2.6.28 /proc/diskstats

2.6.28 /proc/diskstats

[Daniel Collins]

Hi

Since 2.6.28, fd0 isn't visible in /proc/diskstats from boot, only after 
access/mount. It is visible in /sys/block/ from boot, so there are no 
issues detecting it. This may also affect other devices such as hda but 
I haven't tested this yet, my test system and QEMU both only have 
ram0-ram15 from boot. I have tested 2.6.30-rc1 and found the same 
behavior. Is this a bug or a deliberate change?

Thanks

Re: 2.6.28 /proc/diskstats

[Andrew Morton]

On Sun, 12 Apr 2009 02:00:08 +0100 Daniel Collins <solemnwarning@solemnwarning.no-ip.org> wrote:

> Hi
> 
> Since 2.6.28, fd0 isn't visible in /proc/diskstats from boot, only after 
> access/mount. It is visible in /sys/block/ from boot, so there are no 
> issues detecting it. This may also affect other devices such as hda but 
> I haven't tested this yet, my test system and QEMU both only have 
> ram0-ram15 from boot. I have tested 2.6.30-rc1 and found the same 
> behavior. Is this a bug or a deliberate change?
> 

Tejun touched it last :)

Re: 2.6.28 /proc/diskstats

[Daniel Collins]

Reposting since I don't think anyone will see it ~150 messages down, 
sorry if this is the wrong way to draw attention to a bug.

Daniel Collins wrote:
> Hi
>
> Since 2.6.28, fd0 isn't visible in /proc/diskstats from boot, only 
> after access/mount. It is visible in /sys/block/ from boot, so there 
> are no issues detecting it. This may also affect other devices such as 
> hda but I haven't tested this yet, my test system and QEMU both only 
> have ram0-ram15 from boot. I have tested 2.6.30-rc1 and found the same 
> behavior. Is this a bug or a deliberate change?
>
> Thanks

Re: 2.6.28 /proc/diskstats

[Daniel Collins]

I've stepped through the commits in linux-2.6, the first revision with 
this behavior is 074a7aca7afa6f230104e8e65eba3420263714a5

Daniel Collins wrote:
> Hi
>
> Since 2.6.28, fd0 isn't visible in /proc/diskstats from boot, only 
> after access/mount. It is visible in /sys/block/ from boot, so there 
> are no issues detecting it. This may also affect other devices such as 
> hda but I haven't tested this yet, my test system and QEMU both only 
> have ram0-ram15 from boot. I have tested 2.6.30-rc1 and found the same 
> behavior. Is this a bug or a deliberate change?
>
> Thanks

Re: 2.6.28 /proc/diskstats

[Sam Ravnborg]

On Mon, Apr 13, 2009 at 02:21:10AM +0100, Daniel Collins wrote:
> I've stepped through the commits in linux-2.6, the first revision with 
> this behavior is 074a7aca7afa6f230104e8e65eba3420263714a5

It would be a good idea to:
1) mention the subject of said path as no-one remember all these
SHA-1 anyway (but include the SHA-1 for reference.
2) copy the author of said patch.

The relevant patch is: "block: move stats from disk to part0"

I have copied Tejun and JEns.

The rest of the mail kept for reference.

	Sam


> 
> Daniel Collins wrote:
> >Hi
> >
> >Since 2.6.28, fd0 isn't visible in /proc/diskstats from boot, only 
> >after access/mount. It is visible in /sys/block/ from boot, so there 
> >are no issues detecting it. This may also affect other devices such as 
> >hda but I haven't tested this yet, my test system and QEMU both only 
> >have ram0-ram15 from boot. I have tested 2.6.30-rc1 and found the same 
> >behavior. Is this a bug or a deliberate change?
> >
> >Thanks
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: 2.6.28 /proc/diskstats

[Tejun Heo]

Daniel Collins wrote:
> Hi
> 
> Since 2.6.28, fd0 isn't visible in /proc/diskstats from boot, only after
> access/mount. It is visible in /sys/block/ from boot, so there are no
> issues detecting it. This may also affect other devices such as hda but
> I haven't tested this yet, my test system and QEMU both only have
> ram0-ram15 from boot. I have tested 2.6.30-rc1 and found the same
> behavior. Is this a bug or a deliberate change?

Hmm... the reason why fd device is skipped before the first open is
because diskstats_show() skips zero sized devices and open is when the
floppy driver sets the capacity.  I think the original code also
skipped zero sized device.  It could be that the original code
triggered revalidation while the current one doesn't.  I'll dig
deeper.

Thanks.

