| 00:00.12 | cr2 | yeah, seems to be ok for RaphaelNAND |
| 00:00.46 | cr2 | all the sizes seem to be 0x20000 aligned. |
| 00:01.08 | *** join/#htc-linux Moku (n=John@f049115225.adsl.alicedsl.de) |
| 00:06.46 | cr2 | patching haret seems to be the easiest solution. but i also need to pass the readonly flag somehow. |
| 00:07.51 | *** join/#htc-linux hairyraven (n=nobody@209.250.241.104) |
| 00:13.58 | *** join/#htc-linux miknix (n=miknix@gentoo/developer/miknix) |
| 00:15.52 | cr2 | good night |
| 00:25.11 | miknix | hum.. |
| 00:25.18 | miknix | no chance to say good night |
| 00:25.50 | miknix | : P |
| 01:02.26 | *** join/#htc-linux NetRipper (n=netrippe@netripper.nl) |
| 03:08.16 | *** join/#htc-linux mrmoku|easter (n=mrmoku@93.104.110.217) |
| 03:52.40 | *** join/#htc-linux Moult (n=Moult@60.54.126.114) |
| 04:04.54 | *** join/#htc-linux Shinto (n=John@f049172219.adsl.alicedsl.de) |
| 04:14.36 | *** join/#htc-linux marex (n=marex@vasut.kolej.mff.cuni.cz) |
| 04:26.49 | *** join/#htc-linux marex (n=marex@vasut.kolej.mff.cuni.cz) |
| 04:42.23 | *** join/#htc-linux timebomb (n=tb@d027099.adsl.hansenet.de) |
| 04:53.25 | *** join/#htc-linux marex (n=marex@vasut.kolej.mff.cuni.cz) |
| 05:53.53 | *** join/#htc-linux Magorium (i=Othello@gateway/tor/x-c62265e908ff3fbe) |
| 06:54.22 | *** join/#htc-linux Othello (i=Othello@gateway/tor/x-8450995257debf56) |
| 08:05.58 | *** join/#htc-linux TheiPirate (i=user@wireless-128-62-144-16.public.utexas.edu) |
| 08:36.34 | *** join/#htc-linux nebi_ (n=nebi@217.142.147.19) |
| 08:37.17 | *** join/#htc-linux TheiPirate (i=user@wireless-128-62-144-16.public.utexas.edu) |
| 08:42.02 | *** join/#htc-linux pleemans (n=toi@d54C2AAB7.access.telenet.be) |
| 08:49.44 | *** join/#htc-linux syn-fin (n=alien@75.74.251.191) |
| 08:54.27 | *** join/#htc-linux dcordes (n=dcordes_@unaffiliated/dcordes) |
| 08:58.01 | *** join/#htc-linux pleemans (n=toi@d54C2AAB7.access.telenet.be) |
| 09:05.48 | *** join/#htc-linux syn-fin (n=alien@c-75-74-251-191.hsd1.fl.comcast.net) |
| 09:08.07 | *** join/#htc-linux syn-fin1 (n=alien@c-75-74-251-191.hsd1.fl.comcast.net) |
| 09:09.27 | TheiPirate | anyone awake |
| 09:17.01 | dcordes | not me |
| 09:26.02 | dcordes | TheiPirate, anything I can help with? |
| 09:27.12 | TheiPirate | im trying to test something but i think it is a network issue nothing else |
| 09:27.14 | TheiPirate | thanks though |
| 09:40.20 | *** join/#htc-linux nebi_ (n=nebi@217.142.147.19) |
| 09:42.55 | *** join/#htc-linux dzo (n=dzo@121-98-128-127.bitstream.orcon.net.nz) |
| 09:42.55 | *** join/#htc-linux BabelO (n=fcr@unaffiliated/babelo) |
| 10:01.46 | *** join/#htc-linux SockPuppet (n=X@unaffiliated/egns) |
| 10:01.50 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 10:02.22 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 10:02.52 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 10:03.53 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 10:04.47 | *** join/#htc-linux StarLite (n=nnscript@s55916ca6.adsl.wanadoo.nl) |
| 10:06.10 | *** join/#htc-linux cr2 (n=cr2@ip-77-25-122-96.web.vodafone.de) |
| 10:08.04 | *** join/#htc-linux Moult (n=Moult@60.54.126.114) |
| 10:09.53 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 10:16.39 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 10:17.18 | tmzt | cr2: trying to make fixup work on raph500 |
| 10:19.25 | tmzt | changing node1 size to 64m doesn't work |
| 10:20.02 | tmzt | since mem=72 works it might be node0 |
| 10:20.18 | tmzt | and linux-size that are wrong |
| 10:24.01 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 10:24.37 | *** join/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 10:24.48 | *** part/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 10:30.39 | cr2 | tmzt: you have 64+128 ? |
| 10:31.00 | cr2 | tmzt: is it possible to write to the second 32MB sram bank ? |
| 10:31.30 | tmzt | I don't know |
| 10:31.41 | tmzt | what does linux-size 89m mean? |
| 10:31.54 | tmzt | how can it be more thab sram |
| 10:32.26 | tmzt | I have 192m reported by dev info |
| 10:34.26 | tmzt | oh, 3 banks? |
| 10:34.43 | tmzt | 32 32 128 |
| 10:34.55 | cr2 | the sram banks are 32+32 |
| 10:35.17 | tmzt | but contig? |
| 10:35.20 | cr2 | but still you have the qcsbl+oemsbl+amss hole in the middle |
| 10:35.29 | tmzt | ok |
| 10:35.35 | cr2 | qcsbl+oemsbl+amss+smem |
| 10:35.40 | cr2 | to be more accurate |
| 10:35.56 | tmzt | so I need more than linux-size |
| 10:36.03 | cr2 | spl+free+qcsbl+oemsbl+amss+smem(+free) |
| 10:36.10 | tmzt | and we need a raph500 mtype |
| 10:36.29 | cr2 | register one |
| 10:36.57 | cr2 | but check if it already exists |
| 10:37.15 | cr2 | smart people registered kaiser 3 times after i've done it ;) |
| 10:38.50 | cr2 | i'm wondering if there is some way to remove mtypes |
| 10:40.26 | TheiPirate | anyone write code for android |
| 10:40.42 | cr2 | TheiPirate: #android |
| 10:44.33 | tmzt | htc-raphael-cdma500? |
| 10:50.00 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 10:50.45 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 10:54.56 | *** join/#htc-linux timebomb (n=tb@d027099.adsl.hansenet.de) |
| 10:58.44 | Dunedan | cr2: I built the latest kernel for blackstone and removed the mem-option from the kernel command line, but I still get 85MB of RAM. Am I doing something wrong? |
| 11:01.07 | tmzt | 2182 htcraphael-cdma500 |
| 11:01.24 | tmzt | (underscore) |
| 11:01.34 | Dunedan | oh, seems like I thought only included for raphael and not for blackstone :-( |
| 11:02.20 | tmzt | what board file are you using? |
| 11:02.30 | Dunedan | should be htcblackstone |
| 11:02.42 | tmzt | checkt the fixup function |
| 11:02.44 | Dunedan | I'll try it with enabled second bank now. |
| 11:04.53 | mickeyl | why do we get 211 and not 256 btw? |
