From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dominik Brodowski Subject: TODO list, patches will follow... Date: Wed, 19 Nov 2003 19:29:28 +0100 Sender: cpufreq-bounces@www.linux.org.uk Message-ID: <20031119182928.GA20576@brodo.de> 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@www.linux.org.uk Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: cpufreq@www.linux.org.uk, davej@codemonkey.org.uk Here are a few TODO items I'm aware of for cpufreq in 2.6. * denotes a patch exists for 2.6.0-test9 and will be sent out in a few moments. Merged in cpufreq-bk -------------------- - [speedstep-smi] fail on failure to aquire ownership - [speedstep-centrino] 1 GHz Centrino support - [powernow-k7] different error message on no PST for CPU - [powernow-k8] resolve one #warning 2.6.0 / <2.6.2 -------------- * [speedstep-smi] Disable smi_detect_freqs() call on systems which do not support it [BUG #1422] * [speedstep-smi] Disable debug output * [speedstep-smi] Spelling fix * [speedstep-lib] Support for 533 MHz FSB * [p4-clockmod] Detect CPU speed without relying on cpu_khz * [core] remove unneeded #ifdefs in include/linux/cpufreq.h * [core] latency in nanoseconds -- sometimes nanoseconds are used, sometimes microseconds, sometimes even something else. * [speedstep-smi] Do something about "cpufreq: change failed" Cause seems to be DMA by IDE and others. Possible soultions are mdelay(250), notice to users, ignoring it, retry etc. >2.6.2 / 2.7 ------------ - [powernow-k8] make it aware of frequency table helpers Have patches, AMD wanted to look into this - [powernow-k7] override for PSTs - [acpi-cpufreq] major rewrite patches exist, Len Brown liked them, but not now... - [core] split up drivers/cpufreq.c into manageable pieces - [core] use ->range() instead of ->setpolicy - [longrun] remove special-handling of "policy" - [ondemand] merge if patent issue is resolved - [core] support of CPU groups (hyperthreading, SMT) use cpumask_t instead of single CPU number - [core] verify ->resume support is OK (I doubt it) 2.7 --- - [sa1100] move cpufreq_get code to arch/arm/sa1100-specific code - [core] remove deprecated interfaces namely proc_intf, /proc/sys/ interface in userspace governor