blob: 3a916a5753f1a31546a471c9575f58e00d1f108a (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
|
- Don't output same info more than once on SMP, just make a note
that there are N identical CPUs.
- feature flag handling code could use more cleanup.
- Make intel cache sizing use cpuid(4) where present.
- Fix benchmarking
- introduce hexdump option (-x) to dump register values in addition to decoding.
- Identify Transmeta CPUs.
- Decode mtrr registers as well as dumping them
- Dump cyrix ARR registers when we pass --mtrr
- --apic
Dump APIC registers
- --gcc option to output best optimisation parameters for that CPU.
- Some CPUs allow the name string to be changed.
If this has happened we die horribly. We need to store the current
name string, perform the cpuid reset sequence, and restore the
name string on exit.
- After identifying CPU we can associate more info with the CPU.
- CPU codename
- Introduction date
- number of transistors
- voltages (I/O, Core, Maximum)
- number of pipelines
- review code for extended natsemi support.
- report L3 (chipset) caches (requires PCI poking)
- might be time to start thinking about XML for CPU descriptions?
- libx86info for other apps
- Be able to tell the user when we have failed to identify the chip, so they can send us information!
|
