__f2fs_replace_block() cannot handle NEW_ADDR
as new_blkaddr properly and a kernel panic occurrs.
Signed-off-by: Daeho Jeong
Tested-by: Shu Tan
Reviewed-by: Chao Yu
---
fs/f2fs/segment.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
still linked in
the global dirty list of f2fs and this will cause a kernel panic.
So, we will prevent the newly created inode from being dirtied during
the FI_NEW_INODE flag of the inode is set. We will make it dirty
right after the flag is cleared.
Signed-off-by: Daeho Jeong
Signed-off-by: Young
still linked in
the global dirty list of f2fs and this will cause a kernel panic.
So, we will prevent the newly created inode from being dirtied during
the FI_NEW_INODE flag of the inode is set. We will make it dirty
right after the flag is cleared.
Signed-off-by: Daeho Jeong
Signed-off-by: Young
__f2fs_replace_block() cannot handle NEW_ADDR
as new_blkaddr properly and a kernel panic occurrs.
Signed-off-by: Daeho Jeong
Tested-by: Shu Tan
---
fs/f2fs/segment.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index c117e09..463f420
Hi Chao,
> Original intention here is to recover status to the timing before
> committing atomic write. As at that timing blkaddr in dnode should be
> cur->old_addr(NEW_ADDR), so we need to change to call:
> f2fs_update_data_blkaddr(&dn, NEW_ADDR);
Ok, I'll change NULL_ADDR to NEW_ADDR.
Thanks
__f2fs_replace_block() cannot handle NEW_ADDR
as new_blkaddr properly and a kernel panic occurrs.
Signed-off-by: Daeho Jeong
Tested-by: Shu Tan
---
fs/f2fs/segment.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index c117e09..0673d08
From: Daeho Jeong
As checkpoint=merge comes in, mount option setting related to
checkpoint had been mixed up. Fixed it.
Signed-off-by: Daeho Jeong
---
fs/f2fs/super.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index
From: Daeho Jeong
It causes a crash to change the ioprio of checkpoint thread not in
checkpoint=merge. I fixed that to prevent setting the ioprio of the
thread when checkpoint=merge is not enabled.
Signed-off-by: Daeho Jeong
---
fs/f2fs/sysfs.c | 8 ++--
1 file changed, 6 insertions(+), 2
fsync_mode, whint_mode and etc.
So, the answer will be "checkpoint=enable". What do you think?
2021년 2월 1일 (월) 오후 9:40, Chao Yu 님이 작성:
>
> On 2021/2/1 8:06, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > As checkpoint=merge comes in, mount option setting relate
heckpoint=disable
remount checkpoint=merge,checkpoint=enable => checkpoint=enable
Like
mount fsync_mode=posix, fsync_mode=strict, fsync_mode=nobarrier =>
fsync_mode=nobarrier
2021년 2월 2일 (화) 오전 5:11, Jaegeuk Kim 님이 작성:
>
> On 02/01, Daeho Jeong wrote:
> > Actually, I think we ne
For less confusion, I am going to separate the "merge" option from
"checkpoint=".
I am adding another "ckpt_merge" option. :)
2021년 2월 2일 (화) 오전 8:33, Daeho Jeong 님이 작성:
>
> The rightmost one is the final option. And checkpoint=merge means
> checkpo
From: Daeho Jeong
As checkpoint=merge comes in, mount option setting related to checkpoint
had been mixed up and it became hard to understand. So, I separated
this option from "checkpoint=" and made another mount option
"checkpoint_merge" for this.
Signed-off-by: Daeho Je
little radical. :)
What do you think?
2021년 2월 2일 (화) 오후 4:40, Chao Yu 님이 작성:
>
> On 2021/2/2 13:18, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > As checkpoint=merge comes in, mount option setting related to checkpoint
> > had been mixed up and it became hard
If I understand it correctly, the only thing I have to do now is
remove "nocheckpoint_merge" now.
Am I correct? :)
2021년 2월 2일 (화) 오후 5:30, Chao Yu 님이 작성:
>
> On 2021/2/2 16:02, Daeho Jeong wrote:
> > I chose the same step with "flush_merge", because it doesn&#x
From: Daeho Jeong
As checkpoint=merge comes in, mount option setting related to checkpoint
had been mixed up and it became hard to understand. So, I separated
this option from "checkpoint=" and made another mount option
"checkpoint_merge" for this.
