On Fri, 2009-10-02 at 15:16 -0700, David Rientjes wrote:
> On Thu, 1 Oct 2009, Lee Schermerhorn wrote:
>
> > Index: linux-2.6.31-mmotm-090925-1435/mm/hugetlb.c
> > ===================================================================
> > --- linux-2.6.31-mmotm-090925-1435.orig/mm/hugetlb.c 2009-09-30 12:48:45.000000000 -0400
> > +++ linux-2.6.31-mmotm-090925-1435/mm/hugetlb.c 2009-10-01 12:13:25.000000000 -0400
> > @@ -1334,29 +1334,71 @@ static struct hstate *kobj_to_hstate(str
> > return NULL;
> > }
> >
> > -static ssize_t nr_hugepages_show(struct kobject *kobj,
> > +static ssize_t nr_hugepages_show_common(struct kobject *kobj,
> > struct kobj_attribute *attr, char *buf)
> > {
> > struct hstate *h = kobj_to_hstate(kobj);
> > return sprintf(buf, "%lu\n", h->nr_huge_pages);
> > }
> > -static ssize_t nr_hugepages_store(struct kobject *kobj,
> > - struct kobj_attribute *attr, const char *buf, size_t count)
> > +static ssize_t nr_hugepages_store_common(bool obey_mempolicy,
> > + struct kobject *kobj, struct kobj_attribute *attr,
> > + const char *buf, size_t len)
> > {
> > int err;
> > - unsigned long input;
> > + unsigned long count;
> > struct hstate *h = kobj_to_hstate(kobj);
> > + NODEMASK_ALLOC(nodemask, nodes_allowed);
> >
>
> [ FYI: I'm not sure clameter@xxxxxxx still works, you may want to try
> cl@xxxxxxxxxxxxxxxxxxxxx ]
>
>
> mm/hugetlb.c: In function 'nr_hugepages_store_common':
> mm/hugetlb.c:1368: error: storage size of '_m' isn't known
> mm/hugetlb.c:1380: warning: passing argument 1 of 'init_nodemask_of_mempolicy' from incompatible pointer type
> mm/hugetlb.c:1382: warning: assignment from incompatible pointer type
> mm/hugetlb.c:1390: warning: passing argument 1 of 'init_nodemask_of_node' from incompatible pointer type
> mm/hugetlb.c:1392: warning: passing argument 3 of 'set_max_huge_pages' from incompatible pointer type
> mm/hugetlb.c:1394: warning: comparison of distinct pointer types lacks a cast
> mm/hugetlb.c:1368: warning: unused variable '_m'
> mm/hugetlb.c: In function 'hugetlb_sysctl_handler_common':
> mm/hugetlb.c:1862: error: storage size of '_m' isn't known
> mm/hugetlb.c:1864: warning: passing argument 1 of 'init_nodemask_of_mempolicy' from incompatible pointer type
> mm/hugetlb.c:1866: warning: assignment from incompatible pointer type
> mm/hugetlb.c:1868: warning: passing argument 3 of 'set_max_huge_pages' from incompatible pointer type
> mm/hugetlb.c:1870: warning: comparison of distinct pointer types lacks a cast
> mm/hugetlb.c:1862: warning: unused variable '_m'
??? This is after your rework of NODEMASK_ALLOC has been applied? I
don't see this when I build the mmotm that the patch is based on.
>
> This can be fixed after my "nodemask: make NODEMASK_ALLOC more general"
> patch is merged and the following is applied as I suggested in
> http://marc.info/?l=linux-mm&m=125270872312494:
Ah, but your patch didn't exist back then :).
I guess I'll tack this onto the end of V9 with a note that it depends on
your patch. Altho' for bisection builds, I might want to break it into
separate patches that apply to the mempolicy and per node attributes
patches, respectively.
Thanks,
Lee
>
> Signed-off-by: David Rientjes <rientjes@xxxxxxxxxx>
> ---
> mm/hugetlb.c | 4 ++--
> 1 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> --- a/mm/hugetlb.c
> +++ b/mm/hugetlb.c
> @@ -1365,7 +1365,7 @@ static ssize_t nr_hugepages_store_common(bool obey_mempolicy,
> int nid;
> unsigned long count;
> struct hstate *h;
> - NODEMASK_ALLOC(nodemask, nodes_allowed);
> + NODEMASK_ALLOC(nodemask_t, nodes_allowed);
>
> err = strict_strtoul(buf, 10, &count);
> if (err)
> @@ -1859,7 +1859,7 @@ static int hugetlb_sysctl_handler_common(bool obey_mempolicy,
> proc_doulongvec_minmax(table, write, buffer, length, ppos);
>
> if (write) {
> - NODEMASK_ALLOC(nodemask, nodes_allowed);
> + NODEMASK_ALLOC(nodemask_t, nodes_allowed);
> if (!(obey_mempolicy &&
> init_nodemask_of_mempolicy(nodes_allowed))) {
> NODEMASK_FREE(nodes_allowed);
> --
> To unsubscribe from this list: send the line "unsubscribe linux-numa" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-numa" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
