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_queue.c
> U sys/vm/uma.h
> U sys/vm/uma_core.c
> C sys/vm/vm_pageout.c
> > rejected hunk @@ -921,10 +932,11 @@
> Summary of conflicts:
> Text conflicts: 1
>
> ... so close in fact that it's a fairly obvious merge even for someone
> like me.
>
> Cheers,
>
> Matthew
>
>
>
More information about the freebsd-fs
mailing list