00:00.44 | rolk | pdev->id seems to be 1,2,3 or 4 in all legal cases. |
00:01.03 | rolk | And msmsdcc_pclks is defined as { NULL, "sdc1_pclk", "sdc2_pclk", "sdc3_pclk", |
00:01.03 | rolk | <PROTECTED> |
00:01.15 | dwaradzyn | rolk: which line is it in msm_sdcc.c? (cant find pc_clk string in file) |
00:01.19 | dzo | pclk is the peripheral bus clock and I don't think we have control of that, i think i just commented out the code that set it, i'll just check |
00:01.24 | dcordes | dwaradzyn, how would a function look that only runs msm_update_screen when kaiser cam button is pressed? |
00:01.32 | rolk | So, for all legal cases, the clocks are identified as "sdc1_pclk", "sdc2_pclk", "sdc3_pclk", |
00:01.32 | rolk | <PROTECTED> |
00:01.36 | dcordes | dwaradzyn, do I use the keypad driver or use the gpio directly? |
00:02.22 | dcordes | rolk, do you know which code dzo uses on the vogue? |
00:02.33 | dcordes | it's the htc-voge branch on ltg |
00:03.23 | dzo | my pc_clk_get_rate looks like: static unsigned pc_clk_get_rate(unsigned id) |
00:03.25 | dzo | { |
00:03.25 | dcordes | xperia usb woks. |
00:03.26 | dzo | <PROTECTED> |
00:03.28 | dzo | <PROTECTED> |
00:03.29 | dzo | <PROTECTED> |
00:03.31 | dzo | <PROTECTED> |
00:03.32 | dzo | <PROTECTED> |
00:03.35 | dzo | <PROTECTED> |
00:03.35 | dzo | } |
00:04.01 | *** join/#htc-linux Shinto (n=John@g227183018.adsl.alicedsl.de) |
00:04.28 | rolk | The pc_clk_get_rate is somewhat similar: |
00:04.29 | rolk | static unsigned pc_clk_get_rate(unsigned id) |
00:04.30 | rolk | { |
00:04.30 | rolk | if(id==SDC1_CLK) |
00:04.30 | rolk | return get_sdcc_host_clock(SDC1_CLK); |
00:04.30 | rolk | if(id==SDC1_PCLK) |
00:04.31 | rolk | return get_sdcc_host_clock(SDC1_PCLK); |
00:04.33 | rolk | if(id==SDC2_CLK) |
00:04.39 | rolk | return get_sdcc_host_clock(SDC2_CLK); |
00:04.41 | rolk | if(id==SDC2_PCLK) |
00:04.42 | rolk | return get_sdcc_host_clock(SDC2_PCLK); |
00:04.43 | rolk | printk("Unknown clock!\n"); |
00:04.45 | rolk | return -EINVAL; |
00:04.47 | rolk | } |
00:05.10 | rolk | But the mmc driver gets an -EINVAL back, so it must be the case that the id it provides is not equal to any of the listed CLKs. |
00:05.56 | dzo | or it could be the return value from get_sdcc_host_clock |
00:08.06 | rolk | Yes, true. |
00:08.53 | rolk | unsigned int get_sdcc_host_clock(unsigned id) { |
00:08.53 | rolk | int n; |
00:08.53 | rolk | int index; |
00:08.53 | rolk | unsigned val0; |
00:08.54 | rolk | unsigned val1; |
00:08.54 | rolk | if(sdcc_host_clock!=0) |
00:08.56 | rolk | return sdcc_host_clock; |
00:08.58 | rolk | index=clock_to_index(id); |
00:09.00 | rolk | val0=readl(MSM_CLK_CTL_BASE + sd_clock_regs[index].offset); |
00:09.02 | rolk | val1=readl(MSM_CLK_CTL_BASE + sd_clock_regs[index].offset+4); |
00:09.04 | rolk | |
00:09.06 | rolk | for(n=0;n<ARRAY_SIZE(sd_clock_parameters);n++) |
00:09.10 | rolk | if (sd_clock_parameters[index][n].val0==val0 && |
00:09.12 | rolk | <PROTECTED> |
00:09.14 | rolk | <PROTECTED> |
00:09.17 | rolk | } |
00:09.19 | rolk | return -EINVAL; |
00:09.25 | rolk | } |
00:09.47 | *** join/#htc-linux camden (n=camden@136-9-178-69.static.gci.net) |
00:10.05 | camden | hello |
00:10.44 | rolk | Could it be that the id is actually SDC3_CLK or SDC4_CLK, and then int clock_to_index(unsigned id) |
00:10.44 | rolk | { |
00:10.44 | rolk | <PROTECTED> |
00:10.44 | rolk | <PROTECTED> |
00:10.44 | rolk | <PROTECTED> |
00:10.44 | rolk | <PROTECTED> |
00:10.46 | rolk | <PROTECTED> |
00:10.48 | rolk | } |
00:10.52 | dzo | i see, clock to index returns 0 for an id of SDC1_PCLK |
00:11.40 | rolk | I should put some more diagnostics there. But its 1:11 now, and I need to work tomorrow. |