-- 
tejun

Re: 2.6.28 /proc/diskstats

[Tejun Heo]

/proc/diskstats used to show stats for all disks whether they're
zero-sized or not and their non-zero partitions.  Commit
074a7aca7afa6f230104e8e65eba3420263714a5 accidentally changed the
behavior such that it doesn't print out zero sized disks.  This patch
implements DISK_PITER_INCL_EMPTY_PART0 flag to partition iterator and
uses it in diskstats_show() such that empty part0 is shown in
/proc/diskstats.

Reported and bisectd by Dianel Collins.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Daniel Collins <solemnwarning@solemnwarning.no-ip.org>
---
It was my bad.  I missed that diskstats showed empty disks too.  This
fixes the problem.

 block/genhd.c         |   12 ++++++++----
 include/linux/genhd.h |    1 +
 2 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/block/genhd.c b/block/genhd.c
index a9ec910..1a4916e 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -98,7 +98,7 @@ void disk_part_iter_init(struct disk_part_iter *piter, struct gendisk *disk,
 
 	if (flags & DISK_PITER_REVERSE)
 		piter->idx = ptbl->len - 1;
-	else if (flags & DISK_PITER_INCL_PART0)
+	else if (flags & (DISK_PITER_INCL_PART0 | DISK_PITER_INCL_EMPTY_PART0))
 		piter->idx = 0;
 	else
 		piter->idx = 1;
@@ -134,7 +134,8 @@ struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter)
 	/* determine iteration parameters */
 	if (piter->flags & DISK_PITER_REVERSE) {
 		inc = -1;
-		if (piter->flags & DISK_PITER_INCL_PART0)
+		if (piter->flags & (DISK_PITER_INCL_PART0 |
+				    DISK_PITER_INCL_EMPTY_PART0))
 			end = -1;
 		else
 			end = 0;
@@ -150,7 +151,10 @@ struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter)
 		part = rcu_dereference(ptbl->part[piter->idx]);
 		if (!part)
 			continue;
-		if (!(piter->flags & DISK_PITER_INCL_EMPTY) && !part->nr_sects)
+		if (!part->nr_sects &&
+		    !(piter->flags & DISK_PITER_INCL_EMPTY) &&
+		    !(piter->flags & DISK_PITER_INCL_EMPTY_PART0 &&
+		      piter->idx == 0))
 			continue;
 
 		get_device(part_to_dev(part));
@@ -1011,7 +1015,7 @@ static int diskstats_show(struct seq_file *seqf, void *v)
 				"\n\n");
 	*/
  
-	disk_part_iter_init(&piter, gp, DISK_PITER_INCL_PART0);
+	disk_part_iter_init(&piter, gp, DISK_PITER_INCL_EMPTY_PART0);
 	while ((hd = disk_part_iter_next(&piter))) {
 		cpu = part_stat_lock();
 		part_round_stats(cpu, hd);
diff --git a/include/linux/genhd.h b/include/linux/genhd.h
index 634c530..a1a28ca 100644
--- a/include/linux/genhd.h
+++ b/include/linux/genhd.h
@@ -214,6 +214,7 @@ static inline void disk_put_part(struct hd_struct *part)
 #define DISK_PITER_REVERSE	(1 << 0) /* iterate in the reverse direction */
 #define DISK_PITER_INCL_EMPTY	(1 << 1) /* include 0-sized parts */
 #define DISK_PITER_INCL_PART0	(1 << 2) /* include partition 0 */
+#define DISK_PITER_INCL_EMPTY_PART0 (1 << 3) /* include empty partition 0 */
 
 struct disk_part_iter {
 	struct gendisk		*disk;

[PATCH] block: include empty disks in /proc/diskstats

[Tejun Heo]

/proc/diskstats used to show stats for all disks whether they're
zero-sized or not and their non-zero partitions.  Commit
074a7aca7afa6f230104e8e65eba3420263714a5 accidentally changed the
behavior such that it doesn't print out zero sized disks.  This patch
implements DISK_PITER_INCL_EMPTY_PART0 flag to partition iterator and
uses it in diskstats_show() such that empty part0 is shown in
/proc/diskstats.

Reported and bisectd by Dianel Collins.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Daniel Collins <solemnwarning@solemnwarning.no-ip.org>
---
It was my bad.  I missed that diskstats showed empty disks too.  This
fixes the problem.

