I've been using this since 2.4.22 and since laptop_mode is in the
kernel since 2.4.23-pre<something> and I haven't seen that anyone else
has implemented this so I decided to post it on the list in case anyone
is interested.
It a patch to modify the journal flush time of reiserfs to support
laptop_mode (same functionality as ext3 has already).
The times are taken from bdflush.
--- kernel-source-2.4.25-pre6/fs/reiserfs/journal.c 2003-08-25 14:44:43.000000000 +0300
+++ kernel-source-2.4.25-pre6.new/fs/reiserfs/journal.c 2003-10-20 03:35:58.000000000 +0200
@@ -58,6 +58,7 @@
#include <linux/stat.h>
#include <linux/string.h>
#include <linux/smp_lock.h>
+#include <linux/fs.h>
/* the number of mounted filesystems. This is used to decide when to
** start and kill the commit thread
@@ -2659,8 +2660,12 @@
/* starting with oldest, loop until we get to the start */
i = (SB_JOURNAL_LIST_INDEX(p_s_sb) + 1) % JOURNAL_LIST_COUNT ;
while(i != start) {
- if (SB_JOURNAL_LIST(p_s_sb)[i].j_len > 0 && ((now - SB_JOURNAL_LIST(p_s_sb)[i].j_timestamp) > SB_JOURNAL_MAX_COMMIT_AGE(p_s_sb) ||
- immediate)) {
+ /* get_buffer_age() / HZ is used since the time returned by
+ * get_buffer_age is in sec * HZ and the journal time is taken in seconds.
+ */
+ if (SB_JOURNAL_LIST(p_s_sb)[i].j_len > 0 &&
+ ((now - SB_JOURNAL_LIST(p_s_sb)[i].j_timestamp) > get_buffer_age() / HZ
+ || immediate)) {
/* we have to check again to be sure the current transaction did not change */
if (i != SB_JOURNAL_LIST_INDEX(p_s_sb)) {
flush_commit_list(p_s_sb, SB_JOURNAL_LIST(p_s_sb) + i, 1) ;
--- kernel-source-2.4.25-pre6/fs/reiserfs/procfs.c 2003-08-25 14:44:43.000000000 +0300
+++ kernel-source-2.4.25-pre6.new/fs/reiserfs/procfs.c 2003-10-20 03:36:11.000000000 +0200
@@ -18,6 +18,7 @@
#include <linux/locks.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
+#include <linux/fs.h>
#if defined( REISERFS_PROC_INFO )
@@ -528,7 +529,11 @@
DJP( jp_journal_trans_max ),
DJP( jp_journal_magic ),
DJP( jp_journal_max_batch ),
- DJP( jp_journal_max_commit_age ),
+ /* get_buffer_age() / HZ is used since the
+ * time returned by get_buffer_age is in sec * HZ
+ * and the journal time is in seconds.
+ */
+ get_buffer_age()/HZ,
DJP( jp_journal_max_trans_age ),
JF( j_1st_reserved_block ),
--- kernel-source-2.4.25-pre6/fs/buffer.c 2003-12-12 01:35:46.000000000 +0200
+++ kernel-source-2.4.25-pre6.new/fs/buffer.c 2003-12-12 01:41:20.000000000 +0200
@@ -1128,6 +1128,12 @@
}
EXPORT_SYMBOL(get_buffer_flushtime);
+inline int get_buffer_age(void)
+{
+ return bdf_prm.b_un.age_buffer;
+}
+EXPORT_SYMBOL(get_buffer_age);
+
/*
* A buffer may need to be moved from one buffer list to another
* (e.g. in case it is not shared any more). Handle this.
--- kernel-source-2.4.25-pre6/include/linux/fs.h 2003-12-12 01:45:20.000000000 +0200
+++ kernel-source-2.4.23-5-pre6.new/include/linux/fs.h 2003-12-12 01:45:42.000000000 +0200
@@ -1260,6 +1260,7 @@
extern void set_buffer_flushtime(struct buffer_head *);
extern inline int get_buffer_flushtime(void);
+extern inline int get_buffer_age(void);
extern void balance_dirty(void);
extern int check_disk_change(kdev_t);
extern int invalidate_inodes(struct super_block *);
Micha Feigin writes:
> I've been using this since 2.4.22 and since laptop_mode is in the
> kernel since 2.4.23-pre<something> and I haven't seen that anyone else
> has implemented this so I decided to post it on the list in case anyone
> is interested.
