From: Stephen White Subject: longhaul driver broken? Date: Sun, 1 Feb 2004 15:44:49 +0000 Sender: cpufreq-bounces+glkc-cpufreq=gmane.org@www.linux.org.uk Message-ID: <20040201154449.GA30954@the.earth.li> Mime-Version: 1.0 Return-path: Content-Disposition: inline List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: cpufreq-bounces+glkc-cpufreq=gmane.org@www.linux.org.uk Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: cpufreq@www.linux.org.uk I'm using the 2.4 backport (from http://ftp.linux.org.uk/pub/linux/cpufreq/) because of issues with drivers for another piece of hardware in my system. However it appears the longhaul driver is rather broken. The first problem is in the longhaul_table code. The 'index' member of each element in the table gets initialised as follows: longhaul_table[k].index = (j << 8); The 'longhaul_target' routine then obtains this value using: new_clock_ratio = longhaul_table[table_index].index & 0xFF; Thus 'new_clock_ratio' is /always/ found to be zero and so the driver has no effect. If I fix this then I can change the speed of my processor, but things still don't behave perfectly. I haven't looked into this, but I think this may in part be due to the rather random ordering of elements in the table, debug reveals: longhaul: longhaul_table[0].frequency = 399 (ratio 30, fsb 133) longhaul: longhaul_table[1].frequency = 532 (ratio 40, fsb 133) longhaul: longhaul_table[2].frequency = 465 (ratio 35, fsb 133) longhaul: longhaul_table[3].frequency = 598 (ratio 45, fsb 133) longhaul: longhaul_table[4].frequency = 731 (ratio 55, fsb 133) longhaul: longhaul_table[5].frequency = 798 (ratio 60, fsb 133) longhaul: longhaul_table[6].frequency = 665 (ratio 50, fsb 133) [ as an aside does anyone know the minimum safe frequency for a VIA Ezra C3 800 (family 7, mode 7, stepping 8). It seems to work at 400Mhz without crashing, but I'm not sure it's supposed to. ] Does anyone have some insight into the longhaul driver, or is there a better version somewhere that I should be using? If this is the code I'm supposed to be using and there's no-one fixing it, is anyone interested in receiving some patches for it from me? Thanks, Stephen