linux-kernel - Re: [patch] epoll use a single inode ...

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: <200703071831.36930.dada1@cosmosbay.com>
Date:	Wed, 7 Mar 2007 18:31:36 +0100
From:	Eric Dumazet <dada1@...mosbay.com>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	Davide Libenzi <davidel@...ilserver.org>,
	Avi Kivity <avi@...o.co.il>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Al Viro <viro@...iv.linux.org.uk>
Subject: Re: [patch] epoll use a single inode ...

On Wednesday 07 March 2007 18:02, Linus Torvalds wrote:
> On Wed, 7 Mar 2007, Eric Dumazet wrote:
> > I would definitly *love* saving dentries for pipes (and sockets too), but
> > how are you going to get the inode ?
>
> Don't use an inode at all.

Lovely :)

>
> > pipes()/sockets() can use read()/write()/rw_verify_area() and thus need
> > file->f_path.dentry->d_inode (so each pipe needs a separate dentry)
>
> No, at least pipes could easily just use "file->f_private_data" instead.
>
> Now, sockets really do want the inode (or it would be really really big
> changes), but pipes really just want a "struct pipe_inode_info" pointer,
> which we could hide away directly in the file descriptor itself.

sockets already uses file->private_data.

But calls to read()/write() (not send()/recv()) still need to go through the 
dentry, before entering socket land.

>
> That's what Davide already did (on my suggestion) for signalfd - there's a
> *single* inode, and the real data is in the per-fd f_private_data.

Please find enclosed the following patch, to prepare this path.

[PATCH] Delay the dentry name generation on sockets and pipes.

1) Introduces a new method in 'struct dentry_operations'. This method called 
d_dname() might be called from d_path() to be able to provide a dentry name 
for special filesystems. It is called without locks.

Future patches (if we succeed in having one common dentry for all pipes) may 
need to change prototype of this method, but we now use :
char *d_dname(struct dentry *dentry, char *buffer, int buflen)

2) Use this new method for sockets : No more sprintf() at socket creation. 
This is delayed up to the moment someone does an access to /proc/pid/fd/...
We also avoid mntput()/mntget() on sock_mnt

3) Use this new method for pipes : No more sprintf() at pipe creation. This is 
delayed up to the moment someone does an access to /proc/pid/fd/...
We also avoid mntput()/mntget() on pipe_mnt

A benchmark consisting of 1.000.000 calls to pipe()/close()/close() gives a 
*nice* speedup on my Pentium(M) 1.6 Ghz :

3.090 s instead of 3.450 s

Signed-off-by: Eric Dumazet <dada1@...mosbay.com>

 fs/dcache.c            |    3 +++
 fs/pipe.c              |   16 +++++++++++-----
 include/linux/dcache.h |    1 +
 net/socket.c           |   15 +++++++++++----
 4 files changed, 26 insertions(+), 9 deletions(-)

View attachment "introduce_d_dname.patch" of type "text/x-diff" (3785 bytes)