00:11.45 | dzo | since we can't get or set the PCLK, try just returning 66MHz |
00:11.57 | rolk | What is your timezone dzo? |
00:12.43 | dcordes | also late here. good night |
00:12.57 | Bally3 | night dcordes |
00:13.30 | dwaradzyn | dcordes: see you today :) |
00:13.59 | Bally3 | so what do you think we need to do then dzo? are we on the right track you think? |
00:14.20 | rolk | I've patched the pc_get_clk_rate function. |
00:14.23 | *** join/#htc-linux kuhlivisj (n=kuhlivis@s5592d82c.adsl.wanadoo.nl) |
00:14.25 | dzo | its 1.15 pm on friday in NZ, we're always ahead of the world!! |
00:14.35 | Bally3 | lol |
00:14.38 | rolk | I will create one more image and then I'm off to sleep. |
00:14.56 | Bally3 | nice one rolk :) |
00:15.02 | dzo | ok, got to work now, bye. |
00:15.16 | imfloflo | night guys goo dwork |
00:15.19 | Bally3 | whens the best time to catch you dzo? |
00:15.30 | imfloflo | oh dzo is here |
00:15.32 | Bally3 | would you come back here tomorrow? |
00:15.35 | imfloflo | so i have a question |
00:15.49 | Bally3 | please.... |
00:15.51 | dzo | email me is best but i'm going away for 2 weeks on sunday. |
00:15.52 | imfloflo | is it possible to tell us how we can do with the Sd issue ? to test it |
00:16.13 | imfloflo | which files? |
00:16.13 | Bally3 | flo: rolk and dzo discussed it already |
00:16.15 | Bally3 | :P |
00:16.33 | imfloflo | ok i had too drunk so :-# |
00:16.36 | dzo | just trial and error and lots of thinking is usually the best way forward.... |
00:16.44 | Bally3 | thanks dzo.. would love to see some progress before you disappear on sunday |
00:16.47 | imfloflo | waiting report on polaris trhead |
00:16.54 | imfloflo | gn all |
00:16.59 | Bally3 | night flo |
00:17.10 | rolk | So, booting.... |
00:17.20 | Bally3 | but you think rolk is on the right track? |
00:17.22 | imfloflo | thks |
00:17.43 | Bally3 | the sd card has stunped everyone here.. |
00:17.45 | dzo | i'll still check my mail when i'm away but won't be able to do much devel. looks like you are getting on top of it anyway. |
00:18.13 | Bally3 | it would be nice to think so.. from what I get told this has been ongoing for months |
00:18.43 | Bally3 | if we get sd card boot then everyone else can get involved.. finally :) |
00:18.56 | dzo | one more idea: have you tried smaller values for the block size and max block count. |
00:19.21 | Bally3 | you mentioned this in your pm.. but it stumped most of everyone.. |
00:19.29 | rolk | Well, this certainly is part of the story, but not quite the grail yet. The PCLK value is now 66000000, as patched but still the MMC times out. |
00:19.33 | Bally3 | is that in the config file or the default,txt? |
00:19.41 | dwaradzyn | i'm not sure it is the right track, mine sd also shows 4294967274 hz, but it basically works. if it were that big error the symptom should be that it does not work at all |
00:20.19 | dzo | when does it time out, on the initial detect or when transferring data? |
00:20.24 | Bally3 | what about the block size and max block count dwaradzyn? |
00:20.27 | rolk | initial detect |
00:20.37 | dwaradzyn | dzo: i played with blocksize and it did not solve anything, but when i tested it i noticed that smaller blocksize will get better stability |
00:20.39 | dwaradzyn | slightly betther but noticable one |
00:20.56 | dzo | the initial detect doesn't even use dma and it is supposed to time out on some of the requests. |
00:21.18 | rolk | Yes, 3 or so. |
00:21.22 | dwaradzyn | dzo: it gets several timeouts during init |
00:21.28 | rolk | But not 10 in a row. |
00:21.57 | dwaradzyn | i have only seen 3 and 4 timeouts during tests |
00:22.19 | Bally3 | so whats causing the instability? |
00:22.45 | rolk | Guys. Its goodnight to you all. I'm off to sleep. I'll try to make some time available for this this weekend. |