Signed-off-by: Daeho Je
ot;data" argument in the below.
f2fs_remount(struct super_block *sb, int *flags, char *data)
If we don't provide the "nocheckpoint_merge" option, how can we turn
off the "checkpoint_merge" option which is turned on in the previous
mount?
2021년 2월 2일 (화) 오후 6:28, Chao
mount -o remount,nocheckpoint_merge /dir"
It would be more convenient to users. What do you think?
2021년 2월 2일 (화) 오후 6:55, Chao Yu 님이 작성:
>
> On 2021/2/2 17:44, Daeho Jeong wrote:
> > When we remount it without the "checkpoint_merge" option, shouldn't we
> > nee
From: Daeho Jeong
As checkpoint=merge comes in, mount option setting related to checkpoint
had been mixed up and it became hard to understand. So, I separated
this option from "checkpoint=" and made another mount option
"checkpoint_merge" for this.
Signed-off-by: Daeho Je
From: Daeho Jeong
As checkpoint=merge comes in, mount option setting related to checkpoint
had been mixed up and it became hard to understand. So, I separated
this option from "checkpoint=" and made another mount option
"checkpoint_merge" for this.
Signed-off-by: Daeho Je
From: Daeho Jeong
Found a new segemnt allocation without f2fs_lock_op() in
expand_inode_data(). So, when we do fallocate() for a pinned file
and trigger checkpoint very frequently and simultaneously. F2FS gets
stuck in the below code of do_checkpoint() forever.
f2fs_sync_meta_pages(sbi, META
From: Daeho Jeong
When we use F2FS_IOC_RELEASE_COMPRESS_BLOCKS ioctl, if we can't find
any compressed blocks in the file even with large file size, the
ioctl just ends up without changing the file's status as immutable.
It makes the user, who expects that the file is immutable when
From: Daeho Jeong
By profiling f2fs compression works, I've found vmap() callings are
bottlenecks of f2fs decompression path. Changing these with
vm_map_ram(), we can enhance f2fs decompression speed pretty much.
[Verification]
dd if=/dev/zero of=dummy bs=1m count=1000
echo 3 > /pro
) 오후 4:55, Chao Yu 님이 작성:
>
> On 2020/8/11 15:15, Gao Xiang wrote:
> > On Tue, Aug 11, 2020 at 12:37:53PM +0900, Daeho Jeong wrote:
> >> From: Daeho Jeong
> >>
> >> By profiling f2fs compression works, I've found vmap() callings are
> >> bottle
Plus, when we use vmap(), vmap() normally executes in a short time
like vm_map_ram().
But, sometimes, it has a very long delay.
2020년 8월 11일 (화) 오후 6:28, Daeho Jeong 님이 작성:
>
> Actually, as you can see, I use the whole zero data blocks in the test file.
> It can maximize the effect of
Sure, I'll update the test condition as you said in the commit message.
FYI, the test is done with 16kb chunk and Pixel 3 (arm64) device.
Thanks,
2020년 8월 11일 (화) 오후 7:18, Gao Xiang 님이 작성:
>
> On Tue, Aug 11, 2020 at 06:33:26PM +0900, Daeho Jeong wrote:
> > Plus, when we
20 at 08:21:23PM +0900, Daeho Jeong wrote:
> > Sure, I'll update the test condition as you said in the commit message.
> > FYI, the test is done with 16kb chunk and Pixel 3 (arm64) device.
