H, I also like this approach, lower 16 bits are used for original system revision. And use upper 16 bit for CPU detection. I don't want to create global variable and use existing one. One question, do you really used 2^16 revisions? I think 4 bits is enough. It means we can use 12 bits for other purpose. Thank you, Kyungmin Park > -----Original Message----- > From: Andy Green [mailto:andy@xxxxxxxxxxx] > Sent: Tuesday, March 10, 2009 4:58 AM > To: Mark Brown > Cc: Dave Jones; Ben Dooks; linux-arm-kernel@xxxxxxxxxxxxxxxxxxxxxx; > cpufreq@xxxxxxxxxxxxxxx; Kyungmin Park > Subject: Re: [PATCH 1/4] S3C: CPU detection support > > Somebody in the thread at some point said: > > On Mon, Mar 09, 2009 at 07:13:54PM +0000, Andy Green wrote: > > > >> We set system_rev from an ATAG in our bootloader.... this trashes that > >> for us. > > > >> Is what we're doing there wrong or is overwriting system_rev here wrong? > > > > That's an issue that was raised with the patch previously - see the > > thread: > > > > http://www.spinics.net/lists/arm-kernel/msg62218.html > > > > I'd say you're not doing anything wrong and the patch should be updated. > > In the meanwhile I adapted the patch slightly to keep the low 16 bits > from the ATAG, since it doesn't use them itself, and added some access > #defines to get the CPU type part or the ATAG part and changed the dozen > or so places in our code using system_rev to use those access defines. > > I'm sure that won't satisfy anyone :-) but it lets me get on for now. > > -Andy -- To unsubscribe from this list: send the line "unsubscribe cpufreq" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