00:23.08 | Bally3 | goodnight rolk... thanks for the efforts mate :) |
00:23.16 | dwaradzyn | rolk: count me in. i'll be here tomorrow about the same time as today |
00:23.57 | Bally3 | dzo.. did you have any instability issues when you were doing this? |
00:24.59 | dzo | yes, lots... have you got mdelay(250); in mmc_sd_init_card ? (in drivers/mmc/core/sd.c) |
00:25.37 | Bally3 | it seems the card is detected but its not stable enough.. where can I change the blocksize and max block count? (sorry not so techical as these guys.,. but good at project managing) :) |
00:25.59 | Bally3 | dwaradzyn:? |
00:26.17 | dwaradzyn | Bally3: about the block size and max blocks - these are constants defined in kernel source, i just played with different values (but only on blocksize) and it was not a answer to sd issue |
00:26.37 | dwaradzyn | dzo: we have mdelay(250) with comment from san there |
00:26.56 | dcordes | tmzt_, thought I should notify you about dwaradzyn having found a way to make msm_fb work with X |
00:27.43 | dzo | in msmsdcc_probe, mmc->max_blk_size= 512; and mmc->max_blk_count = 127; try smaller numbers. |
00:27.44 | Bally3 | did you get the kaiser then dzo? |
00:28.06 | dzo | it's on it's way but will arrive when i'm away |
00:28.14 | Bally3 | doh! |
00:29.08 | Bally3 | pity there isnt another way to get you what you need to help us |
00:29.26 | dzo | ok, i've really got to go now.... bye all. |
00:29.47 | dwaradzyn | dzo: bye, and happy holidays our hero :) |
00:29.48 | Bally3 | thanks for all your help dzo :) |
00:30.34 | Bally3 | and thats all she wrote.. as they say in the states :P |
00:30.36 | Bally3 | what do you think dwaradzyn? |
00:30.45 | Bally3 | do we have any more to go on? |
00:31.19 | Bally3 | I think you and rolk need to pair up... you seem to be the most head up on this |
00:31.42 | Bally3 | I'll get others involved.. bring them on here |
00:32.17 | Bally3 | I think the sd card issue will get fixed .. I'm confident of it.. damn if it wasnt for compiling I'd be at it right now :P |
00:32.38 | dwaradzyn | Bally3: we do not have concrete things, but this timing thing (pc_clk_get_rate) will be worth checking |
00:32.43 | Bally3 | but I need to get a decent linux environment set up first |
00:33.21 | dwaradzyn | Bally3: i can get you mine - it has everything there, and it is vmware image (about 5gb) |
00:33.33 | Bally3 | no we dont.. but this is the second outing with dzo and he is still pointing to timing and block sizes.. if that isnt it then we'll need a complete change of direction |
00:33.42 | Bally3 | GREAT |
00:33.47 | Bally3 | how do I ge tit? |
00:34.25 | dwaradzyn | i was thinking about posting it on rapidshare and putting links on xda |
00:34.37 | Bally3 | I need to learn linux anyway.. what I do know should be enough to get me started.. |
00:35.15 | Bally3 | that would be a great idea.. I'll post it on the first page along with the files we're working with.. and tell everyone who wants to get involved to use the same setup/files |
00:35.26 | dwaradzyn | lets put it that way - you can be beta tester, before releasing it on xda |
00:35.32 | dwaradzyn | lets go prv on details |
00:35.37 | Bally3 | there are lots of people eager to get involved.. they just dont know where to start |
00:35.44 | Bally3 | ok |
00:40.56 | tmzt_ | dcordes: hey, is everything moving to goog 2.6.27 and trout? do we know how to make that kernel/board work with minimal changes? |
00:41.59 | tmzt_ | cr2: if you are still here, if you have any information on asic3_mmc so we can get a working kernel for BA over the next few weeks??, can you paste it here |
00:43.55 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
