Thread (18 messages) 18 messages, 6 authors, 2014-08-06
STALE4358d

Re: [PATCH] cpufreq: tests: Providing cpufreq regression test

From: Lukasz Majewski <hidden>
Date: 2014-07-18 11:59:33
Also in: linux-arm-kernel, linux-samsung-soc

Hi Sachin,
Hi Lukasz,

I tested this script on 4210 based Origen board.
This is the output:
./cpufreq_freq_test.sh
CURRENT GOVERNOR: performance
SET GOVERNOR: performance
######################################
TEST AVAILABLE FREQS
######################################
FREQ: 1200000 sleep: invalid number '0.1'
[    5.918347] random: gzip urandom read with 61 bits of entropy
available OK
FREQ: 1000000 sleep: invalid number '0.1'
         OK
FREQ: 800000 sleep: invalid number '0.1'
         OK
FREQ: 500000 sleep: invalid number '0.1'
         OK
FREQ: 200000 sleep: invalid number '0.1'
         OK
######################################
TEST FREQS SWITCHING
######################################
REFERENCE FREQ: 1200000
FREQ: 1200000 ----> FREQ: 1200000 sleep: invalid number '0.1'
         OK
FREQ: 1200000 ----> FREQ: 1000000 sleep: invalid number '0.1'
         OK
FREQ: 1200000 ----> FREQ: 800000 sleep: invalid number '0.1'
         OK
FREQ: 1200000 ----> FREQ: 500000 sleep: invalid number '0.1'
         OK
FREQ: 1200000 ----> FREQ: 200000 sleep: invalid number '0.1'
         OK
REFERENCE FREQ: 1000000
FREQ: 1000000 ----> FREQ: 1200000 sleep: invalid number '0.1'
         OK
FREQ: 1000000 ----> FREQ: 1000000 sleep: invalid number '0.1'
         OK
FREQ: 1000000 ----> FREQ: 800000 sleep: invalid number '0.1'
         OK
FREQ: 1000000 ----> FREQ: 500000 sleep: invalid number '0.1'
         OK
FREQ: 1000000 ----> FREQ: 200000 sleep: invalid number '0.1'
         OK
REFERENCE FREQ: 800000
FREQ: 800000 ----> FREQ: 1200000 sleep: invalid number '0.1'
         OK
FREQ: 800000 ----> FREQ: 1000000 sleep: invalid number '0.1'
         OK
FREQ: 800000 ----> FREQ: 800000 sleep: invalid number '0.1'
         OK
FREQ: 800000 ----> FREQ: 500000 sleep: invalid number '0.1'
         OK
FREQ: 800000 ----> FREQ: 200000 sleep: invalid number '0.1'
         OK
REFERENCE FREQ: 500000
FREQ: 500000 ----> FREQ: 1200000 sleep: invalid number '0.1'
         OK
FREQ: 500000 ----> FREQ: 1000000 sleep: invalid number '0.1'
         OK
FREQ: 500000 ----> FREQ: 800000 sleep: invalid number '0.1'
         OK
FREQ: 500000 ----> FREQ: 500000 sleep: invalid number '0.1'
         OK
FREQ: 500000 ----> FREQ: 200000 sleep: invalid number '0.1'
         OK
REFERENCE FREQ: 200000
FREQ: 200000 ----> FREQ: 1200000 sleep: invalid number '0.1'
         OK
FREQ: 200000 ----> FREQ: 1000000 sleep: invalid number '0.1'
         OK
FREQ: 200000 ----> FREQ: 800000 sleep: invalid number '0.1'
         OK
FREQ: 200000 ----> FREQ: 500000 sleep: invalid number '0.1'
         OK
FREQ: 200000 ----> FREQ: 200000 sleep: invalid number '0.1'
         OK
######################################
ERRORS: 0
######################################

Though it says 0 errors, what does the "invalid number..." signify?
I guess that this message is caused by your default sleep
implementation.

Could you type 'sleep 0.1' and then 'sleep 1' in your console on the
target system?
Is the "invalid number" not present with the second case?


On Fri, Jul 18, 2014 at 3:53 PM, Lukasz Majewski
[off-list ref] wrote:
quoted
This commit adds first regression test "cpufreq_freq_test.sh" for
the cpufreq subsystem.

Signed-off-by: Lukasz Majewski <redacted>
---
 drivers/cpufreq/tests/README               |  23 +++++
 drivers/cpufreq/tests/cpufreq_freq_test.sh | 149
+++++++++++++++++++++++++++++ 2 files changed, 172 insertions(+)
 create mode 100644 drivers/cpufreq/tests/README
 create mode 100755 drivers/cpufreq/tests/cpufreq_freq_test.sh
diff --git a/drivers/cpufreq/tests/README
b/drivers/cpufreq/tests/README new file mode 100644
index 0000000..66638d2
--- /dev/null
+++ b/drivers/cpufreq/tests/README
@@ -0,0 +1,23 @@
+This file contains list of cpufreq's available regression tests
with a short +usage description.
+
+1. cpufreq_freq_test.sh
+
+Description:
+------------
+This test is supposed to test if cpufreq attributes exported by
sysfs are
s/test/script would be better
Yes, you are right.
quoted
+exposing a correct values.
s/ exposing a correct values / exposing correct values
quoted
+
+It can work with or without boost enabled and helps spotting
errors related to
s/ helps spotting / helps in spotting
Thanks for feedback.
<snip>
quoted
+
+set +x
+
+COLOUR_RED="\33[31m"
+COLOUR_BLUE="\33[34m"
+COLOUR_GREEN="\33[32m"
+COLOUR_DEFAULT="\33[0m"
+
+T_PATCH=/sys/devices/system/cpu/cpu0/cpufreq
Shouldn't this be called PATH instead of PATCH?
Hmm.... It really should be PATH :-).
quoted
+BOOST_PATCH=/sys/devices/system/cpu/cpufreq
ditto and rest of the places in the document.
Ok, I will correct that.

-- 
Best regards,

Lukasz Majewski

Samsung R&D Institute Poland (SRPOL) | Linux Platform Group
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help