>
> Yeah, anyway, it'd better to lock the freq and offline the little
> cores
From: Daeho Jeong
By profiling f2fs compression works, I've found vmap() callings have
unexpected hikes in the execution time in our test environment and
those are bottlenecks of f2fs decompression path. Changing these with
vm_map_ram(), we can enhance f2fs decompression speed pretty
From: Daeho Jeong
By profiling f2fs compression works, I've found vmap() callings have
unexpected hikes in the execution time in our test environment and
those are bottlenecks of f2fs decompression path. Changing these with
vm_map_ram(), we can enhance f2fs decompression speed pretty
From: Daeho Jeong
Added "ckpt_thread_ioprio" sysfs node to give a way to change checkpoint
merge daemon's io priority. Its default value is "be,3", which means
"BE" I/O class and I/O priority "3". We can select the class between "rt"
and
From: Daeho Jeong
We've added a new mount option "checkpoint=merge", which creates a
kernel daemon and makes it to merge concurrent checkpoint requests as
much as possible to eliminate redundant checkpoint issues. Plus, we
can eliminate the sluggish issue caused by slow checkpoint
Got it~ :)
2021년 1월 11일 (월) 오후 6:50, Jaegeuk Kim 님이 작성:
>
> On 01/11, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > We've added a new mount option "checkpoint=merge", which creates a
> > kernel daemon and makes it to merge concurrent checkpoint
2021년 1월 11일 (월) 오후 6:34, Chao Yu 님이 작성:
>
> On 2021/1/11 13:15, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > We've added a new mount option "checkpoint=merge", which creates a
> > kernel daemon and makes it to merge concurrent checkpoint
more making sense.
Thanks,
2020년 12월 8일 (화) 오전 5:31, Eric Biggers 님이 작성:
>
> On Sat, Dec 05, 2020 at 01:26:26PM +0900, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > I found out f2fs_free_dic() is invoked in a wrong timing, but
> > f2fs_verify_bio() still neede
12월 8일 (화) 오후 12:51, Eric Biggers 님이 작성:
>
> On Tue, Dec 01, 2020 at 01:08:02PM +0900, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > We will add a new "compress_mode" mount option to control file
> > compression mode. This supports "fs"
From: Daeho Jeong
We've added a new mount option "checkpoint=merge", which creates a
kernel daemon and makes it to merge concurrent checkpoint requests as
much as possible to eliminate redundant checkpoint issues. Plus, we
can eliminate the sluggish issue caused by slow checkpoint
Sorry, I think you were definitely confused.
There is no change for 2/2. So I didn't send it.
Next time, I am going to send it all, even though some of them don't
have changes.
2021년 1월 20일 (수) 오전 4:08, Jaegeuk Kim 님이 작성:
>
> Is there v4 2/2?
>
> On 01/19, Daeho Jeong
From: Daeho Jeong
We've added a new mount option "checkpoint=merge", which creates a
kernel daemon and makes it to merge concurrent checkpoint requests as
much as possible to eliminate redundant checkpoint issues. Plus, we
can eliminate the sluggish issue caused by slow checkpoint
From: Daeho Jeong
Added "ckpt_thread_ioprio" sysfs node to give a way to change checkpoint
merge daemon's io priority. Its default value is "be,3", which means
"BE" I/O class and I/O priority "3". We can select the class between "rt"
and
From: Daeho Jeong
Added acc_compr_inodes to show accumulated compressed inode count and
acc_compr_blocks to show accumulated secured block count with
compression in sysfs. These can be re-initialized to "0" by writing "0"
value in one of both.
Signed-off-by: Daeho Jeon
From: Daeho Jeong
Added acc_compr_inodes to show accumulated compressed inode count and
acc_compr_blocks to show accumulated secured block count with
compression in sysfs. These can be re-initialized to "0" by writing "0"
value in one of both.
Signed-off-by: Daeho Jeong
--
From: Daeho Jeong
We will add a new "compress_mode" mount option to control file
compression mode. This supports "fs" and "user". In "fs" mode (default),
f2fs does automatic compression on the compression enabled files.
In "user" mode, f2fs di
From: Daeho Jeong
Added two ioctl to decompress/compress explicitly the compression
enabled file in "compress_mode=user" mount option.
Using these two ioctls, the users can make a control of compression
and decompression of their files.
Signed-off-by: Daeho Jeong
---
v2: reformed c
description about these in
Documentation/filesystems/f2fs.rst and I implemented tests internally.
2020년 11월 24일 (화) 오전 3:48, Eric Biggers 님이 작성:
>
> On Mon, Nov 23, 2020 at 12:17:51PM +0900, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > Added two ioctl to decompress/compress ex
Jaegeuk,
Got it.
Eric,
Yep.