00:46.59 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
00:49.38 | *** join/#htc-linux ionstorm (n=ion@ip68-228-225-247.ph.ph.cox.net) |
00:54.13 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
00:54.37 | Bally3 | night all |
00:58.24 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:03.16 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:08.29 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:11.42 | *** join/#htc-linux axe_hm (n=weechat@217.116.62.236) |
01:14.11 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:18.37 | *** join/#htc-linux Tinyboom (n=nahh@178.80-202-153.nextgentel.com) |
01:21.21 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:27.14 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:30.55 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:34.24 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:37.29 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:41.00 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:44.08 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:46.26 | infernixx | just fyi |
01:46.42 | infernixx | swetland mentioned that they got PIO working first for SD |
01:46.55 | infernixx | and it took quite some time after that to get DMA working |
01:47.46 | infernixx | im not up to speed on kaiser sd development but i hope PIO is the main focus |
01:47.57 | infernixx | since DMA is apparently a PITA on msm7k |
01:50.58 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:54.17 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
01:57.23 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
02:00.44 | *** join/#htc-linux kring (n=kring@211.155.128.98) |
02:00.46 | *** join/#htc-linux ilovethisplace (n=ilovethi@152.sub-70-198-251.myvzw.com) |
02:04.21 | *** join/#htc-linux BHSPitMonkey (n=stephen@unaffiliated/bhspitmonkey) |
03:06.27 | *** join/#htc-linux apt_ (i=ibot@rikers.org) |
03:06.27 | *** topic/#htc-linux is HTC Linux Channel: Find logs at http://apt.rikers.org/%23htc-linux/ | please check http://handhelds.org/moin/moin.cgi/HTC_2dPhones | http://wiki.xda-developers.com/index.php?pagename=Xanadux | <cr2> let's define a common setup. |
03:17.03 | *** join/#htc-linux axe|hm (n=weechat@217.116.62.236) |
03:18.44 | *** join/#htc-linux paulproteus (n=paulprot@2002:db69:2513:0:0:0:0:1) |
03:25.54 | *** join/#htc-linux paulproteus (n=paulprot@2002:db69:2513:0:0:0:0:1) |
03:25.54 | *** join/#htc-linux Tinyboom (n=nahh@178.80-202-153.nextgentel.com) [NETSPLIT VICTIM] |
03:25.54 | *** join/#htc-linux swetland (n=swetland@nat/google/x-ac2328226c12fbc1) |
03:25.55 | *** join/#htc-linux Poohba (n=poohba@c-71-58-20-66.hsd1.nj.comcast.net) [NETSPLIT VICTIM] |
03:25.55 | *** join/#htc-linux nizox (n=none@2a01:e35:8a13:a2b0:21c:c0ff:fe25:ff68) [NETSPLIT VICTIM] |
03:25.55 | *** join/#htc-linux toer (i=tore@179.81-166-86.customer.lyse.net) |
03:36.35 | *** join/#htc-linux axe|hm (n=weechat@217.116.62.236) |
03:42.32 | *** join/#htc-linux apt (n=ibot@rikers.org) |
03:42.32 | *** topic/#htc-linux is HTC Linux Channel: Find logs at http://apt.rikers.org/%23htc-linux/ | please check http://handhelds.org/moin/moin.cgi/HTC_2dPhones | http://wiki.xda-developers.com/index.php?pagename=Xanadux | <cr2> let's define a common setup. |
05:13.46 | tmzt_ | NetRipper: hello? are you in #android? |
05:13.55 | tmzt_ | hello, |
05:38.05 | *** join/#htc-linux goxboxlive (n=goxboxli@mail2.hjellnesconsult.no) |
05:43.59 | maejrep | <NetRipper> maejrep, can you pastebin your default.txt used for haret? <-- http://privatepaste.com/1bXeGkIukx |
05:45.27 | maejrep | I've tried the RAMSIZE line both commented and uncommented, and I've tried each of the INITRD lines. They all get to the same place and stop. If I comment out the RAMADDR line, it never gets past the haret "booting linux" progress bar in winmo |