| 11:04.56 | *** join/#htc-linux Moult (n=Moult@124.13.233.148) |
| 11:05.05 | mickeyl | who's eating the remaining 45MB? :) |
| 11:05.46 | tmzt | amss I think |
| 11:06.21 | tmzt | video, audio and some other stuff |
| 11:07.55 | *** join/#htc-linux Echo31 (i=52f0c236@gateway/web/ajax/mibbit.com/x-caa5cde46a1ba6cc) |
| 11:08.21 | mickeyl | k |
| 11:10.34 | *** join/#htc-linux PoohbaLT1 (n=BlkPoohb@c-69-141-71-249.hsd1.nj.comcast.net) |
| 11:13.19 | Dunedan | may it possible, that the first bank of blackstone is larger than 89MB? Because with mem=108 I got more than 89MB of RAM. |
| 11:15.14 | cr2 | Dunedan: the amss part of sdram is protected (>=0x17300000 on raph), the rest can be used as you like. |
| 11:15.34 | cr2 | Dunedan: the sram is more fragmented. |
| 11:16.10 | cr2 | 128-13=115MB(+128)=243MB |
| 11:17.25 | Dunedan | yeah |
| 11:17.29 | Dunedan | got 212MB now |
| 11:17.39 | cr2 | spl is 512K, so 8-0.5=7.5MB sram can be used too (+32 on raph500) |
| 11:18.09 | cr2 | so the full amount of "usable" ram is 243+7.5=250.5MB |
| 11:18.45 | Dunedan | Somebody with git-access could add the second bank in board-htcblackstone.c. It's working. |
| 11:18.54 | cr2 | but the "nadroid" kernel reserves some ram for some other, not well understood purposes. |
| 11:19.27 | Dunedan | well, I don't want android. I just want pure debian. ;-) |
| 11:19.36 | cr2 | i think we can drop some of them for non-android distro |
| 11:20.22 | cr2 | and the rest can be optimised in size, but we need some better understanding of the constraints. |
| 11:21.07 | cr2 | Dunedan: that's why i'm for the non-android kernel config. |
| 11:21.07 | tmzt | does X need/use double bufering? |
| 11:21.39 | cr2 | no |
| 11:22.04 | cr2 | the memory allocation is up to the Xserver itself |
| 11:22.19 | tmzt | so what do I need for raph? |
| 11:22.22 | cr2 | if it wants to do double-buffering |
| 11:22.24 | tmzt | raph500 |
| 11:22.26 | *** join/#htc-linux MethoS (n=clemens@host-091-097-241-235.ewe-ip-backbone.de) |
| 11:22.56 | cr2 | tmzt: try to write/read the second sram bank it haret. |
| 11:23.13 | cr2 | tmzt: if it will work, you can add the thirt |
| 11:23.29 | cr2 | s/thirt/third memory bank/ |
| 11:24.43 | cr2 | tmzt: we also need to clarify the GPU0 and GPU1 purpose. |
| 11:37.01 | *** join/#htc-linux kiozen (n=oeichler@rgnb-5d87ccc6.pool.einsundeins.de) |
| 11:38.32 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 11:38.43 | Echo31 | I am new, French, and I want improved the development of HTC athena |
| 11:40.03 | Echo31 | so, I am a beginner for using IRC |
| 11:50.00 | tmzt | cr2: I think those are for offscreen buffers for 3d |
| 11:50.32 | tmzt | how do I write in haret? |
| 11:51.08 | tmzt | Echo31: hello, do you mean linux on athena? |
| 11:51.25 | *** join/#htc-linux Luzdegas_ (n=X@23.Red-81-34-72.dynamicIP.rima-tde.net) |
| 11:56.09 | Echo31 | yes, the kernel development dedicated for athena seems stopped |
| 11:57.32 | mickeyl | i'm afraid the handhelds.org view of the world killed many kernel ports |
| 12:03.28 | Echo31 | I found only two sources : www.handlelds.org and http://github.com/mistadman/android-athena/ |
| 12:04.25 | tmzt | what's the second one? anybody know who work on it? |
| 12:04.49 | cr2 | mickeyl: not really the hh.org itself, but the fact that there is no coordinated kernel development in this area. |
| 12:05.33 | cr2 | tmzt: "android"-specific kernel development sounds silly from the very beginning. |
| 12:05.55 | mickeyl | cr2: true, with "view of the world" i meant out-of-kernel-development with ignoring upstream, not hh.org as an institute (which had its own problems, but that's another story) |
| 12:07.03 | cr2 | mickeyl: linux-arm was very separared from linux-x86 too |
| 12:07.37 | mickeyl | sure, but where other communities already worked with upstream (zaurus as a prime example), the folks surrounding hh.org kernel tree just didn't care |
| 12:07.49 | mickeyl | it doesn't matter anyways these days |
| 12:07.56 | mickeyl | just another form of lost momentum |
| 12:08.10 | mickeyl | that's why it's even more important to do better these days :) |
| 12:11.21 | cr2 | yes. |
| 12:11.40 | cr2 | Echo31: what is the latest kernel version for htc magician ? |
| 12:11.53 | *** join/#htc-linux timebomb (n=tb@c211099.adsl.hansenet.de) |
| 12:13.20 | tmzt | cr2: it's apparently 2.6.24 based |
| 12:13.33 | cr2 | Echo31: you should take the latest kernel code for magician first. |
| 12:13.41 | tmzt | magician? |
| 12:13.42 | cr2 | tmzt: i think it must be better |
| 12:14.13 | tmzt | magician is upstream, lates is 2.6.30-rcX |
| 12:14.31 | cr2 | tmzt: ok, so it should be the base. |
| 12:14.37 | tmzt | but it's not complete I think |
| 12:14.48 | cr2 | does it boot ? |
| 12:14.51 | tmzt | it appears mistadmin used android 2.6.24 |
| 12:15.04 | tmzt | don't know, don't have magician |
| 12:15.20 | tmzt | but I used it to build motoq kernel and it worked on there |
| 12:15.25 | *** join/#htc-linux marex (n=marex@vasut.kolej.mff.cuni.cz) |
| 12:15.58 | tmzt | github history makes no sense |
| 12:16.27 | cr2 | tmzt: porting hh.org athena code to magician 2.6.30 should be easy. |
| 12:16.54 | cr2 | tmzt: it's only the new htc-egpio interface that needs to be changed. |
| 12:17.46 | Echo31 | <PROTECTED> |
| 12:19.37 | cr2 | Echo31: yes, i have written it, and it works. but. it should not be the base for the further development. |
| 12:21.18 | Echo31 | yes, should I create a new project (git) to share these sources |
| 12:25.33 | Echo31 | how can i improve the ati w228x ? |
| 12:25.35 | cr2 | Echo31: i suggest adapting the magician code, so it will be completely rewritten, and once the new kernel boots it should be added to 2.6.30-rcX. |
| 12:25.49 | cr2 | pH5 will certanly support us. |
| 12:26.08 | cr2 | Echo31: which part of w228x ? |
