From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <7cf9cd7e97048602033c63bed8973292@terzarima.net> To: 9fans@cse.psu.edu Subject: Re: [9fans] terminal types and photos? From: Charles Forsyth In-Reply-To: <116f90107ed4a05b196780472bf971ef@collyer.net> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Date: Mon, 15 Mar 2004 11:52:17 +0000 Topicbox-Message-UUID: 3085e912-eacd-11e9-9e20-41e7f4b1d025 >>True enough. What is an Xscale anyway? A faster ARM? it's a family of Intel ARM5-based processors with a (fairly) standard core and a range of application-specific modules surrounding it. in its PXA25x form, it's similar to the Strongarm but many of the bugs are different, some of the peripherals are different, and most of the registers are in different places even when the bit values are the same. at least they redesigned the DMA. that A bit B bit dance was just confused. the PXA250 didn't quite give the power or speed improvements over SA1110 that were expected, though it's all right. the PXA255 supposedly manages it. if i sound a bit grumpy it's because some of the PXA bugs are fairly fundamental. for instance, if fancy clock switching and sleep modes are big selling points for the device, and supposedly enabled by setting a few bits, you'd expect them to work, wouldn't you? especially since they were embarrassingly wrong in the StrongARM. silly you. if anything, the situation is worse. the errors are decently documented in the errata, but the fix is as fiddly as ever, and bigger. the cache might corrupt things. still, it's not too hard to get it to work. in general, however, i've found that IBM does a better job, with the smaller powerpcs. i don't know why. generally things work as documented, the documentation is well-organised, and the bugs in the errata tend to be understandably obscure. fairly tidy architecturally as well. my big IBM powerpc processor hasn't arrived yet so i've yet to see about that.