2021-09-14 04:02:15

by Yue Hu

[permalink] [raw]
Subject: [PATCH] erofs: fix compacted_2b if compacted_4b_initial > totalidx

From: Yue Hu <[email protected]>

Currently, the whole indexes will only be compacted 4B if
compacted_4b_initial > totalidx. So, the calculated compacted_2b
is worthless for that case. It may waste CPU resources.

No need to update compacted_4b_initial as mkfs since it's used to
fulfill the alignment of the 1st compacted_2b pack and would handle
the case above.

We also need to clarify compacted_4b_end here. It's used for the
last lclusters which aren't fitted in the previous compacted_2b
packs.

Some messages are from Xiang.

Signed-off-by: Yue Hu <[email protected]>
---
fs/erofs/zmap.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index 9fb98d8..aeed404 100644
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -369,7 +369,8 @@ static int compacted_load_cluster_from_disk(struct z_erofs_maprecorder *m,
if (compacted_4b_initial == 32 / 4)
compacted_4b_initial = 0;

- if (vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B)
+ if ((vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B) &&
+ compacted_4b_initial <= totalidx) {
compacted_2b = rounddown(totalidx - compacted_4b_initial, 16);
else
compacted_2b = 0;
--
1.9.1


2021-09-14 04:39:00

by Gao Xiang

[permalink] [raw]
Subject: Re: [PATCH] erofs: fix compacted_2b if compacted_4b_initial > totalidx

On Tue, Sep 14, 2021 at 11:59:15AM +0800, Yue Hu wrote:
> From: Yue Hu <[email protected]>
>
> Currently, the whole indexes will only be compacted 4B if
> compacted_4b_initial > totalidx. So, the calculated compacted_2b
> is worthless for that case. It may waste CPU resources.
>
> No need to update compacted_4b_initial as mkfs since it's used to
> fulfill the alignment of the 1st compacted_2b pack and would handle
> the case above.
>
> We also need to clarify compacted_4b_end here. It's used for the
> last lclusters which aren't fitted in the previous compacted_2b
> packs.
>
> Some messages are from Xiang.
>
> Signed-off-by: Yue Hu <[email protected]>

Looks good to me,
Reviewed-by: Gao Xiang <[email protected]>

(although I think the subject title would be better changed into
"clear compacted_2b if compacted_4b_initial > totalidx"
since 'fix'-likewise words could trigger some AI bot for stable
kernel backporting..)

Thanks,
Gao Xiang

> ---
> fs/erofs/zmap.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
> index 9fb98d8..aeed404 100644
> --- a/fs/erofs/zmap.c
> +++ b/fs/erofs/zmap.c
> @@ -369,7 +369,8 @@ static int compacted_load_cluster_from_disk(struct z_erofs_maprecorder *m,
> if (compacted_4b_initial == 32 / 4)
> compacted_4b_initial = 0;
>
> - if (vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B)
> + if ((vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B) &&
> + compacted_4b_initial <= totalidx) {
> compacted_2b = rounddown(totalidx - compacted_4b_initial, 16);
> else
> compacted_2b = 0;
> --
> 1.9.1

2021-09-14 05:04:22

by Yue Hu

[permalink] [raw]
Subject: Re: [PATCH] erofs: fix compacted_2b if compacted_4b_initial > totalidx

On Tue, 14 Sep 2021 12:37:14 +0800
Gao Xiang <[email protected]> wrote:

> On Tue, Sep 14, 2021 at 11:59:15AM +0800, Yue Hu wrote:
> > From: Yue Hu <[email protected]>
> >
> > Currently, the whole indexes will only be compacted 4B if
> > compacted_4b_initial > totalidx. So, the calculated compacted_2b
> > is worthless for that case. It may waste CPU resources.
> >
> > No need to update compacted_4b_initial as mkfs since it's used to
> > fulfill the alignment of the 1st compacted_2b pack and would handle
> > the case above.
> >
> > We also need to clarify compacted_4b_end here. It's used for the
> > last lclusters which aren't fitted in the previous compacted_2b
> > packs.
> >
> > Some messages are from Xiang.
> >
> > Signed-off-by: Yue Hu <[email protected]>
>
> Looks good to me,
> Reviewed-by: Gao Xiang <[email protected]>
>
> (although I think the subject title would be better changed into
> "clear compacted_2b if compacted_4b_initial > totalidx"

Yeah, 'clear' is much better for this change.