| 12:28.29 | Echo31 | The mmc host device is still under debug mode. |
| 12:31.07 | cr2 | it's not working anyway |
| 12:31.31 | cr2 | and it is a modification of the asic3-mmc code |
| 12:31.53 | cr2 | which will not work on kernel >2.6.24 too |
| 12:32.42 | cr2 | Echo31: but the usb host works, so you can use an usb stick. |
| 12:32.58 | cr2 | for development and the rootfs |
| 12:33.31 | cr2 | i've even connected a 300GB hdd through the powered hub. |
| 12:33.58 | cr2 | usb1.1 is a bit slow, but still bearable. and better than nothing. |
| 12:34.09 | *** join/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 12:36.55 | Echo31 | I know, i run correctly the athena kernel with the system (Angstrom-x11-gpe-image) installed from mass storage usb device |
| 12:38.13 | Echo31 | also, i use the keyboard and ethernet interface on the usb host of athena. |
| 12:40.10 | *** part/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 12:44.16 | Dunedan | Did anybody got X working on blackstone? |
| 12:47.19 | Echo31 | Do we have enough technical information for this asics3-mmc modification? |
| 12:51.59 | tmzt | Dud: is it 16bit or 18bit? and do you have the update thread enabled in kernel |
| 12:52.05 | tmzt | Dunedan: |
| 12:52.44 | Dunedan | erm, dunno how much bit. And the update thread should be enabled. |
| 12:52.58 | tmzt | fbset |
| 12:53.33 | Dunedan | afaik fbset doesn't work, but wait I moment, I'll try again |
| 12:54.13 | tmzt | just the ouput from it |
| 12:55.38 | Dunedan | mode "480x800" |
| 12:55.40 | Dunedan | <PROTECTED> |
| 12:55.41 | Dunedan | <PROTECTED> |
| 12:55.43 | Dunedan | <PROTECTED> |
| 12:55.44 | Dunedan | endmode |
| 12:56.32 | tmzt | remove the *2 in kernel |
| 12:56.33 | cr2 | Dunedan: pixclock ? |
| 12:56.44 | tmzt | and it's 16, good |
| 12:56.54 | Dunedan | tmzt: which *2? |
| 12:56.57 | Dunedan | cr2: ? |
| 12:57.00 | cr2 | tmzt: epson is 16bit |
| 12:57.04 | tmzt | msmfb.c |
| 12:57.16 | Dunedan | I'll have a look. |
| 12:57.17 | cr2 | Dunedan: there is no frequency or pixclock in the output ? |
| 12:57.31 | Dunedan | cr2: that's the whole output of "fbset" |
| 12:57.45 | Dunedan | http://dpaste.com/35474/ <- my current Xorg.0.log (if that helps) |
| 12:57.48 | tmzt | that's the busybox version probably |
| 12:58.32 | Dunedan | it's the standard debian version |
| 12:59.20 | tmzt | ok |
| 13:00.00 | Dunedan | tmzt: addr = ((msmfb->xres * (yoffset + y) + x) * 2); <- this *2? |
| 13:00.02 | tmzt | you said fso works on blak? |
| 13:00.07 | tmzt | no |
| 13:00.31 | Dunedan | fso works. I can call the phone and do calls with the phone. |
| 13:00.46 | tmzt | virtual-yres |
| 13:00.59 | tmzt | which version? from where? |
| 13:01.06 | Dunedan | fb_info->var.yres_virtual = msmfb->yres * 2; <- this one? |
| 13:01.11 | tmzt | yes |
| 13:02.08 | *** join/#htc-linux dcordes-kais (n=dcordes-@ip-77-25-228-21.web.vodafone.de) |
| 13:02.22 | Dunedan | tmzt: I use the fso version from the DebianFSO-Team (http://wiki.debian.org/Teams/DebianFSO). It's a bit outdated, but it works. But I'll ask them to upload a newer version. |
| 13:02.36 | cr2 | tmzt: hehe. quality code here. |
| 13:02.41 | tmzt | check fso-userland |
| 13:03.14 | tmzt | there's a newer build linked from there |
| 13:03.32 | *** mode/#htc-linux [+o dcordes-kais] by ChanServ |
| 13:03.35 | cr2 | tmzt: isn't there some msmfb->depth value ? |
| 13:04.36 | cr2 | (II) FBDEV(0): hardware: msmfb (video memory: 2048kB) |
| 13:04.51 | *** join/#htc-linux wirelessdreamer (n=dreamer@chrobd01.vailsys.com) |
| 13:05.19 | tmzt | cr2: sorry? |
| 13:05.36 | cr2 | tmzt: does msmfb have the bpp parameter ? |
| 13:05.41 | tmzt | depth would be 16? |
| 13:06.03 | tmzt | this is about have double virtual y resolution for double buffering |
| 13:06.09 | cr2 | anything is better than just *2 |
| 13:06.10 | tmzt | X doens't like it |
| 13:06.28 | cr2 | (EE) FBDEV(0): FBIOPUT_VSCREENINFO: Invalid argument |
| 13:06.38 | cr2 | is this ioctl documented somewhere ? |
| 13:06.45 | tmzt | that's not bytes, it's pixels |
| 13:06.51 | tmzt | probably |
| 13:07.06 | tmzt | that's what fbset uses |
| 13:07.10 | cr2 | android-specific hack anyway. |
| 13:07.32 | tmzt | so you remove *2 from that line and X works |
| 13:07.34 | cr2 | hm. if fbset works, why does X uses wrong parameter ? |
| 13:07.49 | tmzt | because it builds an internal mode |
| 13:07.58 | tmzt | and then tries to set it |
| 13:07.58 | cr2 | tmzt: then it should be #ifdef ANDROID |
| 13:08.19 | cr2 | i mean *2 |
| 13:08.42 | cr2 | ah, it's PUTSCREENINFO |
| 13:09.03 | Dunedan | http://dpaste.com/35480/ <- some more fbset-output |
| 13:09.41 | Dunedan | (kernel without *2 is still building) |
| 13:09.59 | cr2 | Dunedan: strange. i think many moons ago fbset told the refreh frequency or pixclock |
| 13:11.55 | tmzt | it's probably the vyres though cr2 |
| 13:12.19 | tmzt | pixclock doesn't seem to matter outside of mdp/mddi |
| 13:12.22 | Dunedan | hm, kernel build fails without *2 :-/ |
| 13:12.32 | tmzt | paste the line/error |
| 13:13.10 | Dunedan | <PROTECTED> |
| 13:13.12 | cr2 | tmzt: but it's the only way to find out the [hv]refresh time |
| 13:13.12 | Dunedan | drivers/video/msm/msm_fb.c:656:3: error: invalid preprocessing directive #fb_info |
| 13:13.15 | Dunedan | make[2]: *** [drivers/video/msm] Fehler 2 |
| 13:13.16 | Dunedan | make[1]: *** [drivers/video] Fehler 2 |
| 13:13.19 | Dunedan | make[3]: *** [drivers/video/msm/msm_fb.o] Fehler 1 |
| 13:13.21 | Dunedan | make: *** [drivers] Fehler 2 |
| 13:13.31 | cr2 | ? |
| 13:13.38 | cr2 | #fb_info ? |
| 13:13.42 | Dunedan | arg |
| 13:13.44 | Dunedan | my fault |
| 13:14.27 | tmzt | no idea |
| 13:15.06 | Dunedan | I used # for a comment ... |
| 13:15.32 | tmzt | oh |
| 13:25.38 | Dunedan | looks better now |
