2007-02-27 20:52:58

by Eric Sandeen

Subject: [PATCH] use for_each_possible_cpu for reservation slots

NR_CPUS is a config-time option that may be much more than your actual
count of runtime possible CPUs...

compile-tested.

Signed-off-by: Eric Sandeen <[email protected]>

Index: linux-2.6-ext4/fs/ext4/balloc.c
===================================================================
--- linux-2.6-ext4.orig/fs/ext4/balloc.c
+++ linux-2.6-ext4/fs/ext4/balloc.c
@@ -1895,16 +1895,17 @@ void ext4_rebalance_reservation(struct e
__u64 chunk;

/* let's know what slots have been used */
- for (i = 0; i < NR_CPUS; i++)
+ for_each_possible_cpu(i) {
if (rs[i].rs_reserved || i == smp_processor_id())
used_slots++;
+ }

/* chunk is a number of block every used
* slot will get. make sure it isn't 0 */
chunk = free + used_slots - 1;
do_div(chunk, used_slots);

- for (i = 0; i < NR_CPUS; i++) {
+ for_each_possible_cpu(i) {
if (free < chunk)
chunk = free;
if (rs[i].rs_reserved || i == smp_processor_id()) {
@@ -1926,7 +1927,7 @@ int ext4_reserve_global(struct super_blo
rs = sbi->s_reservation_slots;

/* lock all slots */
- for (i = 0; i < NR_CPUS; i++) {
+ for_each_possible_cpu(i) {
spin_lock(&rs[i].rs_lock);
free += rs[i].rs_reserved;
}
@@ -1937,8 +1938,9 @@ int ext4_reserve_global(struct super_blo
rc = 0;
}

- for (i = 0; i < NR_CPUS; i++)
+ for_each_possible_cpu(i) {
spin_unlock(&rs[i].rs_lock);
+ }

return rc;
}
@@ -1979,12 +1981,14 @@ int ext4_reserve_init(struct super_block
struct ext4_reservation_slot *rs;
int i;

- rs = kmalloc(sizeof(struct ext4_reservation_slot) * NR_CPUS, GFP_KERNEL);
+ rs = kmalloc(sizeof(struct ext4_reservation_slot) *
+ highest_possible_processor_id()+1,
+ GFP_KERNEL);
if (rs == NULL)
return -ENOMEM;
sbi->s_reservation_slots = rs;

- for (i = 0; i < NR_CPUS; i++) {
+ for_each_possible_cpu(i) {
spin_lock_init(&rs[i].rs_lock);
rs[i].rs_reserved = 0;
}

2007-03-12 17:21:39

by Eric Sandeen

Subject: Re: [PATCH] use for_each_possible_cpu for reservation slots

Eric Sandeen wrote:
> NR_CPUS is a config-time option that may be much more than your actual
> count of runtime possible CPUs...
>
> compile-tested.
>
> Signed-off-by: Eric Sandeen <[email protected]>

Ted, I don't see this in the git tree yet.

Yea? Nay? Comments?

Thanks,
-Eric