05:47.00 | tmzt_ | maejrep: you have the raph100? |
05:47.07 | maejrep | raph800 |
05:47.09 | maejrep | cdma/sprint |
05:47.24 | tmzt_ | is that one supposed to be working? |
05:48.03 | maejrep | yes, hillsdale was able to get the mddi client working, which NetRipper included in one of his recent patches in git |
05:48.15 | maejrep | it used to at least boot to a console for me, and now I can't even get that far |
05:48.35 | maejrep | with touch of jobo's RGB444 change, it appears to fix the color issue |
05:48.53 | tmzt_ | any messages appear on the screen? you said it stops at booting linux |
05:49.04 | maejrep | no that's not the problem |
05:49.09 | maejrep | that only happens with the RAMADDR line |
05:49.30 | tmzt_ | ok, so there is not enough memory for your initrd? |
05:49.35 | maejrep | the problem I'm having is after booting past the TCP/ether stage, there are 4 something_pwrsink_set:STUB! lines |
05:49.38 | maejrep | and that's the last thing I see |
05:50.01 | maejrep | s/with/without/ |
05:50.58 | maejrep | the same initrd I'm using now, I've used in the past (say, about 1-2 weeks ago), and was able to boot to the console, calibrate the screen, and use the onscreen keyboard |
05:51.06 | tmzt_ | ok, panics if you use no initrd? |
05:51.10 | tmzt_ | ok, it |
05:51.56 | maejrep | but since I've updated to latest git, which includes NetRipper's version of hillsdale's original patch to get the framebuffer working, lavender.t's patch for usb_ether, and some other minor fixes |
05:52.04 | maejrep | it doesn't get past those 4 STUB lines |
05:52.11 | maejrep | I haven't tried it with no initrd .. is that valid? :x |
05:52.40 | maejrep | i mean, i know it's valid, but since there's no root fs to mount, I assumed it would fail |
05:52.42 | tmzt_ | well, it should panic but it's a standard message so it helps |
05:53.07 | maejrep | as in, it'll help determine if it's hanging loading the initrd or if it hangs before that? |
05:53.10 | maejrep | 1 second, |
05:53.35 | maejrep | http://www.netripper.com/raphael/20081202_dmesg.txt |
05:53.37 | tmzt_ | it's not clear whether you are getting to 'freeing init memory' which is when the kernel boot process ends and /sbin/init (or /init) is launched in userspace |
05:54.06 | maejrep | my boot looks more or less like that, except after the 4 "trout_pwrsink_set:STUB!" lines, it just stops |
05:54.20 | maejrep | yes, I see Freeing init memory: 100K |
05:54.37 | tmzt_ | and found initramfs before that? |
05:55.13 | maejrep | "Unpacking initramfs... done" ? |
05:55.30 | *** join/#htc-linux surgex0 (i=surge@pool-98-118-165-224.bflony.fios.verizon.net) |
05:55.41 | tmzt_ | ok, trout, I thought NetRipper was using halibut |
05:55.51 | maejrep | he's using trout for everything except the keyboard, afaik |
05:56.08 | tmzt_ | but sd also does not work on this kernel? |
05:56.31 | maejrep | well, I get similar messages to what's in NetRipper's dmesg, re the sd |
05:56.43 | maejrep | but haven't ever tried to use it |
05:56.56 | maejrep | 1 sec, I'll boot it now |
05:57.50 | maejrep | yeah, I see: [ 1.407873] mmc0: Qualcomm MSM SDCC at 0x00000000e1000000 irq 24,0 dma -1 |
05:57.56 | maejrep | up to "Polling status mode enabled" |
05:58.26 | tmzt_ | sorry, just trying to understand what NetRipper was saying yesterday |
05:58.33 | maejrep | and, in fact, I don't see "Freeing init memory" |
05:59.03 | maejrep | I'll copy NetRipper's dmesg and point out what lines of his I also see |
06:01.38 | *** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
06:03.51 | maejrep | http://privatepaste.com/a3f1WyIOQW |