Thanks.

> since 'fix'-likewise words could trigger some AI bot for stable
> kernel backporting..)
>
> Thanks,
> Gao Xiang
>
> > ---
> > fs/erofs/zmap.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
> > index 9fb98d8..aeed404 100644
> > --- a/fs/erofs/zmap.c
> > +++ b/fs/erofs/zmap.c
> > @@ -369,7 +369,8 @@ static int compacted_load_cluster_from_disk(struct z_erofs_maprecorder *m,
> > if (compacted_4b_initial == 32 / 4)
> > compacted_4b_initial = 0;
> >
> > - if (vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B)
> > + if ((vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B) &&
> > + compacted_4b_initial <= totalidx) {
> > compacted_2b = rounddown(totalidx - compacted_4b_initial, 16);
> > else
> > compacted_2b = 0;
> > --
> > 1.9.1

2021-09-14 05:13:29

by Gao Xiang

[permalink] [raw]
Subject: Re: [PATCH] erofs: fix compacted_2b if compacted_4b_initial > totalidx

On Tue, Sep 14, 2021 at 12:57:48PM +0800, Yue Hu wrote:
> On Tue, 14 Sep 2021 12:37:14 +0800
> Gao Xiang <[email protected]> wrote:
>
> > On Tue, Sep 14, 2021 at 11:59:15AM +0800, Yue Hu wrote:
> > > From: Yue Hu <[email protected]>
> > >
> > > Currently, the whole indexes will only be compacted 4B if
> > > compacted_4b_initial > totalidx. So, the calculated compacted_2b
> > > is worthless for that case. It may waste CPU resources.
> > >
> > > No need to update compacted_4b_initial as mkfs since it's used to
> > > fulfill the alignment of the 1st compacted_2b pack and would handle
> > > the case above.
> > >
> > > We also need to clarify compacted_4b_end here. It's used for the
> > > last lclusters which aren't fitted in the previous compacted_2b
> > > packs.
> > >
> > > Some messages are from Xiang.
> > >
> > > Signed-off-by: Yue Hu <[email protected]>
> >
> > Looks good to me,
> > Reviewed-by: Gao Xiang <[email protected]>
> >
> > (although I think the subject title would be better changed into
> > "clear compacted_2b if compacted_4b_initial > totalidx"
>
> Yeah, 'clear' is much better for this change.

Ok, I will update this when later applying.

Thanks,
Gao Xiang

>
> Thanks.
>
> > since 'fix'-likewise words could trigger some AI bot for stable
> > kernel backporting..)
> >
> > Thanks,
> > Gao Xiang
> >
> > > ---
> > > fs/erofs/zmap.c | 3 ++-
> > > 1 file changed, 2 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
> > > index 9fb98d8..aeed404 100644
> > > --- a/fs/erofs/zmap.c
> > > +++ b/fs/erofs/zmap.c
> > > @@ -369,7 +369,8 @@ static int compacted_load_cluster_from_disk(struct z_erofs_maprecorder *m,
> > > if (compacted_4b_initial == 32 / 4)
> > > compacted_4b_initial = 0;
> > >
> > > - if (vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B)
> > > + if ((vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B) &&
> > > + compacted_4b_initial <= totalidx) {
> > > compacted_2b = rounddown(totalidx - compacted_4b_initial, 16);
> > > else
> > > compacted_2b = 0;
> > > --
> > > 1.9.1

2021-09-23 11:51:37

by Gao Xiang

[permalink] [raw]
Subject: Re: [PATCH] erofs: fix compacted_2b if compacted_4b_initial > totalidx

