Modify

Opened 11 years ago

Closed 11 years ago

#1690 closed defect (fixed)

ar7 TNET7200 clock initialisation hangs kernel

Reported by: hydra Owned by: florian
Priority: high Milestone:
Component: kernel Version:
Keywords: Cc: mailinglist.openwrt@…

Description

I've been trying to track down a kernel boot problem on my Speedcom W701V AR7 device and have got as far as finding the function call that's causing it to die.

build_mipsel/linux-2.6-ar7/linux-2.6.19.2/arch/mips/ar7/clock.c

        prom_printf("!CLOCK3!\n");
        if (ar7_dsp_clock == 250000000)
                tnetd7200_set_clock(ar7_ref_clock, &clocks->dsp,
                                    bootcr, ar7_dsp_clock);

        prom_printf("!CLOCK4!\n");

When the system boots it displays "!CLOCK3!" followed by some corrupted garbage when monitoring the boot process via an attached serial cable.

This may be related to r1649, but at first glance doesn't appear to be directly related other than it's the same source code file (clock.c). The reporter of r1649 has a DG834G(v2) which (I think) has a TNETD7300 CPU. My T-Com Speedport W701V definitely has a TNETD7200 CPU (as can be seen from the photos on the wiki page I created here: http://wiki.openwrt.org/OpenWrtDocs/Hardware/T-Com/SpeedportW701V )

Attachments (0)

Change History (11)

comment:1 Changed 11 years ago by nabcore

I, the reporter of r1649, confirm I have a TNETD7300

comment:2 Changed 11 years ago by jb24

I can also confirm that my T-Com Speedport W900V definitely has a TNETD7200ZDW CPU.
see here ->http://wiki.openwrt.org/OpenWrtDocs/Hardware/T-Com/SpeedportW900V

comment:3 Changed 11 years ago by hydra

oops, we both mean #1649, not r1649 above

comment:4 Changed 11 years ago by florian

kernel developpers themselves think that volatile should not be used, see this thread : http://www.mail-archive.com/linux-kernel%40vger.kernel.org/msg154629.html

Might be worth the try to revert changeset 7119

comment:5 Changed 11 years ago by hydra

that was tried using a fresh svn checkout followed by an svn merge -r 7119:7118 on the checkout. I confirmed the reverse diff had been applied via svn diff.

comment:6 Changed 11 years ago by Dan

Maybe the information on http://forums.whirlpool.net.au/forum-replies.cfm?t=511748&p=9 is useful for a dev.

comment:7 Changed 11 years ago by florian

  • Owner changed from developers to florian
  • Status changed from new to assigned

comment:8 Changed 11 years ago by florian

  • Resolution set to fixed
  • Status changed from assigned to closed

Fixed with [7487], though I agree wreduce with the forum thread that we should probably reduce the DSP clock, however I am not sure about the right frequency to put in there.

comment:9 Changed 11 years ago by loswillios

  • Resolution fixed deleted
  • Status changed from closed to reopened

attached kernel with printks to find out the correct frequency. Don't know how to include it in openwrt-image to actually boot it.

comment:11 Changed 11 years ago by florian

  • Resolution set to fixed
  • Status changed from reopened to closed

Should be fixed with [8036]

Add Comment

Modify Ticket

Action
as closed .
The resolution will be deleted. Next status will be 'reopened'.
Author


E-mail address and user name can be saved in the Preferences.

 
Note: See TracTickets for help on using tickets.