| 13:25.57 | tmzt | works? |
| 13:26.36 | Dunedan | maybe |
| 13:26.58 | Dunedan | I still have a black screen now. |
| 13:27.12 | Dunedan | http://dpaste.com/35487/ <- new log |
| 13:27.41 | tmzt | that might be ok |
| 13:27.52 | tmzt | do you have usbnet? |
| 13:28.01 | tmzt | if so, try starting an xterm |
| 13:28.13 | Dunedan | but how? |
| 13:28.27 | tmzt | do you have usbnet? |
| 13:28.30 | tmzt | ssh in |
| 13:28.39 | tmzt | export DISPLAY=:0 |
| 13:28.40 | Dunedan | sure |
| 13:28.42 | tmzt | xterm& |
| 13:29.00 | Dunedan | debian-blackstone:/var/log# export DISPLAY=:0 |
| 13:29.01 | Dunedan | debian-blackstone:/var/log# xterm |
| 13:29.03 | Dunedan | Warning: This program is an suid-root program or is being run by the root user. |
| 13:29.06 | Dunedan | in this environment. You may get a more descriptive message by running the |
| 13:29.09 | Dunedan | The full text of the error or warning message cannot be safely formatted |
| 13:29.11 | Dunedan | program as a non-root user or by removing the suid bit on the executable. |
| 13:29.13 | Dunedan | xterm Xt error: Can't open display: %s |
| 13:29.32 | tmzt | ps aux PIPE X |
| 13:29.40 | tmzt | ps aux PIPE grep X |
| 13:29.51 | tmzt | more covert channle stuff |
| 13:30.05 | tmzt | not much data |
| 13:30.13 | Dunedan | nothing. So X might not be running?! |
| 13:30.20 | tmzt | right |
| 13:30.23 | tmzt | chvt 1 |
| 13:30.46 | Dunedan | ok |
| 13:30.53 | *** join/#htc-linux WyrM (n=wyrm@20158152015.user.veloxzone.com.br) |
| 13:31.13 | Dunedan | (EE) FBDEV(0): FBIOBLANK: Invalid argument |
| 13:31.14 | Dunedan | waiting for X server to begin accepting connections |
| 13:31.16 | Dunedan | giving up. |
| 13:31.17 | Dunedan | xinit: Connection reset by peer (errno 104): unable to connect to X server |
| 13:31.19 | Dunedan | xinit: No such process (errno 3): Server error. |
| 13:31.21 | Dunedan | xauth: (argv):1: bad display name "debian-blackstone:0" in "remove" command |
| 13:32.54 | tmzt | forget xinit for right now |
| 13:33.03 | tmzt | open a second ssh session |
| 13:33.23 | tmzt | hey WyrM |
| 13:33.25 | Dunedan | I currently have a screen session. |
| 13:33.34 | cr2 | hm. FBIOBLANK should work |
| 13:33.38 | tmzt | you need to windows |
| 13:33.44 | Dunedan | ok |
| 13:33.45 | tmzt | that's wierd though |
| 13:34.22 | Dunedan | I have two connections now. |
| 13:34.53 | tmzt | strace Xorg :0 |
| 13:36.48 | *** join/#htc-linux sabrod (n=sab@82.253.17.97) |
| 13:37.07 | Dunedan | Do you want the whole output or just the last lines? |
| 13:38.00 | tmzt | I want to see what the argument to IOBLANK is |
| 13:38.23 | Dunedan | ioctl(7, FBIOBLANK, 0) = -1 EINVAL (Invalid argument) |
| 13:38.56 | Dunedan | http://dpaste.com/35490/ <- here is the whole output |
| 13:39.17 | tmzt | cr2: what does that do? |
| 13:39.27 | tmzt | why is it being called now |
| 13:45.18 | *** join/#htc-linux cr2 (n=cr2@ip-77-25-122-96.web.vodafone.de) |
| 13:49.23 | cr2 | Dunedan: not implemented in msmfb.c ? |
| 13:49.37 | cr2 | where is the ioctl handler in msmfb.c ? |
| 13:49.51 | Dunedan | don't ask me. I don't no such things. ;-) |
| 13:50.05 | cr2 | :) |
| 13:51.41 | *** join/#htc-linux dzo (n=dzo@121-98-128-127.bitstream.orcon.net.nz) |
| 13:54.14 | tmzt | cr2: it seems some are passed up to another driver though I'm not quite sure |
| 13:55.02 | cr2 | it seems to be a generic fb call |
| 13:55.16 | tmzt | sorry, I mean that not all ioctls are registered with the device and there is a generic handler for the others |
| 13:55.17 | cr2 | so probably some pointer is missing. |
| 13:55.22 | tmzt | yeah |
| 13:55.24 | cr2 | ok |
| 13:55.44 | tmzt | oh, blank just clears the pallet doesn't it? |
| 13:55.58 | tmzt | but why is X calling it now |
| 13:56.14 | cr2 | no, it must be some dev-specific call |
| 13:56.20 | tmzt | unless this is X version which defaults to black background |
| 13:56.34 | cr2 | because of the acceleration |
| 13:56.48 | cr2 | standard x background should not be black ? |
| 13:57.16 | tmzt | it's normally grey |
| 13:57.25 | cr2 | hm. time to recheck how pxafb implements blank |
| 13:57.37 | cr2 | yes, the patterned grey |
| 13:57.39 | tmzt | it clears palette |
| 13:58.02 | tmzt | sets all entries to black |
| 13:58.26 | *** join/#htc-linux MethoS (n=clemens@dyndsl-085-016-165-162.ewe-ip-backbone.de) |
| 13:58.58 | cr2 | ok. |
| 13:59.04 | cr2 | need to look at others. |
| 14:00.29 | cr2 | the code comes from sa1100fb |
| 14:00.38 | cr2 | <PROTECTED> |
| 14:00.40 | cr2 | 772 * turn off the video (maybe by colormap tricks), but don't mess with the |
| 14:00.41 | cr2 | 773 * video itself: think of it semantically between on and Stand-By. |
| 14:00.43 | cr2 | 774 |
| 14:01.04 | cr2 | <PROTECTED> |
| 14:01.13 | cr2 | <PROTECTED> |
| 14:01.28 | cr2 | does msm_fb have fb_ops ? |
| 14:02.32 | Dunedan | yes |
| 14:02.39 | Dunedan | static struct fb_ops msmfb_ops = { |
| 14:02.41 | Dunedan | <PROTECTED> |
| 14:02.42 | Dunedan | <PROTECTED> |
| 14:02.44 | Dunedan | <PROTECTED> |
| 14:02.45 | Dunedan | <PROTECTED> |
| 14:02.46 | cr2 | <PROTECTED> |
| 14:02.48 | Dunedan | <PROTECTED> |
| 14:02.53 | Dunedan | <PROTECTED> |
| 14:02.55 | Dunedan | <PROTECTED> |
| 14:02.57 | cr2 | does full suspend. tis one is an accelerated driver. |
| 14:02.58 | Dunedan | <PROTECTED> |
| 14:03.00 | Dunedan | <PROTECTED> |
| 14:03.03 | Dunedan | }; |
| 14:03.31 | cr2 | Dunedan: ok, but no .fb_blank ? |
| 14:03.51 | Dunedan | seems so |
| 14:04.02 | cr2 | ok, needs to be implemented. |
| 14:04.12 | tmzt | wow |
| 14:04.21 | tmzt | but I didn't have that issue |
| 14:04.39 | tmzt | why is it missing? |
| 14:04.51 | cr2 | *android* |
| 14:05.27 | cr2 | is not and accerelated fb driver ;) |
| 14:06.25 | cr2 | it was my flamebait long ago. |