> It a patch to modify the journal flush time of reiserfs to support
> laptop_mode (same functionality as ext3 has already).
> The times are taken from bdflush.
Support for reiserfs laptop mode is in 2.6 now. It is done by adding new
mount option "commit=N" that sets commit interval in seconds.
It doesn't look right to just plainly set reiserfs commit interval to be
the same as the ext3 commit interval.
>
> --- kernel-source-2.4.25-pre6/fs/reiserfs/journal.c 2003-08-25 14:44:43.000000000 +0300
> +++ kernel-source-2.4.25-pre6.new/fs/reiserfs/journal.c 2003-10-20 03:35:58.000000000 +0200
> @@ -58,6 +58,7 @@
> #include <linux/stat.h>
> #include <linux/string.h>
> #include <linux/smp_lock.h>
> +#include <linux/fs.h>
>
> /* the number of mounted filesystems. This is used to decide when to
> ** start and kill the commit thread
> @@ -2659,8 +2660,12 @@
> /* starting with oldest, loop until we get to the start */
> i = (SB_JOURNAL_LIST_INDEX(p_s_sb) + 1) % JOURNAL_LIST_COUNT ;
> while(i != start) {
> - if (SB_JOURNAL_LIST(p_s_sb)[i].j_len > 0 && ((now - SB_JOURNAL_LIST(p_s_sb)[i].j_timestamp) > SB_JOURNAL_MAX_COMMIT_AGE(p_s_sb) ||
> - immediate)) {
> + /* get_buffer_age() / HZ is used since the time returned by
> + * get_buffer_age is in sec * HZ and the journal time is taken in seconds.
> + */
> + if (SB_JOURNAL_LIST(p_s_sb)[i].j_len > 0 &&
> + ((now - SB_JOURNAL_LIST(p_s_sb)[i].j_timestamp) > get_buffer_age() / HZ
> + || immediate)) {
Nikita.
On Mon, Jan 19, 2004 at 01:37:59PM +0300, Nikita Danilov wrote:
> Micha Feigin writes:
> > I've been using this since 2.4.22 and since laptop_mode is in the
> > kernel since 2.4.23-pre<something> and I haven't seen that anyone else
> > has implemented this so I decided to post it on the list in case anyone
> > is interested.
> > It a patch to modify the journal flush time of reiserfs to support
> > laptop_mode (same functionality as ext3 has already).
> > The times are taken from bdflush.
>
> Support for reiserfs laptop mode is in 2.6 now. It is done by adding new
> mount option "commit=N" that sets commit interval in seconds.
>
First of all its nice to know that laptop mode has finally made it
into 2.6. On the other hand setting commit=N on mount is not a good
solution since you want different flush times for when laptop mode is
activated and when it is disabled.
When laptop mode is disabled the default of 5 seconds is good since the
disk is always spinning. When laptop mode is enabled you want to change
the journal flush time to the linux buffer flush time so that the
journal won't keep waking up the disk. Its a bigger risk of loosing the
data so you don't want the longer journal flush time when laptop mode
isn't activated.
> It doesn't look right to just plainly set reiserfs commit interval to be
> the same as the ext3 commit interval.
I am not setting it to the same value as ext3. When laptop mode isn't
activated it is set to the default value used by reiserfs if bdflush
isn't modified (it does give you the ability to play with the flush
interval if you want even when reiserfs isn't activated).
If laptop mode is activated the flush time is set to the linux buffer
flush time so that the journal won't wake up the disk.
