linux-ext4 - RE: RE: Re: [PATCH] ext4: change sequential discard handling on commit complete phase into parallel manner

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20170531022240epcms1p27bf64e351cc61e84378ea8de38e192ac@epcms1p2>
Date:   Wed, 31 May 2017 02:22:40 +0000
From:   Daeho Jeong <daeho.jeong@...sung.com>
To:     Daeho Jeong <daeho.jeong@...sung.com>, Jan Kara <jack@...e.cz>
CC:     "jack@...e.com" <jack@...e.com>,
        "hch@...radead.org" <hch@...radead.org>,
        "tytso@....edu" <tytso@....edu>,
        "linux-ext4@...r.kernel.org" <linux-ext4@...r.kernel.org>
Subject: RE: RE: Re: [PATCH] ext4: change sequential discard handling on
 commit complete phase into parallel manner

Hi Jan,

> Hi Jan,
 
> > Hum, these games with several callbacks, lists, etc. look awkward and
> > unnecessary. It think they mostly come from the fact that we call separate
> > freeing callback for each extent to free which doesn't fit the needs of
> > async discard well.
 
> > So instead of adding post_cb_list and several callback functions, it would
> > seem easier to have just one callback structure instead of one for every
> > extent. Then the structure would contain a list of extents that need to be
> > freed freed. So something like:
 
> > struct ext4_free_data {
> >         struct ext4_journal_cb_entry efd_jce;
> >         struct list_head efd_extents;
> > }
 
> > struct ext4_freed_extent {
> >         struct list_head efe_list;
> >         struct rb_node efe_node;
> >         ext4_group_t efe_group;
> >         ext4_grpblk_t efe_start_cluster;
> >         ext4_grpblk_t efe_count;
> >         tid_t efe_tid;
> > }
 
> > When commit happens, we can just walk the efd_extents list while efe_tid is
> > equal tid of the transaction for which the callback was called and submit all
> > discard requests. You can use bio chaining implemented in
> > __blkdev_issue_discard() which XFS already uses and so the result of all
> > the discards you submit will be just one bio. Then you walk the list of
> > extents again and free them in the buddy bitmaps. And finally, you wait for
> > the bio to complete. All will be then happening in one function and it will
> > be much easier to understand.
 
> It's right. the patch didn't look neat because of a few callbacks and the
> post callback list. I will modify the patch as your suggestion. It will
> look better.
 
> Thank you very much. :-)

It's a little difficult to decide when we have to add new ext4_free_data entry for
a transaction for the first time and how do we know whether the ext4_free_data entry
for a transaction is already added or not? I think that it is a bad idea to search in
t_private_list of the transaction for that, because there might be the different
type of callback entries in the future.

And how do we find the exact ext4_free_data entry for a newly created ext4_freed_extent?
We only know which transcation is related to the ext4_freed_extent, so we could use this
but I don't have any good idea for that. 

Do you have any idea?

Thank you.