linux-kernel - Re: [PATCH] Centralise definitions of sector_t and blkcnt

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: <Pine.LNX.4.64.0612041941220.3476@woody.osdl.org>
Date:	Mon, 4 Dec 2006 19:44:56 -0800 (PST)
From:	Linus Torvalds <torvalds@...l.org>
To:	Matthew Wilcox <matthew@....cx>
cc:	linux-arch@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] Centralise definitions of sector_t and blkcnt_t

On Mon, 4 Dec 2006, Matthew Wilcox wrote:
> 
> CONFIG_LBD and CONFIG_LSF are spread into asm/types.h for no particularly
> good reason.  Centralising the definition in linux/types.h means that arch
> maintainers don't need to bother adding it, as well as fixing the problem
> with x86-64 users being asked to make a decision that has absolutely no
> effect.  The H8/300 porters seem particularly confused since I'm not aware
> of any microcontrollers that need to support 2TB filesystems.

Applied, since this is a good cleanup regardless.

I'd still be open to switching things around further, and allow even 
64-bit architectures to say that they only want 32-bit sector_t's and page 
indexes (ie remove the "depends on !64BIT" and make the "unsigned long" 
case actually be "u32" instead, so that it literally switches between 
32-bit or 64-bit values _regardless_ or architecture).

I don't know how big a deal it is, but I could imagine that we could 
actually save memory in a smaller "struct page", for example, on 64bit 
architectures by just using a 4-byte index.

For now, the !64BIT makes sense simply because a 64-bit architecture 
probably doesn't care, and might as well just use 64 bits anyway (ie you 
tend to have tons of memory there anyway). And I suspect it doesn't 
actually even help on 64-bits due to structure alignment etc issues, but 
am too lazy to go check.

Just thought I'd mention the possibility, in other words.

		Linus
-
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/