06:04.09 | maejrep | that's essentially what I see on the screen when it freezes |
06:04.40 | maejrep | it gets to line 39, then pauses for about 9 seconds, prints 3 more lines, pauses for 10 seconds, then the rest prints out |
06:05.18 | maejrep | at line 57, where I marked with ***, that line is mostly cut off, where I can only see the top half of each letter, basically |
06:05.36 | maejrep | then the last 4 STUB lines, and a new line with just a "_" printed on the left |
06:06.42 | maejrep | netripper's dmesg does not have the 10-second pauses -- at most it's like 1 second, and his gets all the way through calibration in about 7 seconds |
06:08.14 | maejrep | re the cutoff at line 57, it's possible that the "Freeing init memory" line is *supposed* to be printed there, under the hctosys.c rtc line but the display glitch causes it to just not show up.. I don't know for sure though |
06:13.27 | tmzt_ | you might be right if something is reading /dev/rtc0 |
06:18.30 | maejrep | you think that would cause the hang? |
06:18.37 | maejrep | NetRipper's prints the same message though |
06:19.17 | tmzt_ | no, but I don't know |
06:19.26 | tmzt_ | it seems this is the clocks |
06:20.19 | tmzt_ | I don't mean the rtc, but the actual clocks |
06:26.19 | maejrep | hmm, so you'd recommend some debugging lines around the rtc / hctosys parts of the kernel? |
06:27.08 | tmzt_ | if this is being called from userspace, I don't see how that will provide more information than you already have |
06:27.22 | tmzt_ | you are using fbcon for the user console? |
06:27.40 | tmzt_ | not druidu console |
06:27.40 | tmzt_ | ? |
06:28.06 | maejrep | yeah, it successfully does the framebuffer console handoff after hillsdale's patch for the additional mddi client |
06:28.53 | tmzt_ | ok, should init or the init scripts be printing anything? |
06:29.11 | tmzt_ | it wouldn't be in dmesg, so you would have to see if it prints on the screen |
06:29.51 | maejrep | I can't actually get a dmesg -- what I see is only what's printed on the screen. the link on netripper's site is his dmesg, which I just modified to match what I see |
06:30.15 | maejrep | The next steps I should be seeing would be the touchscreen calibration |
06:30.39 | maejrep | but even if I just go through the steps of calibration without being able to see anything, it doesn't respond after this |
06:41.35 | *** join/#htc-linux marex (n=marex@vasut.kolej.mff.cuni.cz) |
06:47.38 | *** join/#htc-linux pleemans (n=toi@d54C2AAB7.access.telenet.be) |
07:14.53 | *** join/#htc-linux kiozen (n=oeichler@p5492991C.dip0.t-ipconnect.de) |
07:17.57 | *** join/#htc-linux Zoolooc (n=fredsiba@nrbg-4dbf836f.pool.einsundeins.de) |
07:26.25 | *** join/#htc-linux dsr_ro (i=dsr@89.39.110.83) |
07:26.34 | *** part/#htc-linux dsr_ro (i=dsr@89.39.110.83) |
07:49.31 | *** join/#htc-linux rmoravcik (n=rmoravci@gtsgw.ttc.cz) |
07:57.29 | *** join/#htc-linux PoohbaLT (n=Poohba@c-98-235-66-242.hsd1.nj.comcast.net) |
07:59.53 | *** join/#htc-linux marmotta (n=skodde@unaffiliated/skodde) |
08:03.35 | *** join/#htc-linux the_sys0p (n=the_sys0@cpe-75-85-249-111.bak.res.rr.com) |
08:15.32 | *** join/#htc-linux imfloflo (i=5bcfd11d@gateway/web/ajax/mibbit.com/x-a52683d1054d8894) |
08:24.43 | *** join/#htc-linux kilian_ (n=kilian@92.66.94.81) |
08:25.25 | *** join/#htc-linux zule (i=foglem@shadowmite.com) |
08:29.26 | *** join/#htc-linux timebomb (n=tb@e176105083.adsl.alicedsl.de) |
08:34.36 | *** join/#htc-linux Xime (n=xime@dag94-3-82-233-170-230.fbx.proxad.net) |
08:44.20 | *** join/#htc-linux BabelO (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
08:52.01 | *** join/#htc-linux stefan_schmidt (n=stefan@w1583.wlan.rz.tu-bs.de) |
08:53.12 | NetRipper | maejrep, uncomment the RAMSIZE and set mem=96M in the cmdline |
