netdev - Re: [PATCH 1/2] pcnet32: only allocate init

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: <20070308014417.GA11717@verizon.net>
Date:	Wed, 7 Mar 2007 17:44:17 -0800
From:	Don Fry <pcnet32@...izon.net>
To:	"Michael K. Edwards" <medwards.linux@...il.com>
Cc:	Ralf Baechle <ralf@...ux-mips.org>, jgarzik@...ox.com,
	netdev@...r.kernel.org
Subject: Re: [PATCH 1/2] pcnet32: only allocate init_block dma consistent

On Wed, Mar 07, 2007 at 12:15:26PM -0800, Michael K. Edwards wrote:
> cacheability of the allocated block.  Maybe it doesn't.  The alpha
> version of pci_alloc_consistent, on the other hand, does interesting
> things to make the memory visible to PCI.  Don, what arches did you
> have in mind when you commented on caching issues?
> 
I had not made any comments on the caching issue until now.
The comments on the caching and performance improvements came from
Thomas Bogendoerfer who submitted this change, and Ralf Baechle.

On some architectures, at least for MIPS, the locks have to be in cached
memory.  The only part of the data structure that has to be in pci
coherent memory is the init block data structure, which the chip reads.
All the rest of the information in the private structure are for the
driver.

Having those variables in cacheable memory "should" improve performance,
I would think.

Don
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html