>
> >
> > --- kernel-source-2.4.25-pre6/fs/reiserfs/journal.c 2003-08-25 14:44:43.000000000 +0300
> > +++ kernel-source-2.4.25-pre6.new/fs/reiserfs/journal.c 2003-10-20 03:35:58.000000000 +0200
> > @@ -58,6 +58,7 @@
> > #include <linux/stat.h>
> > #include <linux/string.h>
> > #include <linux/smp_lock.h>
> > +#include <linux/fs.h>
> >
> > /* the number of mounted filesystems. This is used to decide when to
> > ** start and kill the commit thread
> > @@ -2659,8 +2660,12 @@
> > /* starting with oldest, loop until we get to the start */
> > i = (SB_JOURNAL_LIST_INDEX(p_s_sb) + 1) % JOURNAL_LIST_COUNT ;
> > while(i != start) {
> > - if (SB_JOURNAL_LIST(p_s_sb)[i].j_len > 0 && ((now - SB_JOURNAL_LIST(p_s_sb)[i].j_timestamp) > SB_JOURNAL_MAX_COMMIT_AGE(p_s_sb) ||
> > - immediate)) {
> > + /* get_buffer_age() / HZ is used since the time returned by
> > + * get_buffer_age is in sec * HZ and the journal time is taken in seconds.
> > + */
> > + if (SB_JOURNAL_LIST(p_s_sb)[i].j_len > 0 &&
> > + ((now - SB_JOURNAL_LIST(p_s_sb)[i].j_timestamp) > get_buffer_age() / HZ
> > + || immediate)) {
>
> Nikita.
>
Micha Feigin wrote:
> On Mon, Jan 19, 2004 at 01:37:59PM +0300, Nikita Danilov wrote:
>
>>Micha Feigin writes:
>> > I've been using this since 2.4.22 and since laptop_mode is in the
>> > kernel since 2.4.23-pre<something> and I haven't seen that anyone else
>> > has implemented this so I decided to post it on the list in case anyone
>> > is interested.
>> > It a patch to modify the journal flush time of reiserfs to support
>> > laptop_mode (same functionality as ext3 has already).
>> > The times are taken from bdflush.
>>
>>Support for reiserfs laptop mode is in 2.6 now. It is done by adding new
>>mount option "commit=N" that sets commit interval in seconds.
>>
>
>
> First of all its nice to know that laptop mode has finally made it
> into 2.6. On the other hand setting commit=N on mount is not a good
> solution since you want different flush times for when laptop mode is
> activated and when it is disabled.
> When laptop mode is disabled the default of 5 seconds is good since the
> disk is always spinning. When laptop mode is enabled you want to change
> the journal flush time to the linux buffer flush time so that the
> journal won't keep waking up the disk. Its a bigger risk of loosing the
> data so you don't want the longer journal flush time when laptop mode
> isn't activated.
>
>
>>It doesn't look right to just plainly set reiserfs commit interval to be
>>the same as the ext3 commit interval.
>
>
> I am not setting it to the same value as ext3. When laptop mode isn't
> activated it is set to the default value used by reiserfs if bdflush
> isn't modified (it does give you the ability to play with the flush
> interval if you want even when reiserfs isn't activated).
> If laptop mode is activated the flush time is set to the linux buffer
> flush time so that the journal won't wake up the disk.
A couple of comments:
* You might want to take a look at Hugang's patch to support laptop mode
on reiserfs in 2.6. This patch was posted somewhere last month. It adds
a "commit=" option to reiserfs, so that you can change this externally
by remounting. I think this solution is a lot cleaner than to have
reiserfs code directly depend on laptop mode.
* If you decide to port that patch, you might want to look at the
documentation for laptop mode in 2.6. The control script that's in there
works for both 2.4 and 2.6, and it automatically remounts your ext3 and
reiserfs filesystems with the appropriate commit options (and remounts
them without the commit options when laptop mode is stopped). (There is
also some ACPI event binding code in the documentation, which
automatically enables laptop_mode when using battery power and disables
it when your laptop is plugged in. I don't know if that works with 2.4
though.)
-- Bart
Bart Samwel writes:
>
[...]
> A couple of comments:
>
> * You might want to take a look at Hugang's patch to support laptop mode
> on reiserfs in 2.6. This patch was posted somewhere last month. It adds
> a "commit=" option to reiserfs, so that you can change this externally
> by remounting. I think this solution is a lot cleaner than to have
> reiserfs code directly depend on laptop mode.
This patch was merged into mainline.
>
>
> -- Bart
Nikita.