2020년 11월 24일 (화) 오전 3:46, Eric Biggers 님이 작성:
>
> On Mon, Nov 23, 2020 at 12:17:50PM +0900, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > We will add a new "compress_mode" mount option to control file
> > compression
2020년 11월 24일 (화) 오전 8:29, Eric Biggers 님이 작성:
>
> On Tue, Nov 24, 2020 at 08:02:21AM +0900, Daeho Jeong wrote:
> > Jaegeuk,
> >
> > My mistake~
> >
> > Eric,
> >
> > What I want is like do_page_cache_ra(), but I used
> > page_cache_ra_unbounde
ode) &&
idx < DIV_ROUND_UP(inode->i_size, PAGE_SIZE);
}
2020년 12월 4일 (금) 오후 12:28, Eric Biggers 님이 작성:
>
> On Fri, Dec 04, 2020 at 09:58:47AM +0900, Daeho Jeong wrote:
> > diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c
> > index 87090da8693d..cd
> Are you sure? I thought that compression (and encryption) apply to the whole
> file, including any Merkle tree blocks past i_size.
This "dic" structure is only for "de"compression, so we don't need to
worry about going beyond i_size case.
> Also, even if you include the i_size check, it's stil
Eric,
I have another question.
I understand enabling the verity can be possible in the middle of I/O.
Is the opposite way also possible? Actually, I couldn't find any
disabling function of it, though.
2020년 12월 4일 (금) 오후 1:31, Daeho Jeong 님이 작성:
>
> > Are you sure? I thought th
ri, Dec 04, 2020 at 01:48:24PM +0900, Daeho Jeong wrote:
> > Eric,
> >
> > I have another question.
> > I understand enabling the verity can be possible in the middle of I/O.
> > Is the opposite way also possible? Actually, I couldn't find any
> > disabling fun
Thanks for the explanation about verity.
I got your point. Thanks~
2020년 12월 4일 (금) 오후 2:18, Eric Biggers 님이 작성:
>
> On Fri, Dec 04, 2020 at 02:00:34PM +0900, Daeho Jeong wrote:
> > I think I don't understand how verity works.
> > Right after verity is enabled on a fi
From: Daeho Jeong
I found out f2fs_free_dic() is invoked in a wrong timing, but
f2fs_verify_bio() still needed the dic info and it triggered the
below kernel panic. It has been caused by the race condition of
pending_pages value between decompression and verity logic, when
the same compression
As I told, this patch has a bug. Ignore this one. So, I'm implementing
another one. :(
2020년 12월 4일 (금) 오후 6:50, Daeho Jeong 님이 작성:
>
> As I told, this patch has a bug. Ignore this one. So, I'm implementing
> another one. :(
>
>
> 2020년 12월 4일 (금) 오후 5:04, Chao Yu 님이
Yep, we need to come back to v1 and enable verity in a unit of cluster.
Plus, as I told you, I'll prevent newly verity enalbed pages from
being merged with verity disabled bio.
Thanks,
2020년 12월 5일 (토) 오전 3:29, Jaegeuk Kim 님이 작성:
>
> On 12/04, Daeho Jeong wrote:
> > Thanks fo
From: Daeho Jeong
I found out f2fs_free_dic() is invoked in a wrong timing, but
f2fs_verify_bio() still needed the dic info and it triggered the
below kernel panic. It has been caused by the race condition of
pending_pages value between decompression and verity logic, when
the same compression
> It looks like it will be better to move this into merge condition?
>
> if (bio && (!page_is_mergeable(sbi, bio,
> *last_block_in_bio, blkaddr) ||
> !f2fs_crypt_mergeable_bio(bio, inode, page->index, NULL)
> ||
>
gt; On 2020/11/23 11:17, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > Added two ioctl to decompress/compress explicitly the compression
> > enabled file in "compress_mode=user-based" mount option.
> >
> > Using these two ioctls, the users can ma
return
"false" to prevent compression write, so we don't use
f2fs_write_compressed_pages() anymore in this case.
Because of this, I manually updated i_compr_blocks. Do you have any
suggestions on this?
2020년 11월 26일 (목) 오후 2:04, Daeho Jeong 님이 작성:
>
> Eric,
>
> do_pag
Yep, you're right~ :)
2020년 12월 3일 (목) 오전 10:16, Chao Yu 님이 작성:
>
> On 2020/12/2 18:54, Daeho Jeong wrote:
> > We might use compress_extension=*,compress_option=user.
> > In this option, we're gonna allocate all the writes in cold zone.