| 14:07.03 | cr2 | if they did follow unix way of doing things, all android stuff should have been on top of the accelerated fb driver. |
| 14:07.42 | tmzt | I think they only had that for omap/emu |
| 14:07.56 | tmzt | the gl version works well on final hardware |
| 14:08.40 | tmzt | this was to implement surfaceflinger which is a fast shared memory video server |
| 14:08.55 | tmzt | without all the event based protocol overhead |
| 14:09.20 | tmzt | and pixelflinger which is the non-gl version/implementation |
| 14:09.40 | cr2 | DRI ? |
| 14:09.51 | tmzt | for arm? |
| 14:10.17 | cr2 | why not. |
| 14:10.33 | tmzt | I would like to see wayland and dri on arm but it's not here yet |
| 14:10.47 | cr2 | a 256MB+ machine like raph is powerful enough. |
| 14:10.57 | tmzt | hopefully kms will standardize things on the desktop a bit |
| 14:11.08 | tmzt | and it will makes it way to handhelds |
| 14:11.22 | tmzt | since we don't have such evil as vbe |
| 14:11.29 | tmzt | and bios calls |
| 14:11.35 | cr2 | hehe. lol |
| 14:12.25 | tmzt | you think mdp can do fast blits without gl stuff? |
| 14:12.27 | cr2 | the idea (video)bios=absolute_evil came into my mind in 1995. |
| 14:12.49 | cr2 | when i have first tried booting linux on alpha ;) |
| 14:12.58 | cr2 | i think so |
| 14:13.05 | tmzt | with pci video card? |
| 14:13.09 | cr2 | it's limited only by the mddi DMA speed |
| 14:13.12 | cr2 | yes |
| 14:13.39 | tmzt | which is fast |
| 14:13.59 | Dunedan | ok, but back to my problem: how difficult will it be to implement .fb_blank? |
| 14:14.13 | tmzt | return 0 |
| 14:14.25 | cr2 | yes, for now. |
| 14:14.40 | cr2 | you can also check other missing entries. |
| 14:14.54 | cr2 | sa1100fb was working with X very well. |
| 14:15.20 | cr2 | hehe. i've even enjoyed twm on 12MB-linux jornada820 :) |
| 14:15.33 | tmzt | and s1dxxxx is very slow, but kristoffer is working on a driver for it |
| 14:15.43 | tmzt | abiword is unusable |
| 14:15.46 | tmzt | hey |
| 14:15.58 | tmzt | j820 is s1d? |
| 14:16.06 | cr2 | 820 has sa1100fb |
| 14:16.16 | cr2 | in a funny config in wince. |
| 14:16.23 | tmzt | oh, ok |
| 14:16.36 | tmzt | pxafb is getting overlay support also |
| 14:16.36 | cr2 | because sa1100 has two separate video DMA channels |
| 14:16.44 | tmzt | right |
| 14:16.56 | cr2 | so wince splits the screen into 2 halves |
| 14:16.57 | tmzt | I mean I think pxa does too |
| 14:17.09 | cr2 | i think sa1100fb does not support it. |
| 14:17.26 | cr2 | and uses only the first dma channel |
| 14:17.41 | cr2 | which loads the in-cpu bus of course. |
| 14:18.38 | cr2 | tmzt: but i have a 16MB RAM extension board, so with 32MB things look much better. |
| 14:18.57 | cr2 | hehe, 2.6.30 on j820 would have been nice... |
| 14:19.21 | tmzt | #jlime is supporting those |
| 14:19.22 | cr2 | it even has usb host, and 1024x768 video out. |
| 14:19.27 | cr2 | 820 ? |
| 14:19.36 | tmzt | with upstream support soon |
| 14:19.38 | tmzt | not sure |
| 14:19.45 | cr2 | 720 i think |
| 14:19.51 | cr2 | but i'll check. |
| 14:19.52 | tmzt | what chipset other than sa does it use? |
| 14:20.13 | cr2 | sa1101 |
| 14:20.35 | tmzt | and wmxxx? |
| 14:20.35 | cr2 | not sa1111 |
| 14:20.52 | cr2 | no, the philips something |
| 14:21.05 | cr2 | http://jornada820.sf.net |
| 14:21.52 | cr2 | it has pcmcia and cf slots. |
| 14:22.05 | tmzt | oh, your not there |
| 14:22.22 | tmzt | those should be same as sl5500 |
| 14:23.49 | cr2 | hm. i have sl5500 somewhere too :) |
| 14:26.46 | cr2 | the kernel should be in hh.org |
| 14:27.42 | tmzt | why would it be there? |
| 14:29.21 | cr2 | because it started in 2002 afair |
| 14:29.57 | cr2 | Dunedan: ok, so what .fb* functions are missing that are included in sa1100fb ? |
| 14:30.08 | tmzt | cr2, can you join #jlime? |
| 14:30.47 | tmzt | but it's lineo, at least 2.4 |
| 14:31.38 | cr2 | the 2.6 kernel patches have a sa1101 implementation bug |
| 14:31.55 | cr2 | but in general sa1101 is much better (=less buggy) than sa1111 |
| 14:32.16 | cr2 | intel took sa1101 from dec and fscked things up. |
| 14:32.42 | cr2 | tmzt: joined |
| 14:32.58 | Dunedan | cr2: there are .fb_set_par, .fb_setcolreg, .fb_blank and .fb_mmap which are not in msmfb_ops and .fb_open, .fb_release, .fb_pan_display and .fb_ioctl which are not in sa1100fb_ops |
| 14:33.29 | *** join/#htc-linux fnord__ (n=fnord@24-151-90-116.static.nwtn.ct.charter.com) |
| 14:33.36 | cr2 | .fb_set_par may be useful |
| 14:33.57 | cr2 | .fb_blank too. .fb_mmap should be checked. |
| 14:34.07 | cr2 | .fb_mmap is cool |
| 14:35.22 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 14:42.31 | cr2 | Dunedan: does your X work with dummy .fb_blank ? |
| 14:42.47 | Dunedan | cr2: currently rebooting the blackstone to try it out ;-) |
| 14:42.57 | cr2 | if the .fb_mmap will work, we certainly want to have it. |
| 14:43.01 | cr2 | :) |
| 14:46.07 | *** join/#htc-linux MethoS (n=clemens@dyndsl-085-016-165-162.ewe-ip-backbone.de) |
| 14:48.46 | Dunedan | hm, doesn't work |
| 14:49.26 | cr2 | Dunedan: error message ? |
| 14:49.54 | tmzt | added to fbops or fb-ioctl? |
| 14:50.04 | Dunedan | fbops |
| 14:50.07 | *** join/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 14:51.05 | Dunedan | Xorg.0.log looks fine now. There are just a few "(EE) FBDEV(0): FBIOPUTCMAP: Invalid argument" in the middle of the log |
| 14:51.43 | Dunedan | http://dpaste.com/35513/ <- current strace output |
| 14:52.40 | tmzt | so that wasn't it |
| 14:52.58 | tmzt | it didn't stop because of fbioblank |
| 14:55.36 | cr2 | .fb_setcolreg probably |
| 14:55.58 | tmzt | which is? |
| 15:04.18 | cr2 | does msmfb support setting the palette ? |
| 15:04.39 | cr2 | for 16/18bit it's not really useful |
| 15:05.42 | cr2 | the missing keyboard and mouse=pointer now ? |
| 15:05.59 | tmzt | what? |
| 15:06.24 | tmzt | they should be automatic |
