LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* [PATCH] btrfs: send: in case of IO error log inode
@ 2022-02-02 20:14 Dāvis Mosāns
  2022-02-04 16:28 ` David Sterba
                   ` (2 more replies)
  0 siblings, 3 replies; 7+ messages in thread
From: Dāvis Mosāns @ 2022-02-02 20:14 UTC (permalink / raw)
  To: linux-btrfs
  Cc: Chris Mason, Josef Bacik, David Sterba, linux-kernel,
	Dāvis Mosāns

Currently if we get IO error while doing send then we abort without
logging information about which file caused issue.
So log inode to help with debugging.
---
 fs/btrfs/send.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index d8ccb62aa7d2..945d9c01d902 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -5000,6 +5000,7 @@ static int put_file_data(struct send_ctx *sctx, u64 offset, u32 len)
 			if (!PageUptodate(page)) {
 				unlock_page(page);
 				put_page(page);
+				btrfs_err(fs_info, "received IO error for inode=%llu", sctx->cur_ino);
 				ret = -EIO;
 				break;
 			}
-- 
2.35.1


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH] btrfs: send: in case of IO error log inode
  2022-02-02 20:14 [PATCH] btrfs: send: in case of IO error log inode Dāvis Mosāns
@ 2022-02-04 16:28 ` David Sterba
  2022-02-05  1:35   ` Dāvis Mosāns
  2022-02-05  1:36 ` [PATCH v2] btrfs: send: in case of IO error log it Dāvis Mosāns
  2022-02-05 18:48 ` [PATCH v3] " Dāvis Mosāns
  2 siblings, 1 reply; 7+ messages in thread
From: David Sterba @ 2022-02-04 16:28 UTC (permalink / raw)
  To: Dāvis Mosāns
  Cc: linux-btrfs, Chris Mason, Josef Bacik, David Sterba, linux-kernel

On Wed, Feb 02, 2022 at 10:14:37PM +0200, Dāvis Mosāns wrote:
> Currently if we get IO error while doing send then we abort without
> logging information about which file caused issue.
> So log inode to help with debugging.

The Signed-off-by tag is missing.

> ---
>  fs/btrfs/send.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
> index d8ccb62aa7d2..945d9c01d902 100644
> --- a/fs/btrfs/send.c
> +++ b/fs/btrfs/send.c
> @@ -5000,6 +5000,7 @@ static int put_file_data(struct send_ctx *sctx, u64 offset, u32 len)
>  			if (!PageUptodate(page)) {
>  				unlock_page(page);
>  				put_page(page);
> +				btrfs_err(fs_info, "received IO error for inode=%llu", sctx->cur_ino);

A message here makes sense. I'd make it more explicit that it's for send
(the word "received" is kind of confusing), the inode number is not
unique identifier so the root id should be also printed, it's available
from the sctx->send_root and maybe also the file offset (taken from the
associated page by page_offset()).

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH] btrfs: send: in case of IO error log inode
  2022-02-04 16:28 ` David Sterba
@ 2022-02-05  1:35   ` Dāvis Mosāns
  0 siblings, 0 replies; 7+ messages in thread
From: Dāvis Mosāns @ 2022-02-05  1:35 UTC (permalink / raw)
  To: David Sterba, Dāvis Mosāns, BTRFS, Chris Mason,
	Josef Bacik, David Sterba, linux-kernel

piektd., 2022. g. 4. febr., plkst. 18:29 — lietotājs David Sterba
(<dsterba@suse.cz>) rakstīja:
>
> On Wed, Feb 02, 2022 at 10:14:37PM +0200, Dāvis Mosāns wrote:
> > Currently if we get IO error while doing send then we abort without
> > logging information about which file caused issue.
> > So log inode to help with debugging.
>
> The Signed-off-by tag is missing.
>

Yeah I forgot.

> > ---
> >  fs/btrfs/send.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
> > index d8ccb62aa7d2..945d9c01d902 100644
> > --- a/fs/btrfs/send.c
> > +++ b/fs/btrfs/send.c
> > @@ -5000,6 +5000,7 @@ static int put_file_data(struct send_ctx *sctx, u64 offset, u32 len)
> >                       if (!PageUptodate(page)) {
> >                               unlock_page(page);
> >                               put_page(page);
> > +                             btrfs_err(fs_info, "received IO error for inode=%llu", sctx->cur_ino);
>
> A message here makes sense. I'd make it more explicit that it's for send
> (the word "received" is kind of confusing), the inode number is not
> unique identifier so the root id should be also printed, it's available
> from the sctx->send_root and maybe also the file offset (taken from the
> associated page by page_offset()).

Thanks! Will submit v2

^ permalink raw reply	[flat|nested] 7+ messages in thread

* [PATCH v2] btrfs: send: in case of IO error log it
  2022-02-02 20:14 [PATCH] btrfs: send: in case of IO error log inode Dāvis Mosāns
  2022-02-04 16:28 ` David Sterba
@ 2022-02-05  1:36 ` Dāvis Mosāns
  2022-02-05 18:50   ` Dāvis Mosāns
  2022-02-05 18:48 ` [PATCH v3] " Dāvis Mosāns
  2 siblings, 1 reply; 7+ messages in thread
From: Dāvis Mosāns @ 2022-02-05  1:36 UTC (permalink / raw)
  To: linux-btrfs
  Cc: Chris Mason, Josef Bacik, David Sterba, linux-kernel,
	Dāvis Mosāns

Currently if we get IO error while doing send then we abort without
logging information about which file caused issue.
So log it to help with debugging.

