| 00:00.32 | m4t | i somewhat figured out how to use a gdb-cross with vmlinux to connect to qemu, and found that the kernel seems to go into a panic/do_blink loop around walnut_probe() |
| 00:02.20 | m4t | i want to see it might panic here: unsigned long root = of_get_flat_dt_root(); |
| 00:02.34 | m4t | which is where i think the addresses in walnut.dts are probed |
| 00:02.57 | m4t | and there *is* a section there describing nvram/rtc, though the panic still seems to occur with it commented out |
| 00:03.51 | m4t | s/see/say/ |
| 00:04.28 | m4t | s/to see/to say/ |
| 00:04.34 | m4t | yea |
| 00:04.36 | m4t | anyhow |
| 00:06.20 | m4t | i just got my psu in the mail so i've moved my efforts onto the actual board, rather than qemu |
| 00:08.24 | GPSFan | m4t: you know that the board described on the elinux wiki is not really a walnut. it is very similar but there are differences. |
| 00:09.32 | m4t | GPSFan yea i got a dht-walnut |
| 00:10.25 | GPSFan | yes and that is not the walnut platform in the kernel |
| 00:12.26 | m4t | was it deprecated |
| 00:12.43 | m4t | ? |
| 00:12.57 | GPSFan | the latest kernel I ran on mine was 2/6/16-rc3. the ppcboot really needed updating. I had a version of 1.1.6 but dfarnsworth had uboot running on it. |
| 00:13.25 | m4t | i got 1.2.0 working |
| 00:13.35 | m4t | i just up-ported the openembedded patch a few releases |
| 00:13.47 | GPSFan | the support code was never really put into mainline, so I don't know it's status. mine runs as a music server 24/7 and has for several years |
| 00:13.51 | m4t | the image that comes out works in the emulator |
| 00:13.57 | m4t | hrm cool |
| 00:14.20 | m4t | i might just try going with the 2.6.23 they have |
| 00:14.46 | m4t | after that it seems like the ppc/walnut disappears and it becomes with powerpc/walnut with walnut.dts rather than a bunch of addresses/registers in walnut.c |
| 00:15.05 | m4t | -with |
| 00:16.08 | m4t | it seems the only thing that was done to work on the dht-walnut in openembedded was a 'remove_todc' patch which removes rtc support |
| 00:16.16 | m4t | so i tried to follow in that direction |
| 00:16.25 | GPSFan | bbl |
| 00:17.11 | m4t | yea, later |
| 00:21.22 | m4t | yea one of them was done by montevista in 00-02 and the recent walnut work was done by ibm in 07 it seems |
| 00:21.37 | m4t | itd be sweet to get a dht_walnut.c and dht_walnut.dts goin |
| 01:32.46 | *** join/#elinux Dr_Who (n=tgall@gentoo/developer/dr-who) |
| 01:44.10 | *** join/#elinux likewhoa (n=y0@ool-45717734.dyn.optonline.net) |
| 04:54.39 | m4t | i am headed the direction of 2.4.37.2 |
| 07:24.24 | *** join/#elinux Christos_N (n=Christos@ppp-94-66-54-74.home.otenet.gr) |
| 07:27.14 | *** join/#elinux Xires (n=Xires@66-190-79-122.dhcp.dntn.tx.charter.com) |
| 07:40.04 | *** join/#elinux Toi_ (n=toi@62.140.137.61) |
| 08:29.04 | *** join/#elinux lyakh (n=lyakh@p57BD1080.dip0.t-ipconnect.de) |
| 09:04.06 | *** join/#elinux lyakh (n=lyakh@p57BD1080.dip0.t-ipconnect.de) |
| 09:07.01 | *** join/#elinux Toi_ (n=toi@37.85.28.81-mn.aerea.nl) |
| 10:01.48 | *** join/#elinux pirho (i=pirho@gateway/gpg-tor/key-0x2CEEC9CB) |
| 12:03.37 | *** join/#elinux Omegamoon (n=egoe@demolitionman.xs4all.nl) |
| 12:20.56 | *** join/#elinux Toi_ (n=toi@62.140.137.27) |
| 13:05.59 | *** join/#elinux Dr_Who (n=tgall@gentoo/developer/dr-who) |
