Thread (2 messages) 2 messages, 1 author, 2014-09-12

state[i].transition_latency and state[i].bus_master_latency set to 0. BUG?

From: Ricardo Ribalda Delgado <hidden>
Date: 2014-08-05 15:38:18
Also in: linux-acpi

Hello!

I am working with an AMD family 14h Processor. And I believe that the
governor is not behaving as expected.

On the _PSS table, the transition_latency and bus_master_latency are set to 0.

The AMDs BIOS development guide (page 60 and 61)
http://developer.amd.com/wordpress/media/2012/10/43170_14h_Mod_00h-0Fh_BKDG.pdf
they recommend to use value of 0.

Is 0 valid value? The governor will work properly with this value?

What could be a good default value?

Is there a way to experimentally measure the latency?

Thanks for your help!

Extract from dmesg:
[    0.843340] Extracting state 0
[    0.843374] State [0]: core_frequency[1650] power[6075]
transition_latency[0] bus_master_latency[0] control[0x0] status[0x0]
[    0.843481] Extracting state 1
[    0.843513] State [1]: core_frequency[1320] power[3712]
transition_latency[0] bus_master_latency[0] control[0x1] status[0x1]
[    0.843618] Extracting state 2
[    0.843652] State [2]: core_frequency[825] power[1472]
transition_latency[0] bus_master_latency[0] control[0x2] status[0x2]
[    0.843757] Extracting state 3
[    0.843790] State [3]: core_frequency[528] power[1008]
transition_latency[0] bus_master_latency[0] control[0x3] status[0x3]


-- 
Ricardo Ribalda
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help