On Tue, Sep 14, 2021 at 12:57:48PM +0800, Yue Hu wrote:
> On Tue, 14 Sep 2021 12:37:14 +0800
> Gao Xiang <[email protected]> wrote:
>
> > On Tue, Sep 14, 2021 at 11:59:15AM +0800, Yue Hu wrote:
> > > From: Yue Hu <[email protected]>
> > >
> > > Currently, the whole indexes will only be compacted 4B if
> > > compacted_4b_initial > totalidx. So, the calculated compacted_2b
> > > is worthless for that case. It may waste CPU resources.
> > >
> > > No need to update compacted_4b_initial as mkfs since it's used to
> > > fulfill the alignment of the 1st compacted_2b pack and would handle
> > > the case above.
> > >
> > > We also need to clarify compacted_4b_end here. It's used for the
> > > last lclusters which aren't fitted in the previous compacted_2b
> > > packs.
> > >
> > > Some messages are from Xiang.
> > >
> > > Signed-off-by: Yue Hu <[email protected]>
> >
> > Looks good to me,
> > Reviewed-by: Gao Xiang <[email protected]>
> >
> > (although I think the subject title would be better changed into
> > "clear compacted_2b if compacted_4b_initial > totalidx"
>
> Yeah, 'clear' is much better for this change.
>
> Thanks.
>
> > since 'fix'-likewise words could trigger some AI bot for stable
> > kernel backporting..)
> >
> > Thanks,
> > Gao Xiang
> >
> > > ---
> > > fs/erofs/zmap.c | 3 ++-
> > > 1 file changed, 2 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
> > > index 9fb98d8..aeed404 100644
> > > --- a/fs/erofs/zmap.c
> > > +++ b/fs/erofs/zmap.c
> > > @@ -369,7 +369,8 @@ static int compacted_load_cluster_from_disk(struct z_erofs_maprecorder *m,
> > > if (compacted_4b_initial == 32 / 4)
> > > compacted_4b_initial = 0;
> > >
> > > - if (vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B)
> > > + if ((vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B) &&
> > > + compacted_4b_initial <= totalidx) {

btw, I've fixed up the build error due to redundant brace '{' when
applying...

Thanks,
Gao Xiang

> > > compacted_2b = rounddown(totalidx - compacted_4b_initial, 16);
> > > else
> > > compacted_2b = 0;
> > > --
> > > 1.9.1

2021-09-23 12:52:10

by Yue Hu

[permalink] [raw]
Subject: Re: [PATCH] erofs: fix compacted_2b if compacted_4b_initial > totalidx

On Thu, 23 Sep 2021 19:49:42 +0800
Gao Xiang <[email protected]> wrote:

> On Tue, Sep 14, 2021 at 12:57:48PM +0800, Yue Hu wrote:
> > On Tue, 14 Sep 2021 12:37:14 +0800
> > Gao Xiang <[email protected]> wrote:
> >
> > > On Tue, Sep 14, 2021 at 11:59:15AM +0800, Yue Hu wrote:
> > > > From: Yue Hu <[email protected]>
> > > >
> > > > Currently, the whole indexes will only be compacted 4B if
> > > > compacted_4b_initial > totalidx. So, the calculated compacted_2b
> > > > is worthless for that case. It may waste CPU resources.
> > > >
> > > > No need to update compacted_4b_initial as mkfs since it's used to
> > > > fulfill the alignment of the 1st compacted_2b pack and would handle
> > > > the case above.
> > > >
> > > > We also need to clarify compacted_4b_end here. It's used for the
> > > > last lclusters which aren't fitted in the previous compacted_2b
> > > > packs.
> > > >
> > > > Some messages are from Xiang.
> > > >
> > > > Signed-off-by: Yue Hu <[email protected]>
> > >
> > > Looks good to me,
> > > Reviewed-by: Gao Xiang <[email protected]>
> > >
> > > (although I think the subject title would be better changed into
> > > "clear compacted_2b if compacted_4b_initial > totalidx"
> >
> > Yeah, 'clear' is much better for this change.
> >
> > Thanks.
> >
> > > since 'fix'-likewise words could trigger some AI bot for stable
> > > kernel backporting..)
> > >
> > > Thanks,
> > > Gao Xiang
> > >
> > > > ---
> > > > fs/erofs/zmap.c | 3 ++-
> > > > 1 file changed, 2 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
> > > > index 9fb98d8..aeed404 100644
> > > > --- a/fs/erofs/zmap.c
> > > > +++ b/fs/erofs/zmap.c
> > > > @@ -369,7 +369,8 @@ static int compacted_load_cluster_from_disk(struct z_erofs_maprecorder *m,
> > > > if (compacted_4b_initial == 32 / 4)
> > > > compacted_4b_initial = 0;
> > > >
> > > > - if (vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B)
> > > > + if ((vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B) &&
> > > > + compacted_4b_initial <= totalidx) {
>
> btw, I've fixed up the build error due to redundant brace '{' when

