linux-kernel - Re: New filesystem for Linux

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: <20061105083416.GA2246@1wt.eu>
Date:	Sun, 5 Nov 2006 09:34:16 +0100
From:	Willy Tarreau <w@....eu>
To:	Mikulas Patocka <mikulas@...ax.karlin.mff.cuni.cz>
Cc:	Linus Torvalds <torvalds@...l.org>, linux-kernel@...r.kernel.org
Subject: Re: New filesystem for Linux

On Sun, Nov 05, 2006 at 05:14:06AM +0100, Mikulas Patocka wrote:
> On Sat, 4 Nov 2006, Linus Torvalds wrote:
> >- you have a _very_ confusing usage of upper-case. Not only are a lot of
> >  functions upper-case, some filenames are also upper-case. What would
> >  otherwise be more readable just ends up being hard to read because it's
> >  so odd and unexpected.
> >
> >  I'm sure there is some logic to it, but it escapes me.
> 
> I'm used to this. I usually make important functions with uppercase 
> letters and nonimportant temporary functions with lowercase letters.
>
> But I see that it contradicts general kernel coding style, so I can change 
> it.

We're more used to have uppercase for macros (or enums) and lowercase for
the rest. That way, when you read NULL, KERN_WARNING, PAGE_CACHE_SIZE,
INIT_LIST_HEAD..., you know that you're dealing with a macro, which is
very convenient.

> BTW do you find uppercase typedefs like
> typedef struct {
> 	...
> } SPADFNODE;
> confusing too?

Yes for the reason above. Also, we don't much use type definitions for
structures, because it's easier to understand "struct spadfnode *node"
in a function declaration than "SPADFNODE *node". Take a look at other
file systems. You'll see lots of "struct inode", "struct buffer_head".
Sometimes, you'll find some types suffixed with "_t", such as "handle_t"
or "spinlock_t". Generally, they are used for very commonly used data
(such as spinlocks), or opaque data which are passed between functions
without any interpretation. But it's more from observation than a rule.

> Uppercase filenames are there because the files are taken from another 
> (not yet released) project. But the kernel driver does not share any code 
> except definitions of disk structures, I saw how badly an attempt to share 
> kernel code affected XFS.

It's better to avoid uppercases in file names. I recently had to help
a user who could not build his kernel because of strange errors. I
finally found out that he was building from "entry.s" instead of "entry.S",
which suggested he copied the tree on a FAT. He confirmed having used a
vfat-formatted USB stick to copy his tree. Such errors are very annoying
to debug.

(...)
> I placed some benchmark on 
> http://artax.karlin.mff.cuni.cz/~mikulas/spadfs/benchmarks/

Not bad at all it seems !

Regards,
Willy

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/