>
> Oh, so all files i
backend_ready(inode))
> f2fs_warn(sbi, "compression algorithm is successfully set, "
> "but current kernel doesn't support this algorithm.");
> out:
> --
> 2.26.2
>
>
>
>
>
> On 2020/10/30 12:10, Dae
From: Daeho Jeong
Added two ioctl to decompress/compress explicitly the compression
enabled file in "compress_mode=user" mount option.
Using these two ioctls, the users can make a control of compression
and decompression of their files.
Signed-off-by: Daeho Jeong
---
v3: cha
From: Daeho Jeong
I found out f2fs_free_dic() is invoked in a wrong timing, but
f2fs_verify_bio() still needed the dic info and it triggered the
below kernel panic. It has been caused by the race condition of
pending_pages value between decompression and verity logic, when
the same compression
We're going to use these just accumulating numbers and periodically
gather and reset them.
Yes, I will add the description like you suggested.
Thanks,
2021년 3월 12일 (금) 오후 6:50, Chao Yu 님이 작성:
>
> On 2021/3/11 10:32, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> >
From: Daeho Jeong
I've added new sysfs nodes to show runtime compression stat since mount.
compr_written_block - show the block count written after compression
compr_saved_block - show the saved block count with compression
compr_new_inode - show the count of inode newly enabled for compre
e (0)
I needed a protection here, because they might be updated in the race condition.
2021년 3월 12일 (금) 오후 9:39, Greg KH 님이 작성:
>
> On Fri, Mar 12, 2021 at 09:25:31PM +0900, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > I've added new sysfs nodes to show runtime compres
lock = 0;
+blocks(3);
+ blocks(2);
sbi->compr_written_block = 3;
sbi->compr_written_block = 2;
Finally, we end up with 2, not 5.
As more threads are participating it, we might miss more counting.
2021년 3월 12일 (금) 오후 11:04, Greg KH 님이 작성:
>
> On Fri, Mar 12, 2021 at 10:5
So, do you want we protect the values here with spin_lock and just
read without spin_lock in sysfs read part?
2021년 3월 12일 (금) 오후 11:37, Daeho Jeong 님이 작성:
>
> As you can see, if we're doing like the below.
>
> sbi->compr_written_block += blocks;
>
> Let'
.
> Q: What is the most annoying thing in e-mail?
>
> A: No.
> Q: Should I include quotations after my reply?
>
> http://daringfireball.net/2007/07/on_top
>
Thanks for letting me know this!
>
> On Fri, Mar 12, 2021 at 11:37:29PM +0900, Daeho Jeong wrote:
> > As you
From: Daeho Jeong
I've added new sysfs nodes to show runtime compression stat since mount.
compr_written_block - show the block count written after compression
compr_saved_block - show the saved block count with compression
compr_new_inode - show the count of inode newly enabled for compre
Oops, it disappeared while versioning up...
2021년 1월 21일 (목) 오후 7:30, Chao Yu 님이 작성:
>
> On 2021/1/14 14:23, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > Added "ckpt_thread_ioprio" sysfs node to give a way to change checkpoint
> > merge daemon&#
From: Daeho Jeong
I've added new sysfs nodes to show runtime compression stat since mount.
compr_written_block - show the block count written after compression
compr_saved_block - show the saved block count with compression
compr_new_inode - show the count of inode newly enabled for compre
From: Daeho Jeong
Added "ckpt_thread_ioprio" sysfs node to give a way to change checkpoint
merge daemon's io priority. Its default value is "be,3", which means
"BE" I/O class and I/O priority "3". We can select the class between "rt"
and
From: Daeho Jeong
We've added a new mount option "checkpoint=merge", which creates a
kernel daemon and makes it to merge concurrent checkpoint requests as
much as possible to eliminate redundant checkpoint issues. Plus, we
can eliminate the sluggish issue caused by slow checkpoint
Got it~! :)
2021년 1월 22일 (금) 오전 9:55, Chao Yu 님이 작성:
>
> If there is no update in v5, it will be better to include my Reviewed-by
> tag, please note that. :)
>
> Thanks,
>
> On 2021/1/21 21:45, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > We'v
Yep~
2021년 1월 22일 (금) 오전 10:46, Jaegeuk Kim 님이 작성:
>
> On 01/21, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > We've added a new mount option "checkpoint=merge", which creates a
> > kernel daemon and makes it to merge concurrent checkpoint
2021년 3월 9일 (화) 오후 6:22, Chao Yu 님이 작성:
>
> On 2021/3/5 10:24, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > Added acc_compr_inodes to show accumulated compressed inode count and
> > acc_compr_blocks to show accumulated secured block count with
>
> I
We've reconsidered this feature and decided to get just runtime info
of them, not persistent on disk.
