[PATCH] btrfs: Use raw numbers when setting passing uptodate parameter to btrfs_writepage_endio_finish_ordered

[PATCH] btrfs: Use raw numbers when setting passing uptodate parameter to btrfs_writepage_endio_finish_ordered

[Nikolay Borisov]

The uptodate parameter of btrfs_writepage_endio_finish_ordered is used
to signal whether an error has occured while writing the given page.
0 signal an error, which is propagated to callees and 1 signifies
success. In end_compressed_bio_write the ->bi_status is checked and
based on it either BLK_STS_OK (0) or BLK_STS_NOTSUPP (1) are used. While
from functional point of view this is ok it's a for the poor reader of
the code, since the block layer values are conflated with the semantics
of the parameter.

Just use plain 0 or 1. No functional changes.

Signed-off-by: Nikolay Borisov <nborisov@suse.com>
---
 fs/btrfs/compression.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
index eb8e20b740d6..f82441f18512 100644
--- a/fs/btrfs/compression.c
+++ b/fs/btrfs/compression.c
@@ -250,7 +250,7 @@ static void end_compressed_bio_write(struct bio *bio)
 	cb->compressed_pages[0]->mapping = cb->inode->i_mapping;
 	btrfs_writepage_endio_finish_ordered(cb->compressed_pages[0],
 			cb->start, cb->start + cb->len - 1,
-			bio->bi_status ? BLK_STS_OK : BLK_STS_NOTSUPP);
+			bio->bi_status ? 0 : 1);
 	cb->compressed_pages[0]->mapping = NULL;
 
 	end_compressed_writeback(inode, cb);
-- 
2.17.1

Re: [PATCH] btrfs: Use raw numbers when setting passing uptodate parameter to btrfs_writepage_endio_finish_ordered

[David Sterba]

On Wed, Mar 20, 2019 at 09:53:16PM +0200, Nikolay Borisov wrote:
> The uptodate parameter of btrfs_writepage_endio_finish_ordered is used
> to signal whether an error has occured while writing the given page.
> 0 signal an error, which is propagated to callees and 1 signifies
> success. In end_compressed_bio_write the ->bi_status is checked and
> based on it either BLK_STS_OK (0) or BLK_STS_NOTSUPP (1) are used. While
> from functional point of view this is ok it's a for the poor reader of
> the code, since the block layer values are conflated with the semantics
> of the parameter.
> 
> Just use plain 0 or 1. No functional changes.
> 
> Signed-off-by: Nikolay Borisov <nborisov@suse.com>
> ---
>  fs/btrfs/compression.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
> index eb8e20b740d6..f82441f18512 100644
> --- a/fs/btrfs/compression.c
> +++ b/fs/btrfs/compression.c
> @@ -250,7 +250,7 @@ static void end_compressed_bio_write(struct bio *bio)
>  	cb->compressed_pages[0]->mapping = cb->inode->i_mapping;
>  	btrfs_writepage_endio_finish_ordered(cb->compressed_pages[0],
>  			cb->start, cb->start + cb->len - 1,
> -			bio->bi_status ? BLK_STS_OK : BLK_STS_NOTSUPP);
> +			bio->bi_status ? 0 : 1);

Essentially reverting 2dbe0c77186c691386b74391039899808a4d3f59, but
using the BLK_STS constants is definetelly wrong. The values match the
intended use of 'uptodate', but otherwise are confusing.

I actually think it can be simplified to bio->bi_staus == BLK_STS_OK.

Re: [PATCH] btrfs: Use raw numbers when setting passing uptodate parameter to btrfs_writepage_endio_finish_ordered

[Nikolay Borisov]

On 21.03.19 г. 16:30 ч., David Sterba wrote:
> On Wed, Mar 20, 2019 at 09:53:16PM +0200, Nikolay Borisov wrote:
>> The uptodate parameter of btrfs_writepage_endio_finish_ordered is used
>> to signal whether an error has occured while writing the given page.
>> 0 signal an error, which is propagated to callees and 1 signifies
>> success. In end_compressed_bio_write the ->bi_status is checked and
>> based on it either BLK_STS_OK (0) or BLK_STS_NOTSUPP (1) are used. While
>> from functional point of view this is ok it's a for the poor reader of
>> the code, since the block layer values are conflated with the semantics
>> of the parameter.
>>
>> Just use plain 0 or 1. No functional changes.
>>
>> Signed-off-by: Nikolay Borisov <nborisov@suse.com>
>> ---
>>  fs/btrfs/compression.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
>> index eb8e20b740d6..f82441f18512 100644
>> --- a/fs/btrfs/compression.c
>> +++ b/fs/btrfs/compression.c
>> @@ -250,7 +250,7 @@ static void end_compressed_bio_write(struct bio *bio)
>>  	cb->compressed_pages[0]->mapping = cb->inode->i_mapping;
>>  	btrfs_writepage_endio_finish_ordered(cb->compressed_pages[0],
>>  			cb->start, cb->start + cb->len - 1,
>> -			bio->bi_status ? BLK_STS_OK : BLK_STS_NOTSUPP);
>> +			bio->bi_status ? 0 : 1);
> 
> Essentially reverting 2dbe0c77186c691386b74391039899808a4d3f59, but
> using the BLK_STS constants is definetelly wrong. The values match the
> intended use of 'uptodate', but otherwise are confusing.
> 
> I actually think it can be simplified to bio->bi_staus == BLK_STS_OK.