Signed-off-by: Dāvis Mosāns <davispuh@gmail.com>
---
 fs/btrfs/send.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index d8ccb62aa7d2..a1fd449a5ecc 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -5000,6 +5000,8 @@ static int put_file_data(struct send_ctx *sctx, u64 offset, u32 len)
 			if (!PageUptodate(page)) {
 				unlock_page(page);
 				put_page(page);
+				btrfs_err(fs_info, "send: IO error at offset=%llu for inode=%llu root=%llu",
+					page_offset(page), sctx->cur_ino, sctx->send_root->root_key.objectid);
 				ret = -EIO;
 				break;
 			}
-- 
2.35.1


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [PATCH v3] btrfs: send: in case of IO error log it
  2022-02-02 20:14 [PATCH] btrfs: send: in case of IO error log inode Dāvis Mosāns
  2022-02-04 16:28 ` David Sterba
  2022-02-05  1:36 ` [PATCH v2] btrfs: send: in case of IO error log it Dāvis Mosāns
@ 2022-02-05 18:48 ` Dāvis Mosāns
  2022-02-08 23:04   ` David Sterba
  2 siblings, 1 reply; 7+ messages in thread
From: Dāvis Mosāns @ 2022-02-05 18:48 UTC (permalink / raw)
  To: linux-btrfs
  Cc: Chris Mason, Josef Bacik, David Sterba, linux-kernel,
	Dāvis Mosāns

Currently if we get IO error while doing send then we abort without
logging information about which file caused issue.
So log it to help with debugging.

Signed-off-by: Dāvis Mosāns <davispuh@gmail.com>
---
 fs/btrfs/send.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index d8ccb62aa7d2..b1f75fde4a19 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -4999,6 +4999,8 @@ static int put_file_data(struct send_ctx *sctx, u64 offset, u32 len)
 			lock_page(page);
 			if (!PageUptodate(page)) {
 				unlock_page(page);
+				btrfs_err(fs_info, "send: IO error at offset=%llu for inode=%llu root=%llu",
+					page_offset(page), sctx->cur_ino, sctx->send_root->root_key.objectid);
 				put_page(page);
 				ret = -EIO;
 				break;
-- 
2.35.1


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2] btrfs: send: in case of IO error log it
  2022-02-05  1:36 ` [PATCH v2] btrfs: send: in case of IO error log it Dāvis Mosāns
@ 2022-02-05 18:50   ` Dāvis Mosāns
  0 siblings, 0 replies; 7+ messages in thread
From: Dāvis Mosāns @ 2022-02-05 18:50 UTC (permalink / raw)
  To: BTRFS; +Cc: Chris Mason, Josef Bacik, David Sterba, linux-kernel

> diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
> index d8ccb62aa7d2..a1fd449a5ecc 100644
> --- a/fs/btrfs/send.c
> +++ b/fs/btrfs/send.c
> @@ -5000,6 +5000,8 @@ static int put_file_data(struct send_ctx *sctx, u64 offset, u32 len)
>                         if (!PageUptodate(page)) {
>                                 unlock_page(page);
>                                 put_page(page);
> +                               btrfs_err(fs_info, "send: IO error at offset=%llu for inode=%llu root=%llu",
> +                                       page_offset(page), sctx->cur_ino, sctx->send_root->root_key.objectid);

Thought more about this and I'm guessing using page after put_page is
probably wrong but it did work fine. Submitted v3.

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v3] btrfs: send: in case of IO error log it
  2022-02-05 18:48 ` [PATCH v3] " Dāvis Mosāns
@ 2022-02-08 23:04   ` David Sterba
  0 siblings, 0 replies; 7+ messages in thread
From: David Sterba @ 2022-02-08 23:04 UTC (permalink / raw)
  To: Dāvis Mosāns
  Cc: linux-btrfs, Chris Mason, Josef Bacik, David Sterba, linux-kernel

On Sat, Feb 05, 2022 at 08:48:23PM +0200, Dāvis Mosāns wrote:
> Currently if we get IO error while doing send then we abort without
> logging information about which file caused issue.
> So log it to help with debugging.
> 
> Signed-off-by: Dāvis Mosāns <davispuh@gmail.com>

Added to misc-next, thanks.

> ---
>  fs/btrfs/send.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
> index d8ccb62aa7d2..b1f75fde4a19 100644
> --- a/fs/btrfs/send.c
> +++ b/fs/btrfs/send.c
> @@ -4999,6 +4999,8 @@ static int put_file_data(struct send_ctx *sctx, u64 offset, u32 len)
>  			lock_page(page);
>  			if (!PageUptodate(page)) {
>  				unlock_page(page);
> +				btrfs_err(fs_info, "send: IO error at offset=%llu for inode=%llu root=%llu",
> +					page_offset(page), sctx->cur_ino, sctx->send_root->root_key.objectid);
>  				put_page(page);

Good point in v2, using page must be before put_page. I've slightly
reformatted the message so the lines fit to the limit.

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2022-02-08 23:08 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-02-02 20:14 [PATCH] btrfs: send: in case of IO error log inode Dāvis Mosāns
2022-02-04 16:28 ` David Sterba
2022-02-05  1:35   ` Dāvis Mosāns
2022-02-05  1:36 ` [PATCH v2] btrfs: send: in case of IO error log it Dāvis Mosāns
2022-02-05 18:50   ` Dāvis Mosāns
2022-02-05 18:48 ` [PATCH v3] " Dāvis Mosāns
2022-02-08 23:04   ` David Sterba

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).