Thanks.

> applying...
>
> Thanks,
> Gao Xiang
>
> > > > compacted_2b = rounddown(totalidx - compacted_4b_initial, 16);
> > > > else
> > > > compacted_2b = 0;
> > > > --
> > > > 1.9.1


2021-09-23 15:15:55

by Chao Yu

[permalink] [raw]
Subject: Re: [PATCH] erofs: fix compacted_2b if compacted_4b_initial > totalidx



On 2021/9/23 19:49, Gao Xiang wrote:
> On Tue, Sep 14, 2021 at 12:57:48PM +0800, Yue Hu wrote:
>> On Tue, 14 Sep 2021 12:37:14 +0800
>> Gao Xiang <[email protected]> wrote:
>>
>>> On Tue, Sep 14, 2021 at 11:59:15AM +0800, Yue Hu wrote:
>>>> From: Yue Hu <[email protected]>
>>>>
>>>> Currently, the whole indexes will only be compacted 4B if
>>>> compacted_4b_initial > totalidx. So, the calculated compacted_2b
>>>> is worthless for that case. It may waste CPU resources.
>>>>
>>>> No need to update compacted_4b_initial as mkfs since it's used to
>>>> fulfill the alignment of the 1st compacted_2b pack and would handle
>>>> the case above.
>>>>
>>>> We also need to clarify compacted_4b_end here. It's used for the
>>>> last lclusters which aren't fitted in the previous compacted_2b
>>>> packs.
>>>>
>>>> Some messages are from Xiang.
>>>>
>>>> Signed-off-by: Yue Hu <[email protected]>
>>>
>>> Looks good to me,
>>> Reviewed-by: Gao Xiang <[email protected]>
>>>
>>> (although I think the subject title would be better changed into
>>> "clear compacted_2b if compacted_4b_initial > totalidx"
>>
>> Yeah, 'clear' is much better for this change.
>>
>> Thanks.
>>
>>> since 'fix'-likewise words could trigger some AI bot for stable
>>> kernel backporting..)
>>>
>>> Thanks,
>>> Gao Xiang
>>>
>>>> ---
>>>> fs/erofs/zmap.c | 3 ++-
>>>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
>>>> index 9fb98d8..aeed404 100644
>>>> --- a/fs/erofs/zmap.c
>>>> +++ b/fs/erofs/zmap.c
>>>> @@ -369,7 +369,8 @@ static int compacted_load_cluster_from_disk(struct z_erofs_maprecorder *m,
>>>> if (compacted_4b_initial == 32 / 4)
>>>> compacted_4b_initial = 0;
>>>>
>>>> - if (vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B)
>>>> + if ((vi->z_advise & Z_EROFS_ADVISE_COMPACTED_2B) &&
>>>> + compacted_4b_initial <= totalidx) {
>
> btw, I've fixed up the build error due to redundant brace '{' when
> applying...

It looks good to me after above fix.

Reviewed-by: Chao Yu <[email protected]>

Thanks,