| 15:06.30 | cr2 | there is no ts support ? |
| 15:06.51 | cr2 | write(0, "(II) GLX: Initialized DRISWRAST G"..., 57) = 57 |
| 15:06.53 | tmzt | there is a xorg-input-tslib package available |
| 15:06.58 | cr2 | i will switch GLX off |
| 15:07.05 | tmzt | yes |
| 15:07.39 | cr2 | xinerama too ? |
| 15:08.06 | cr2 | EE) FBDEV(0): FBIOPUTCMAP: Inval |
| 15:08.09 | tmzt | what packages are you using? |
| 15:08.10 | cr2 | a lot of them. |
| 15:08.29 | cr2 | but why does it want to set a cmap on a 16bit device ?? |
| 15:08.51 | tmzt | I don't know |
| 15:09.03 | tmzt | it should be TrueColor right? |
| 15:09.07 | cr2 | ioctl(7, FBIOGET_VSCREENINFO, 0x1bee48) = 0 |
| 15:09.09 | cr2 | ioctl(7, FBIOBLANK, 0x1) = 0 |
| 15:09.14 | cr2 | the blank is ok now. |
| 15:09.15 | cr2 | yes |
| 15:09.48 | cr2 | ioctl(7, FBIOPUTCMAP, 0xbef7a468) = -1 EINVAL (Invalid argument) |
| 15:10.00 | cr2 | add a dummy here too ? |
| 15:10.41 | Dunedan | Shall I try it? |
| 15:11.15 | cr2 | it's strange, but why not. |
| 15:12.11 | tmzt | is it failing to get a screen? |
| 15:12.29 | cr2 | the w100fb has some cmap support |
| 15:13.01 | cr2 | fb_alloc_cmap(&info->cmap, 256, 0) |
| 15:13.27 | cr2 | tmzt: it's probably for the 8bit visual |
| 15:14.05 | cr2 | called in _probe btw |
| 15:14.33 | tmzt | right |
| 15:14.35 | tmzt | ok |
| 15:15.02 | tmzt | this be mode enum and xrandr1.1 if new enough Xorg |
| 15:15.14 | tmzt | but not relevant |
| 15:15.45 | tmzt | Dunedan: what are the last messages in X.log.0? |
| 15:16.01 | cr2 | check static int w100fb_setcolreg |
| 15:16.43 | Dunedan | (II) Initializing built-in extension RANDR |
| 15:16.45 | Dunedan | (II) Initializing built-in extension COMPOSITE |
| 15:16.46 | Dunedan | (II) Initializing built-in extension DAMAGE |
| 15:16.47 | Dunedan | SELinux: Disabled on system, not enabling in X server |
| 15:16.49 | Dunedan | (II) AIGLX: Screen 0 is not DRI2 capable |
| 15:16.51 | Dunedan | (II) AIGLX: Screen 0 is not DRI capable |
| 15:16.55 | tmzt | last |
| 15:16.56 | Dunedan | (II) AIGLX: Loaded and initialized /usr/lib/dri/swrast_dri.so |
| 15:16.59 | tmzt | messages |
| 15:16.59 | Dunedan | (II) GLX: Initialized DRISWRAST GL provider for screen 0 |
| 15:17.18 | cr2 | info->pseudo_palette = kmalloc(sizeof (u32) * MAX_PALETTES, GFP_KERNEL); |
| 15:18.15 | cr2 | Dunedan: i'd just copypaste the w100 code |
| 15:18.49 | tmzt | Xorg doesn't care |
| 15:19.11 | tmzt | the question is whether it created the screen successfuly or not |
| 15:20.03 | Dunedan | tmzt: http://dpaste.com/35523/ <- complete log |
| 15:22.20 | tmzt | it looks fine |
| 15:22.42 | tmzt | is is crashing? |
| 15:22.49 | Dunedan | yes |
| 15:23.06 | cr2 | the putcmap is an error, though |
| 15:23.18 | tmzt | ok, what is the last thing in strace |
| 15:23.20 | cr2 | Dunedan: create a dummy ? |
| 15:23.24 | tmzt | for that mode |
| 15:25.46 | cr2 | truecolor is default, so it should not be a problem |
| 15:28.15 | Dunedan | erm |
| 15:28.28 | Dunedan | Do we have any driver available which implements putcmap? |
| 15:28.38 | cr2 | w100 |
| 15:28.51 | cr2 | but create a dummy for now |
| 15:29.14 | tmzt | that might confuse it more though |
| 15:29.24 | cr2 | but it will not crash |
| 15:29.30 | tmzt | if it things cmap is supported |
| 15:29.40 | tmzt | where is it crashing? |
| 15:29.57 | Dunedan | static struct fb_ops w100fb_ops = { |
| 15:29.59 | Dunedan | <PROTECTED> |
| 15:30.00 | Dunedan | <PROTECTED> |
| 15:30.02 | Dunedan | <PROTECTED> |
| 15:30.03 | Dunedan | <PROTECTED> |
| 15:30.05 | Dunedan | <PROTECTED> |
| 15:30.08 | Dunedan | <PROTECTED> |
| 15:30.10 | Dunedan | <PROTECTED> |
| 15:30.13 | Dunedan | <PROTECTED> |
| 15:30.16 | Dunedan | <PROTECTED> |
| 15:30.18 | Dunedan | }; |
| 15:30.20 | Dunedan | hm |
| 15:30.23 | Dunedan | sry, should use dpaste |
| 15:30.23 | cr2 | .fb_setcolreg = w100fb_setcolreg, |
| 15:30.26 | Dunedan | shouldn't be putcmap there? |
| 15:30.31 | Dunedan | ah ok |
| 15:34.20 | Dunedan | compiling now once again |
| 15:34.46 | cr2 | does blac100 support serial console ? |
| 15:34.55 | cr2 | hm, maybe i'd ask myself. |
| 15:35.12 | *** part/#htc-linux Moult (n=Moult@124.13.233.148) |
| 15:35.18 | cr2 | but i has the gpio table rearranged. as it's usual by htc |
| 15:38.10 | cr2 | tmzt: can we mmap the msm framebuffer ? |
| 15:38.22 | tmzt | through pmem? |
| 15:38.36 | tmzt | I think either gpu0 or gpu1 is framebuffer |
| 15:38.44 | tmzt | the other is registers |
| 15:44.12 | cr2 | i think gpu0 and gpu1 are redundant |
| 15:44.38 | tmzt | check pmem.c? |
| 15:44.42 | cr2 | the fbmem is a separate area |
| 15:45.47 | *** part/#htc-linux Echo31 (n=olivier@mir31-4-82-240-194-54.fbx.proxad.net) |
| 15:45.59 | cr2 | #define MSM_FB_BASE MSM_PMEM_GPU1_BASE + MSM_PMEM_GPU1_SIZE |
| 15:46.01 | cr2 | 35 #define MSM_FB_SIZE 0x200000 /* 2mb */ |
| 15:46.48 | cr2 | where is pmem.c ? |
| 15:47.07 | tmzt | drivers/android or mach-msm |
| 15:47.22 | tmzt | that's where I got thatfrom I think |
| 15:47.36 | tmzt | oh |
| 15:47.47 | tmzt | so fb is after gpu1 |
| 15:47.51 | tmzt | not the same |
| 15:47.55 | cr2 | static struct resource resources_msm_fb[] = { |
| 15:47.56 | cr2 | 551 { |
| 15:47.58 | cr2 | 552 .start = MSM_FB_BASE, |
| 15:47.59 | cr2 | 553 .end = MSM_FB_BASE + MSM_FB_SIZE, |
| 15:48.01 | cr2 | 554 .flags = IORESOURCE_MEM, |
| 15:48.02 | cr2 | 555 }, |
| 15:48.04 | cr2 | 55 |
| 15:48.49 | cr2 | the hw3d grp registers are at 0xa0000000 |
| 15:49.01 | cr2 | but it's this 3d stuff i think. |
| 15:49.09 | tmzt | yeah |
| 15:49.16 | tmzt | find pmem? |
| 15:49.45 | cr2 | not in the linuxtogo git viewer |
| 15:50.24 | tmzt | not in htc-msm-2.6.25? |
| 15:50.26 | cr2 | struct pmem_region region = {.offset = 0x0, .len = HW3D_REGS_LEN}; |
| 15:51.31 | tmzt | drivers/misc |
| 15:52.29 | cr2 | misc ;) |