(Oops, forgot to write SUBJECT, resending...)

 block/genhd.c         |   12 ++++++++----
 include/linux/genhd.h |    1 +
 2 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/block/genhd.c b/block/genhd.c
index a9ec910..1a4916e 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -98,7 +98,7 @@ void disk_part_iter_init(struct disk_part_iter *piter, struct gendisk *disk,
 
 	if (flags & DISK_PITER_REVERSE)
 		piter->idx = ptbl->len - 1;
-	else if (flags & DISK_PITER_INCL_PART0)
+	else if (flags & (DISK_PITER_INCL_PART0 | DISK_PITER_INCL_EMPTY_PART0))
 		piter->idx = 0;
 	else
 		piter->idx = 1;
@@ -134,7 +134,8 @@ struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter)
 	/* determine iteration parameters */
 	if (piter->flags & DISK_PITER_REVERSE) {
 		inc = -1;
-		if (piter->flags & DISK_PITER_INCL_PART0)
+		if (piter->flags & (DISK_PITER_INCL_PART0 |
+				    DISK_PITER_INCL_EMPTY_PART0))
 			end = -1;
 		else
 			end = 0;
@@ -150,7 +151,10 @@ struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter)
 		part = rcu_dereference(ptbl->part[piter->idx]);
 		if (!part)
 			continue;
-		if (!(piter->flags & DISK_PITER_INCL_EMPTY) && !part->nr_sects)
+		if (!part->nr_sects &&
+		    !(piter->flags & DISK_PITER_INCL_EMPTY) &&
+		    !(piter->flags & DISK_PITER_INCL_EMPTY_PART0 &&
+		      piter->idx == 0))
 			continue;
 
 		get_device(part_to_dev(part));
@@ -1011,7 +1015,7 @@ static int diskstats_show(struct seq_file *seqf, void *v)
 				"\n\n");
 	*/
  
-	disk_part_iter_init(&piter, gp, DISK_PITER_INCL_PART0);
+	disk_part_iter_init(&piter, gp, DISK_PITER_INCL_EMPTY_PART0);
 	while ((hd = disk_part_iter_next(&piter))) {
 		cpu = part_stat_lock();
 		part_round_stats(cpu, hd);
diff --git a/include/linux/genhd.h b/include/linux/genhd.h
index 634c530..a1a28ca 100644
--- a/include/linux/genhd.h
+++ b/include/linux/genhd.h
@@ -214,6 +214,7 @@ static inline void disk_put_part(struct hd_struct *part)
 #define DISK_PITER_REVERSE	(1 << 0) /* iterate in the reverse direction */
 #define DISK_PITER_INCL_EMPTY	(1 << 1) /* include 0-sized parts */
 #define DISK_PITER_INCL_PART0	(1 << 2) /* include partition 0 */
+#define DISK_PITER_INCL_EMPTY_PART0 (1 << 3) /* include empty partition 0 */
 
 struct disk_part_iter {
 	struct gendisk		*disk;

Re: [stable] [PATCH] block: include empty disks in /proc/diskstats

[Chris Wright]

* Tejun Heo (tj@kernel.org) wrote:
> /proc/diskstats used to show stats for all disks whether they're
> zero-sized or not and their non-zero partitions.  Commit
> 074a7aca7afa6f230104e8e65eba3420263714a5 accidentally changed the
> behavior such that it doesn't print out zero sized disks.  This patch
> implements DISK_PITER_INCL_EMPTY_PART0 flag to partition iterator and
> uses it in diskstats_show() such that empty part0 is shown in
> /proc/diskstats.
> 
> Reported and bisectd by Dianel Collins.
> 
> Signed-off-by: Tejun Heo <tj@kernel.org>
> Reported-by: Daniel Collins <solemnwarning@solemnwarning.no-ip.org>

looks like this isn't upstream yet.

thanks,
-chris

Re: [stable] [PATCH] block: include empty disks in /proc/diskstats

[Jens Axboe]

On Tue, Apr 21 2009, Chris Wright wrote:
> * Tejun Heo (tj@kernel.org) wrote:
> > /proc/diskstats used to show stats for all disks whether they're
> > zero-sized or not and their non-zero partitions.  Commit
> > 074a7aca7afa6f230104e8e65eba3420263714a5 accidentally changed the
> > behavior such that it doesn't print out zero sized disks.  This patch
> > implements DISK_PITER_INCL_EMPTY_PART0 flag to partition iterator and
> > uses it in diskstats_show() such that empty part0 is shown in
> > /proc/diskstats.
> > 
> > Reported and bisectd by Dianel Collins.
> > 
> > Signed-off-by: Tejun Heo <tj@kernel.org>
> > Reported-by: Daniel Collins <solemnwarning@solemnwarning.no-ip.org>
> 
> looks like this isn't upstream yet.

It'll go upstream today or tomorrow.

-- 
Jens Axboe