#19444: 1-wire functions are broken by broken __udelay() at ramips
-------------------------+-------------------------------------------------
Reporter: SergeiP | Owner: developers
Type: defect | Status: new
Priority: high | Milestone: Chaos Calmer (trunk)
Component: kernel | Version: Trunk
Resolution: | Keywords: w1, 1-wire, udelay, calibrating
| delay loop
-------------------------+-------------------------------------------------
Comment (by SergeiP):
This look suspiciously
[ 0.000000] CPU Clock: 360MHz
>>[ 0.000000] systick: running - mult: 214748, shift: 32
[ 0.000000] Calibrating delay loop... 479.23 BogoMIPS (lpj=2396160)
....
>>[ 0.130000] Switched to clocksource systick
It looks like delay loop is calculated at timer speed 50 Hz,
and later timer switched to selected value (100 Hz).
===
Ops!
This assumption is wrong.
calculated lpj value is changed with timer period settings.
So ether timer tick period is twice longer of nominal,
ether processor clock is really slowed down later ...
--
Ticket URL: <https://dev.openwrt.org/ticket/19444#comment:7>
OpenWrt <http://openwrt.org>
Opensource Wireless Router Technology
_______________________________________________
openwrt-tickets mailing list
[email protected]
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-tickets