Yes that would work, are you going to fix up the patch or shall I resend?

> 

Re: [PATCH] btrfs: Use raw numbers when setting passing uptodate parameter to btrfs_writepage_endio_finish_ordered

[David Sterba]

On Thu, Mar 21, 2019 at 04:31:53PM +0200, Nikolay Borisov wrote:
> 
> 
> On 21.03.19 г. 16:30 ч., David Sterba wrote:
> > On Wed, Mar 20, 2019 at 09:53:16PM +0200, Nikolay Borisov wrote:
> >> The uptodate parameter of btrfs_writepage_endio_finish_ordered is used
> >> to signal whether an error has occured while writing the given page.
> >> 0 signal an error, which is propagated to callees and 1 signifies
> >> success. In end_compressed_bio_write the ->bi_status is checked and
> >> based on it either BLK_STS_OK (0) or BLK_STS_NOTSUPP (1) are used. While
> >> from functional point of view this is ok it's a for the poor reader of
> >> the code, since the block layer values are conflated with the semantics
> >> of the parameter.
> >>
> >> Just use plain 0 or 1. No functional changes.
> >>
> >> Signed-off-by: Nikolay Borisov <nborisov@suse.com>
> >> ---
> >>  fs/btrfs/compression.c | 2 +-
> >>  1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
> >> index eb8e20b740d6..f82441f18512 100644
> >> --- a/fs/btrfs/compression.c
> >> +++ b/fs/btrfs/compression.c
> >> @@ -250,7 +250,7 @@ static void end_compressed_bio_write(struct bio *bio)
> >>  	cb->compressed_pages[0]->mapping = cb->inode->i_mapping;
> >>  	btrfs_writepage_endio_finish_ordered(cb->compressed_pages[0],
> >>  			cb->start, cb->start + cb->len - 1,
> >> -			bio->bi_status ? BLK_STS_OK : BLK_STS_NOTSUPP);
> >> +			bio->bi_status ? 0 : 1);
> > 
> > Essentially reverting 2dbe0c77186c691386b74391039899808a4d3f59, but
> > using the BLK_STS constants is definetelly wrong. The values match the
> > intended use of 'uptodate', but otherwise are confusing.
> > 
> > I actually think it can be simplified to bio->bi_staus == BLK_STS_OK.
> 
> Yes that would work, are you going to fix up the patch or shall I resend?

I'll fix it.

Re: [PATCH] btrfs: Use raw numbers when setting passing uptodate parameter to btrfs_writepage_endio_finish_ordered

[Anand Jain]

On 3/21/19 11:25 PM, David Sterba wrote:
> On Thu, Mar 21, 2019 at 04:31:53PM +0200, Nikolay Borisov wrote:
>>
>>
>> On 21.03.19 г. 16:30 ч., David Sterba wrote:
>>> On Wed, Mar 20, 2019 at 09:53:16PM +0200, Nikolay Borisov wrote:
>>>> The uptodate parameter of btrfs_writepage_endio_finish_ordered is used
>>>> to signal whether an error has occured while writing the given page.
>>>> 0 signal an error, which is propagated to callees and 1 signifies
>>>> success. In end_compressed_bio_write the ->bi_status is checked and
>>>> based on it either BLK_STS_OK (0) or BLK_STS_NOTSUPP (1) are used. While
>>>> from functional point of view this is ok it's a for the poor reader of
>>>> the code, since the block layer values are conflated with the semantics
>>>> of the parameter.
>>>>
>>>> Just use plain 0 or 1. No functional changes.
>>>>
>>>> Signed-off-by: Nikolay Borisov <nborisov@suse.com>
>>>> ---
>>>>   fs/btrfs/compression.c | 2 +-
>>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>>
>>>> diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
>>>> index eb8e20b740d6..f82441f18512 100644
>>>> --- a/fs/btrfs/compression.c
>>>> +++ b/fs/btrfs/compression.c
>>>> @@ -250,7 +250,7 @@ static void end_compressed_bio_write(struct bio *bio)
>>>>   	cb->compressed_pages[0]->mapping = cb->inode->i_mapping;
>>>>   	btrfs_writepage_endio_finish_ordered(cb->compressed_pages[0],
>>>>   			cb->start, cb->start + cb->len - 1,
>>>> -			bio->bi_status ? BLK_STS_OK : BLK_STS_NOTSUPP);
>>>> +			bio->bi_status ? 0 : 1);
>>>
>>> Essentially reverting 2dbe0c77186c691386b74391039899808a4d3f59, but
>>> using the BLK_STS constants is definetelly wrong. The values match the
>>> intended use of 'uptodate', but otherwise are confusing.
>>>
>>> I actually think it can be simplified to bio->bi_staus == BLK_STS_OK.
>>
>> Yes that would work, are you going to fix up the patch or shall I resend?
> 
> I'll fix it.

bio->bi_staus == BLK_STS_OK ? 1 : 0);

Is better. Its confusing. Thanks for the fix.

Reviewed-by: Anand Jain <anand.jain@oracle.com>