| 15:52.39 | tmzt | yeah, like mem |
| 15:52.47 | tmzt | but we want the board trout |
| 15:53.40 | cr2 | <PROTECTED> |
| 15:54.09 | cr2 | but i'm worried that it's needed for adsp and (maybe) mdp |
| 15:54.56 | tmzt | 8m |
| 15:55.11 | tmzt | 8mb |
| 15:55.39 | tmzt | gpu1 size |
| 15:55.48 | tmzt | definitely not registers |
| 15:56.49 | cr2 | the fbmem can be probably kmalloc'd |
| 15:57.12 | tmzt | not fixed? |
| 15:57.14 | cr2 | is it possible to kmalloc 2MB ? |
| 15:57.42 | cr2 | wince uses a different locaton. and g1 anyway. |
| 15:57.56 | tmzt | other fb driver kmalloc, don't know about size limits |
| 15:58.25 | cr2 | i think 128K was a limit. but maybe i'm an oldtimer :) |
| 15:58.27 | tmzt | is it just so it's in sram? |
| 15:58.38 | cr2 | maybe |
| 15:58.50 | cr2 | but the current code has it in sdram anyway. |
| 15:59.16 | tmzt | why? |
| 15:59.28 | tmzt | current google code?? |
| 15:59.35 | cr2 | because NetRipper started from halibut |
| 15:59.43 | tmzt | oh |
| 15:59.58 | cr2 | and the g1 code has it in the second sram bank -> does not exist on raph100 |
| 16:00.21 | cr2 | you can still see some stray halibut references in the code |
| 16:00.30 | tmzt | I know |
| 16:00.59 | tmzt | raph100 has less sram then g1? |
| 16:01.08 | cr2 | for non-android we may kmalloc, or put it into sram. |
| 16:01.17 | tmzt | I thought the only differences were rf |
| 16:01.19 | cr2 | yes. 64 vs 32 |
| 16:02.08 | tmzt | how much is available on raph100 then? |
| 16:02.32 | tmzt | gpu memory is in sdram also? |
| 16:02.37 | cr2 | 32 -spl -amss-smem-oemsbl-qcsbl |
| 16:02.49 | cr2 | NetRipper split it |
| 16:03.03 | cr2 | gpu0 is in sram, gpu1 and fbmem are in sdram |
| 16:03.24 | tmzt | ok |
| 16:03.26 | cr2 | i guess for non-android we may drop gpu0 and gpu1 and put fbmem into sram |
| 16:03.50 | tmzt | unless gpu0 is fixed |
| 16:03.59 | tmzt | if it's the registers |
| 16:04.02 | cr2 | freeing 17MB, heh |
| 16:04.26 | *** join/#htc-linux dcordes (n=dcordes_@unaffiliated/dcordes) |
| 16:04.31 | tmzt | I don't think it really has to be that much |
| 16:04.43 | cr2 | 18MB |
| 16:05.00 | tmzt | gpu1 I mean |
| 16:05.16 | tmzt | it could have been dynamic |
| 16:05.24 | cr2 | msm_fb does not use it, so it may go away. |
| 16:05.37 | tmzt | yeah |
| 16:05.58 | *** join/#htc-linux BHSPitLappy (n=BHSPitLa@unaffiliated/bhspitmonkey) |
| 16:06.01 | tmzt | unless we can mmiotrace the 2d primatives |
| 16:06.06 | Dunedan | xorg still fails :-/ |
| 16:06.28 | cr2 | tmzt: you can't mmtrace dm |
| 16:07.37 | tmzt | what are the last few lines of strace? |
| 16:07.37 | tmzt | I mean the gpu0 writes |
| 16:07.37 | tmzt | or a wrapper for libhgl or debugging in pmem driver (possibly) |
| 16:07.40 | Dunedan | http://dpaste.com/35552/ <- current xorg log |
| 16:08.00 | Dunedan | http://dpaste.com/35553/ <- current strace output |
| 16:08.15 | tmzt | I mean on linux not ce/haret |
| 16:10.32 | tmzt | segfault |
| 16:10.37 | tmzt | gdb it |
| 16:10.47 | tmzt | it's not ioctls |
| 16:31.17 | Dunedan | hm, don't know how to trace it. gdb doesn't print out something useful, even with xorg debug packages installed. |
| 16:33.45 | tmzt | not even a partial bt? |
| 16:33.58 | tmzt | try bt |
| 16:36.19 | Dunedan | after the crash I couldn't get a bt |
| 16:37.20 | tmzt | but you were running ingdb? |
| 16:37.21 | Dunedan | Currently I just do "gdb Xorg" and then inside gdb "run". Is this the right proceeding? |
| 16:37.24 | *** join/#htc-linux timebomb (n=tb@c211099.adsl.hansenet.de) |
| 16:37.29 | tmzt | yes |
| 16:38.36 | Dunedan | http://dpaste.com/35563/ |
| 16:41.32 | tmzt | something is wrong |
| 16:41.58 | tmzt | check the options for Xorg, there should be a few that will help here |
| 17:05.09 | *** join/#htc-linux hollo (n=hollo@3e6b7b2c.rev.stofanet.dk) |
| 17:16.01 | *** join/#htc-linux BoSJo (n=bo@164.84-48-222.nextgentel.com) |
| 17:51.37 | *** join/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 17:52.12 | *** part/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 18:01.41 | *** join/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 18:04.16 | *** part/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 18:23.56 | *** join/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 18:26.48 | *** part/#htc-linux exception13 (n=exceptio@testdrive.kgts.ru) |
| 18:26.54 | *** join/#htc-linux tsdogs (n=tsdogs@net203-187-146.mclink.it) |
| 18:29.42 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 18:33.58 | *** join/#htc-linux XD (i=t3st1fy@ircop.com) |
| 18:34.28 | *** join/#htc-linux _XD (i=t3st1fy@ircop.com) |
| 18:59.35 | *** join/#htc-linux spenguin[work] (n=penguin@203.199.114.33) |
| 19:21.56 | *** join/#htc-linux zycho (n=zycho@dslb-088-070-254-107.pools.arcor-ip.net) |
| 19:25.34 | cr2 | kiozen: are you here ? |
| 19:26.25 | kiozen | jep |
| 19:26.59 | kiozen | cr2: |
| 19:27.21 | cr2 | new M and rootfs |
| 19:27.55 | cr2 | we can't run X on raph, maybe q(topia) will work better ? |
| 19:28.46 | kiozen | you mean via framebuffer? |
| 19:28.50 | cr2 | yes |
| 19:29.17 | kiozen | you tried old fs? |
| 19:29.44 | cr2 | hmm. i have it on the usb stick only |
| 19:30.35 | cr2 | drivers/video/msm/mddi_client_epson.c:99: warning: ISO C90 forbids mixed declarations and code |
| 19:32.50 | cr2 | kiozen: qlandkarte.org does not have favicon :) |
| 19:33.00 | kiozen | cr2: it's still on the server http://www.qlandkarte.org/rootfs.tar.bz2 |
| 19:33.06 | kiozen | cr2: who cares |
| 19:35.14 | cr2 | ok |
| 19:35.32 | cr2 | what is the debian name for gdal and proj4 ? |
| 19:35.47 | cr2 | i'll compile GT on nc10 |
| 19:38.02 | cr2 | hm. already installed |
| 19:45.23 | cr2 | cmpiling |
| 19:51.51 | cr2 | produces weird messages at startup |