I am re-writing it.
Thanks,
2021년 3월 10일 (수) 오전 10:31, Chao Yu 님이 작성:
>
> On 2021/3/9 21:00, Daeho Jeong wrote:
> > 2021년 3월 9일 (화) 오후 6:22, Chao Yu 님이 작성:
> >>
>
ollowing commit:
>
> commit 5fdb322ff2c2b4ad519f490dcb7ebb96c5439af7
> Author: Daeho Jeong
> Date: Thu Dec 3 15:56:15 2020 +0900
>
> f2fs: add F2FS_IOC_DECOMPRESS_FILE and F2FS_IOC_COMPRESS_FILE
>
> The analysis is as follows:
>
> 4025 static int redirty_blocks(str
From: Daeho Jeong
Fixed null page reference when find_lock_page() fails in
redirty_blocks().
Signed-off-by: Daeho Jeong
Reported-by: Colin Ian King
Fixes: 5fdb322ff2c2 ("f2fs: add F2FS_IOC_DECOMPRESS_FILE and
F2FS_IOC_COMPRESS_FILE")
---
fs/f2fs/file.c | 4 +++-
1 file changed, 3
Yes, it's better~ :)
2021년 1월 5일 (화) 오전 10:44, Chao Yu 님이 작성:
>
> On 2021/1/5 9:28, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > Fixed null page reference when find_lock_page() fails in
> > redirty_blocks().
> >
> > Signed-off-by: Daeho Je
From: Daeho Jeong
Fixed null page reference when find_lock_page() fails in
redirty_blocks().
Signed-off-by: Daeho Jeong
Reported-by: Colin Ian King
Fixes: 5fdb322ff2c2 ("f2fs: add F2FS_IOC_DECOMPRESS_FILE and
F2FS_IOC_COMPRESS_FILE")
---
v2: changed error value and break the loop
2021년 1월 15일 (금) 오후 6:22, Chao Yu 님이 작성:
>
> On 2021/1/14 14:23, Daeho Jeong wrote:
> > From: Daeho Jeong
> >
> > We've added a new mount option "checkpoint=merge", which creates a
> > kernel daemon and makes it to merge concurrent checkpoint
2021년 1월 15일 (금) 오후 11:00, Daeho Jeong 님이 작성:
>
> 2021년 1월 15일 (금) 오후 6:22, Chao Yu 님이 작성:
> >
> > On 2021/1/14 14:23, Daeho Jeong wrote:
> > > From: Daeho Jeong
> > >
> > > We've added a new mount option "checkpoint=merge", which cr
From: Daeho Jeong
Added "ckpt_thread_ioprio" sysfs node to give a way to change checkpoint
merge daemon's io priority. Its default value is "be,3", which means
"BE" I/O class and I/O priority "3". We can select the class between "rt"
and
Sorry, ignore this.
I've sent this wrong. :(
2021년 1월 16일 (토) 오전 9:58, Daeho Jeong 님이 작성:
>
> From: Daeho Jeong
>
> Added "ckpt_thread_ioprio" sysfs node to give a way to change checkpoint
> merge daemon's io priority. Its default value is "be,3", whi
From: Daeho Jeong
We've added a new mount option "checkpoint=merge", which creates a
kernel daemon and makes it to merge concurrent checkpoint requests as
much as possible to eliminate redundant checkpoint issues. Plus, we
can eliminate the sluggish issue caused by slow checkpoint
Got it~
Thanks,
2021년 1월 5일 (화) 오후 9:06, Markus Elfring 님이 작성:
>
> > Fixed null page reference when find_lock_page() fails in
> > redirty_blocks().
>
> I suggest to choose an other imperative wording for this change description.
