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  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]
Date:   Thu, 25 Feb 2021 21:42:54 -0700
From:   Andreas Dilger <adilger@...ger.ca>
To:     harshad shirwadkar <harshadshirwadkar@...il.com>
Cc:     Благодаренко Артём 
        <artem.blagodarenko@...il.com>,
        Ext4 Developers List <linux-ext4@...r.kernel.org>,
        "Theodore Y. Ts'o" <tytso@....edu>,
        Alex Zhuravlev <bzzz@...mcloud.com>,
        Shuichi Ihara <sihara@....com>
Subject: Re: [PATCH v2 4/5] ext4: improve cr 0 / cr 1 group scanning

On Feb 25, 2021, at 9:06 PM, harshad shirwadkar <harshadshirwadkar@...il.com> wrote:
> 
> Hi Andreas,
> 
> On Thu, Feb 25, 2021 at 7:43 PM Andreas Dilger <adilger@...ger.ca> wrote:
>> 
>> Yes, modulo using the existing "mb_min_to_scan" parameter for this.
>> I think 4 or 8 or 10 groups is reasonable (512MB, 1GB, 1.25GB),
>> since if it needs a seek anyway then we may as well find a good
>> group for this.
> If I understand it right, the meaning of mb_min_to_scan is the number
> of *extents* that the allocator should try to find before choosing the
> best one. However, what we want here is the number of *groups* that
> the allocator should travel linearly before trying to optimize the
> search. So, even if mb_min_to_scan is set to 1, by the current
> definition of it, it means that the allocator may still traverse the
> entire file system if it doesn't find a match. Is my understanding
> right?

Sorry, you are right.  It is the number of extents to scan and not
the number of groups.

Cheers, Andreas






Download attachment "signature.asc" of type "application/pgp-signature" (874 bytes)

Powered by blists - more mailing lists