[LSF/MM ATTEND] block: multipage bvec upstreaming & next step

[LSF/MM ATTEND] block: multipage bvec upstreaming & next step

[Ming Lei]

Hi Guys,

The multipage bvec patchset[1] has been posted out for a while, and
xfstest(ext4, xfs and btrfs) have been run and no regression is
observed, and I hope having talk about this in person would help
moving it forward.

Secondly I'd like to discuss with guys about the following next steps
of mp-bvec:

1) cleanup raid1/raid5/raid10 for removing the only singlepage bvec path
2) optimizing block layer with sort of 'complete' multipage bvec implementation
3) use multpage bvec direclty in some drivers
4) pass multipage segment from iov iterator helpers / get_user_pages()

Also I am intereted in Kirill's topic about THP for filesystem too.


[1] https://lkml.org/lkml/2017/1/15/266


Thanks,
Ming Lei

Re: [LSF/MM ATTEND] block: multipage bvec upstreaming & next step

[Coly Li]

On 2017/1/17 上午11:56, Ming Lei wrote:
> Hi Guys,
> 
> The multipage bvec patchset[1] has been posted out for a while, and
> xfstest(ext4, xfs and btrfs) have been run and no regression is
> observed, and I hope having talk about this in person would help
> moving it forward.
> 
> Secondly I'd like to discuss with guys about the following next steps
> of mp-bvec:
> 
> 1) cleanup raid1/raid5/raid10 for removing the only singlepage bvec path

Hi Ming,

I am interested on this topic. Could you please provide a hint that what
is the proper way to remove single page bvec from md raid code ? And
will dm code involved in as well ?

Thanks.

Coly

Re: [LSF/MM ATTEND] block: multipage bvec upstreaming & next step

[Ming Lei]

Hi Coly,

On Thu, Jan 19, 2017 at 3:57 PM, Coly Li <colyli@suse.de> wrote:
> On 2017/1/17 上午11:56, Ming Lei wrote:
>> Hi Guys,
>>
>> The multipage bvec patchset[1] has been posted out for a while, and
>> xfstest(ext4, xfs and btrfs) have been run and no regression is
>> observed, and I hope having talk about this in person would help
>> moving it forward.
>>
>> Secondly I'd like to discuss with guys about the following next steps
>> of mp-bvec:
>>
>> 1) cleanup raid1/raid5/raid10 for removing the only singlepage bvec path
>
> Hi Ming,
>
> I am interested on this topic. Could you please provide a hint that what
> is the proper way to remove single page bvec from md raid code ?

Basically speaking the raid code can't access the bvec table directly like
referencing .bi_vcnt or .bi_io_vec once multipage bvec is supported,
especially after pages are added to the bio via bio_add_page().

I am looking at the raid1 code, and maybe we can discuss a bit
about the cleanup.

> And
> will dm code involved in as well ?

No, nothing at all, and no such usage in dm code.


Thanks,
Ming Lei