| 19:51.55 | cr2 | ------------ +proj=tmerc +lat_0=0 +lon_0=12 +k=1 +x_0=4500000 +y_0=0 +ellps=bessel +datum=potsdam +units=m +no_defs +towgs84=606.0,23.0,413.0 |
| 20:05.41 | *** join/#htc-linux j0b0 (n=jobo@5ED40048.cable.ziggo.nl) |
| 20:09.06 | *** join/#htc-linux AstainHellbring (n=AstainHe@unaffiliated/astainhellbring) |
| 20:15.17 | *** join/#htc-linux marex (n=marex@vasut.kolej.mff.cuni.cz) |
| 20:18.11 | cr2 | hi marex |
| 20:19.10 | *** join/#htc-linux timebomb (n=tb@c211099.adsl.hansenet.de) |
| 20:22.51 | marex | cr2, hi, sorry I havent had time |
| 20:23.34 | marex | cr2, if someone feels like testing my "method", he can try in qemu |
| 20:26.04 | cr2 | marex: ok, write a patch for .suspend, it's not going to hurt anyway, because we do not support mdoc |
| 20:26.33 | marex | cr2, I tried, the LCD doesnt kick in |
| 20:26.52 | marex | well, is there any GPIO I can switch and see some effect ? |
| 20:27.02 | marex | you see, I'd like to test if the resume code is even executed |
| 20:27.36 | cr2 | hmm, it's not a phone, so there is no vibra. the LED control is complex. |
| 20:28.50 | marex | cr2, yes |
| 20:29.01 | marex | there's a very valid problem |
| 20:29.28 | marex | also, the LCD is somehow weird with the SPI interface |
| 20:29.57 | cr2 | the pwm1 is not handled properly |
| 20:30.29 | cr2 | and the spi bitbang needs spinlocks |
| 20:32.12 | marex | you wanna say it basically sucks |
| 20:32.25 | marex | and you have no idea what will happen once it resumes |
| 20:33.05 | cr2 | marex: you can pull the blue keypad gpio |
| 20:33.23 | marex | isnt it on the CPLD too ? |
| 20:33.32 | cr2 | yes, but it does not matter |
| 20:33.55 | marex | well during suspend, it's shut down (the backlight) |
| 20:34.04 | cr2 | just write directly, the mmu is off |
| 20:34.17 | marex | cr2, write what and where ? |
| 20:34.23 | cr2 | i mean at resume |
| 20:34.38 | cr2 | 0xa0100000 |
| 20:34.39 | marex | cr2, are you sure the CPLD is enabled at that time ? |
| 20:34.48 | marex | cr2, huh ? |
| 20:35.08 | marex | ok ... once again ... write what to which address ? |
| 20:35.10 | cr2 | you can always write there. the cpld does not need some init. |
| 20:35.15 | marex | I know 0xa0100000 is wakeup address |
| 20:35.16 | cr2 | hmm. wiki |
| 20:35.24 | cr2 | wakeup is 0x0 |
| 20:35.27 | marex | hmm ... later ;) |
| 20:35.33 | cr2 | ok |
| 20:35.38 | marex | cr2, if the DoCG4 is dead, it jumps into RAM |
| 20:35.42 | marex | to that address |
| 20:36.43 | cr2 | yes. |
| 20:37.04 | cr2 | so put the strh to gpio there |
| 20:37.18 | marex | well and I'd like to put some indication it actually is there |
| 20:38.08 | marex | I'll write some assembler code and put it there, I just need to know what exactly and where should I store in the code to get some effect |
| 20:38.17 | marex | I'll see the HW wiki probably later then |
| 20:39.15 | cr2 | GPIOD0 0x0001 kbd backlight(on=1) ? |
| 20:39.33 | marex | what's that question mark doing there? |
| 20:39.36 | cr2 | it must be in the source |
| 20:39.51 | cr2 | is it the blue kbd led ? |
| 20:39.54 | marex | I used arm:devel (mainline kernel) last time I hacked on N560 |
| 20:40.10 | cr2 | i'll look at hh.org kernel |
| 20:40.13 | marex | cr2, just as I thought ... it's uncertain ;) |
| 20:40.28 | cr2 | *Dpin +0x06 |
| 20:40.29 | *** join/#htc-linux diogene31 (n=rj@mur31-2-82-243-122-54.fbx.proxad.net) |
| 20:40.31 | marex | cr2, I'll look into it later as I already said |
| 20:40.45 | cr2 | since i've added it in hh.og, it's certain |
| 20:41.11 | cr2 | 0x10000000 1 CPLD |
| 20:41.53 | cr2 | ldrh [0x10000006] |
| 20:42.04 | cr2 | [0x10000006] | 1 |
| 20:42.20 | cr2 | strh [0x10000006] |
| 20:42.26 | cr2 | should not be too difficult |
| 20:43.13 | marex | mov r0, #0x10000000 ; mov r1, #0x1 ; str r1, [r0, #0x06] |
| 20:44.27 | *** join/#htc-linux swc|666 (n=swc@unaffiliated/swc666/x-4934821) |
| 20:46.42 | *** join/#htc-linux FuzzyPolarBear (n=Polar@fuzzy.msi.vxu.se) |
| 20:52.25 | swc|666 | <Dunedan> well, I don't want android. I just want pure debian. ;-) ... roger that :>] |
| 21:19.37 | *** join/#htc-linux AstainHellbring (n=AstainHe@unaffiliated/astainhellbring) |
| 21:19.44 | *** join/#htc-linux MethoS- (n=clemens@host-091-096-213-231.ewe-ip-backbone.de) |
| 21:25.45 | AstainHellbring | anyone with windows and pgp/gpg encryption expirence? |
| 21:26.34 | cr2 | swc|666: lol. anti-android team here :) |
| 21:27.23 | swc|666 | :D |
| 21:28.08 | cr2 | AstainHellbring: the commandline should not differ |
| 21:28.42 | AstainHellbring | cr2 I'm doing email encryption stuff... |
| 21:30.32 | cr2 | swc|666: at least we can free 18MB more |
| 21:30.51 | cr2 | swc|666: and use the second128M bank. |
| 21:31.40 | AstainHellbring | cr2 whats the latest progress on data and cell stuffs in the kernel? |
| 21:32.33 | cr2 | it breaks the ppp connection |
| 21:32.40 | cr2 | i have not tried it myself. |
| 21:34.46 | swc|666 | cr2, awesome! |
| 21:35.57 | tmzt | AstainHellbring: thunderbird and enigmail |
| 21:58.03 | *** join/#htc-linux Amynecka (n=bedrunec@hellmachine.klfree.cz) [NETSPLIT VICTIM] |
| 21:58.03 | *** join/#htc-linux madCoder- (n=madcoder@c-71-225-238-170.hsd1.pa.comcast.net) [NETSPLIT VICTIM] |
| 21:58.03 | *** join/#htc-linux methril|fix_part (n=Methril@213.27.233.98) |
| 21:58.03 | *** join/#htc-linux bodhiand (n=bodhi@89.121.200.106) [NETSPLIT VICTIM] |
| 22:49.43 | *** join/#htc-linux syn-fin (n=alien@c-75-74-251-191.hsd1.fl.comcast.net) |
| 22:59.26 | *** join/#htc-linux lpotter (n=lpotter@218.185.108.156) |
| 23:03.47 | *** join/#htc-linux wirelessdreamer (n=dreamer@chrobd01.vailsys.com) |
| 23:04.11 | *** join/#htc-linux timebomb (n=tb@c211099.adsl.hansenet.de) |
| 23:04.23 | *** join/#htc-linux syn-fin1 (n=alien@c-75-74-251-191.hsd1.fl.comcast.net) |
| 23:58.35 | *** join/#htc-linux swc|666 (n=swc@unaffiliated/swc666/x-4934821) |