>
> See also:
> https://git.kernel.org/pub/scm/linux/kernel/git/torv
From: Daeho Jeong
By Colin's static analysis, we found out there is a null page reference
under low memory situation in redirty_blocks. I've made the page finding
loop stop immediately and return an error not to cause further memory
pressure when we run into a failure to find a page
Thanks for the notice~
2021년 1월 6일 (수) 오전 10:04, Chao Yu 님이 작성:
>
> Daeho,
>
> FYI
>
> https://www.spinics.net/lists/kernel/msg3595944.html
>
> On 2021/1/5 20:04, Markus Elfring wrote:
> >> Fixed null page reference when find_lock_page() fails in
> >> redirty_blocks().
> >
> > I suggest to choose
misaligned
reads reading more pages in advance.
Thanks,
2020년 11월 27일 (금) 오전 2:49, Eric Biggers 님이 작성:
>
> On Thu, Nov 26, 2020 at 02:04:41PM +0900, Daeho Jeong wrote:
> > Eric,
> >
> > do_page_cache_ra() is defined in mm/internal.h for internal use
> > between in mm, so we
Re-thinking about this, page_cache_sync_readahead() is not good for
our situation, it might end up with cluster misaligned reads which
trigger internal duplicated cluster reads.
2020년 11월 27일 (금) 오전 8:46, Daeho Jeong 님이 작성:
>
> Chao,
>
> Got it~
>
> Eric,
>
> Actually
Sure, but I don't think we need to expose compr_inode and compr_block right now.
2020년 11월 27일 (금) 오후 6:44, Chao Yu 님이 작성:
>
> Daeho,
>
> How about updating this patch based on below patch?
>
> f2fs: introduce a new per-sb directory in sysfs
>
> On 2020/1
From: Daeho Jeong
We will add a new "compress_mode" mount option to control file
compression mode. This supports "fs-based" and "user-based".
In "fs-based" mode (default), f2fs does automatic compression on
the compression enabled files. In "
From: Daeho Jeong
Added two ioctl to decompress/compress explicitly the compression
enabled file in "compress_mode=user-based" mount option.
Using these two ioctls, the users can make a control of compression
and decompression of their files.
Signed-off-by: Daeho Jeong
---
fs/f
know.
--
From: Daeho Jeong
commit 4327ba52afd03fc4b5afa0ee1d774c9c5b0e85c5 upstream.
If a EXT4 filesystem utilizes JBD2 journaling and an error occurs, the
journaling will be aborted first and the error number will be recorded
into JBD2 superblock and, finally, the system will
Hi,
I got a trouble while I am developing a patch for EXT4 filesytem.
Actually, I used a percpu rw semaphore in my patch and there was no
problem when I built EXT4 filesystem as built-in mode. However, when
kbuild auto build robot built EXT4 filesystem with my patch as a "module",
a bulid error oc
e
mnt_want_write_file() handles all the things we need.
It checks FMODE_WRITER, which is set in do_dentry_open() when
FMODE_WRITE is already set, and does the stuff that file_start_write()
is doing. This is why the other filesystem system calls use it.
What do you think?
2020년 6월 10일 (수) 오후 12:5
for write already.
in ext4/ioctl.c
case EXT4_IOC_SWAP_BOOT:
{
int err;
if (!(filp->f_mode & FMODE_WRITE))
return -EBADF;
err = mnt_want_write_file(filp);
if (err)
return err;2
목) 오전 9:00, Eric Biggers 님이 작성:
>
> On Thu, Jun 11, 2020 at 08:53:10AM +0900, Daeho Jeong wrote:
> > > > > > Using FMODE_WRITE is more proper for this case, since we're going to
> > > > > > modify the data. But I think mnt_want_write_file() is still req
Ok, I got it. Thanks for quick response~ :)
2020년 6월 11일 (목) 오전 10:56, Eric Biggers 님이 작성:
>
> On Thu, Jun 11, 2020 at 09:23:23AM +0900, Daeho Jeong wrote:
> > Yes, I saw the implementation in vfs_write().
> > But if we use mnt_want_write_file() here, it'll call mnt_clone
From: Daeho Jeong
Added a new ioctl to send discard commands or/and zero out
to whole data area of a regular file for security reason.
Signed-off-by: Daeho Jeong
---
fs/f2fs/f2fs.h | 8 +++
fs/f2fs/file.c | 143 +
2 files changed, 151
1 - 100 of 201 matches
Mail list logo
