linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Peter Staubach <staubach@redhat.com>
To: Al Viro <viro@ftp.linux.org.uk>
Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH 1/2] handling 64bit values for st_ino]
Date: Thu, 10 Nov 2005 08:52:15 -0500	[thread overview]
Message-ID: <4373508F.9060004@redhat.com> (raw)
In-Reply-To: <20051110134336.GE7992@ftp.linux.org.uk>

Al Viro wrote:

>On Thu, Nov 10, 2005 at 07:57:21AM -0500, Peter Staubach wrote:
>  
>
>>Has this potential degradation been measured?  This is a lot of extra
>>complexity which needs to justified by the resulting performance.
>>    
>>
>
>What extra complexity?
>
>  
>

Two different sized types to describe inode numbers, different paths, etc.
Having two of something, when just one would suffice, is usually more
complicated.

The simplest implementation would be to have one, 64 bit ino_t, and then
only places which need to know about 32 bit ino_t would know about them.
These places would be restricted to system call interfaces with defined
32 bit ino_t's and file systems which only support 32 bit ino_t's.

>>>	Fix is pretty cheap and consists of two parts:
>>>1) widen struct kstat ->ino to u64, add a macro (check_inumber()) to
>>>be used in callers of ->getattr() that want to store ->ino in possibly
>>>narrower fields and care about overflows (stuff like sys_old_stat() with
>>>its 16bit st_ino clearly doesn't ;-)
>>>      
>>>
>
>  
>
>>It seems to me that a type with a name which better matches the intended
>>semantics would be a better choice than u64.  Even something like ino64_t
>>would help file systems maintainers to correctly implement the appropriate
>>support.
>>    
>>
>
>Why the hell would fs maintainers needs to touch their code at all?
>Have you actually read that patches?
>

Yes, Al, I read the patches.  I didn't say anything about current file
systems needing make changes to implement this support.  You've already
proposed the changes to do so.  I am more concerned with new file systems
coming down the road and file system maintenance starting the second after
your changes are integrated.  You aren't going to be doing that support,
so we will be depending upon people who aren't as aware of the intended
semantics to do the support.  Maintenance of a file system implementation
is already complex enough, so anything that can be done to simplify the
job is a good thing.

       ps

  reply	other threads:[~2005-11-10 13:52 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-11-10  0:30 [PATCH 1/2] handling 64bit values for st_ino] Al Viro
2005-11-10 12:57 ` Peter Staubach
2005-11-10 13:43   ` Al Viro
2005-11-10 13:52     ` Peter Staubach [this message]
2005-11-10 14:04       ` Al Viro
  -- strict thread matches above, loose matches on Subject: below --
2005-11-10 16:50 Steve French

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4373508F.9060004@redhat.com \
    --to=staubach@redhat.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=viro@ftp.linux.org.uk \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).