| 13:16.17 | *** join/#elinux GPSFan (n=kenm@64.92.145.112) |
| 13:21.02 | *** join/#elinux orges (n=yet@88.251.254.198) |
| 13:29.18 | *** join/#elinux thraxisp (n=thraxisp@24.139.16.154) |
| 13:45.14 | *** join/#elinux Dr_Who (n=tgall@gentoo/developer/dr-who) |
| 14:21.55 | *** join/#elinux Toi_ (n=toi@62.140.137.26) |
| 14:22.04 | m4t | ## Transferring control to Linux (at address 00000000) ... |
| 14:22.04 | m4t | Warning, memory limited to 768 Mb, use CONFIG_HIGHMEM to reach 4079 Mb |
| 14:22.04 | m4t | Linux version 2.4.37.2 (matt@deep-thought) (gcc version 4.4.0 (GCC) ) #1 Sun Jul 12 09:53:28 EDT 2009 |
| 14:22.04 | m4t | IBM Walnut port (C) 2000-2002 MontaVista Software, Inc. (source@mvista.com) |
| 14:22.04 | m4t | On node 0 totalpages: 196608 |
| 14:22.05 | m4t | zone(0): 196608 pages. |
| 14:22.07 | m4t | zone(1): 0 pages. |
| 14:22.09 | m4t | zone(2): 0 pages. |
| 14:22.11 | m4t | Kernel command line: console=ttyS0,9600 |
| 14:22.13 | m4t | Warning: real time clock seems stuck! |
| 14:22.15 | m4t | :--) |
| 14:22.53 | m4t | i had it set up real nice, modifed to say IBM DHT-Walnut port and everything |
| 14:23.01 | m4t | but i didnt get console output |
| 14:26.07 | m4t | this world of walnut thing is kind of time consuming |
| 14:26.12 | m4t | WoW |
| 14:56.57 | m4t | hrm so i got rid of the rtc code |
| 14:57.18 | m4t | and set a few of the ppc_md.times to NULL, mimicing a few other board.c files |
| 14:58.26 | m4t | i dont get any further than Kernel command line: console=ttyS0,9600 |
| 14:58.29 | m4t | though |
| 15:17.24 | *** join/#elinux dijenerate (n=dijenera@64.210.44.91) |
| 15:19.30 | *** join/#elinux Christos_N (n=Christos@ppp-94-66-46-5.home.otenet.gr) |
| 15:32.16 | m4t | freezes right at sti(); |
| 15:32.19 | m4t | whatever that is |
| 16:57.29 | *** join/#elinux lyakh (n=lyakh@p57BD1080.dip0.t-ipconnect.de) |
| 17:14.49 | *** join/#elinux Dr_Who (n=tgall@gentoo/developer/dr-who) |
| 18:19.29 | *** join/#elinux djszapi_ (n=djszapi@94-21-75-156.pool.digikabel.hu) |
| 18:19.32 | djszapi_ | Hello! |
| 18:20.54 | djszapi_ | Is there any pragma operator facility with gcc, like this: _Pragma("interrupt")? __attribute__ ((interrupt_handler)) <- maybe this ? |
| 18:21.14 | *** join/#elinux booxter (n=booxter@212.98.182.22) |
| 18:31.13 | *** join/#elinux Christos_N (n=Christos@ppp-94-66-46-5.home.otenet.gr) |
| 19:53.30 | *** join/#elinux orges (n=yet@88.251.254.198) |
| 20:35.28 | *** join/#elinux Dr_Who (n=tgall@gentoo/developer/dr-who) |
| 22:49.58 | *** join/#elinux Dr_Who (n=tgall@gentoo/developer/dr-who) |
| 22:50.11 | MonMotha | djszapi_: there is an interrupt "attribute" on many architectures |
| 22:50.21 | MonMotha | but you can't turn it into a pragma that I know of |
| 22:50.40 | MonMotha | gcc's use of "pragma" is pretty limited. They're not fans of how "non-standard" it is. |
| 22:51.19 | MonMotha | what many people do is use macros to turn them into fake keywords so you can do e.g. "void handler my_interrupt() { ... }" |
| 22:51.40 | MonMotha | the Linux kernel takes that approach with several attributes and does something somewhat similar for interrupts. |
| 23:34.24 | *** join/#elinux thraxisp (n=thraxisp@24.139.16.154) |
| 23:41.37 | *** join/#elinux Dr_Who (n=tgall@gentoo/developer/dr-who) |