08:54.18 | NetRipper | maejrep, and try this package: http://www.netripper.com/raphael/20081204-01_raph_diam_angstrom_usb/ |
09:36.36 | *** join/#htc-linux Babel1 (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
09:51.37 | *** join/#htc-linux LanceHaig (i=c2d8ed19@foresight/member/lhaig) |
09:51.43 | *** part/#htc-linux LanceHaig (i=c2d8ed19@foresight/member/lhaig) |
10:23.13 | *** join/#htc-linux BabelO_ (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
10:29.01 | Marajin | mornin BabelO_ |
10:29.12 | BabelO_ | morning Marajin |
10:29.22 | Marajin | how's it going? |
10:30.06 | BabelO_ | Marajin: good, and you ? |
10:30.26 | Marajin | not bad, trying to find a usable Windows GPS nav software though |
10:30.26 | BabelO_ | i m playing with microsoft virtual server, cause we have a bug on it :( |
10:30.47 | Marajin | what's the bug? |
10:42.17 | BabelO_ | Marajin: the bug is related to bad USB driver support |
10:42.39 | BabelO_ | Marajin: there is no USB on virtual server, and got some software bug related to that |
10:50.56 | Marajin | ah |
10:51.02 | Marajin | that's a bit pathetic |
11:06.00 | *** join/#htc-linux mokolade (n=mokolade@85.221.112.173) |
11:34.17 | *** join/#htc-linux Guimli (n=guimli@ecu69-1-82-231-127-213.fbx.proxad.net) |
11:37.35 | *** join/#htc-linux infernixx (n=nix@unaffiliated/infernix) |
11:37.53 | *** join/#htc-linux imfloflo (i=5bcfd11d@gateway/web/ajax/mibbit.com/x-c98a048a640a0e6d) |
11:45.19 | *** join/#htc-linux timebomb (n=tb@p5B3E594C.dip.t-dialin.net) |
11:59.07 | *** join/#htc-linux marex (n=marex@vasut.kolej.mff.cuni.cz) |
12:22.17 | *** join/#htc-linux PoohbaLT (n=Poohba@c-98-235-66-242.hsd1.nj.comcast.net) |
12:42.09 | kiozen | http://www.redferret.net/?p=11407 |
12:44.21 | kiozen | and another usb joke for BabelO_ http://news.softpedia.com/news/Two-Gadgets-That-Will-Have-You-Fired-82801.shtml :) |
13:23.44 | *** join/#htc-linux kimhoon (n=kimhoon@s559116c1.adsl.wanadoo.nl) |
13:24.31 | *** join/#htc-linux goxboxlive (n=goxboxli@24.84-48-212.nextgentel.com) |
13:30.50 | *** join/#htc-linux LunohoD (n=alex@e180070185.adsl.alicedsl.de) |
13:32.53 | *** join/#htc-linux kuhlivisj (n=kuhlivis@s5592d82c.adsl.wanadoo.nl) |
13:37.11 | *** join/#htc-linux LunohoD (n=alex@e180070185.adsl.alicedsl.de) |
13:51.18 | *** join/#htc-linux LunohoD (n=alex@e180070185.adsl.alicedsl.de) |
14:04.24 | *** join/#htc-linux BabelO_ (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
14:13.10 | *** join/#htc-linux pichurri (n=pishuri@194.230.154.96) |
14:13.20 | pichurri | hello everybody |
14:13.40 | pichurri | I have a htc blackstone, also known as touch hd or t8282... |
14:13.52 | pichurri | I'm trying to run linux / android on it |
14:14.24 | pichurri | I took as base, the work done on the htc raphael and diamond |
14:14.31 | pichurri | I get to boot android |
14:15.39 | pichurri | and angstrom |
14:16.08 | pichurri | I can connect via usb to the angstrom and execute commands...via ssh |
14:16.33 | pichurri | but the screen is scrambled, in both cases |
14:16.39 | pichurri | android and angstrom |
14:16.51 | pichurri | any ideas |
14:17.03 | pichurri | I've been over the framebuffer and touchscreen code |
14:17.30 | Bushman | hmm... |
14:17.44 | Bushman | but windows works fine? |
14:17.45 | pichurri | but I'm not a kernel / driver developer (yet), so far I changed the resolution from 640x480 to fit the blackstone 800x480... |
14:18.06 | Bushman | hmm... |
14:18.07 | pichurri | yeah, on windows screen works correctly |
14:18.35 | pichurri | also |
14:18.38 | pichurri | with haret |
14:18.50 | Bushman | huh? |
14:19.00 | Bushman | what do you mean? |
14:19.28 | pichurri | I've dumped the mmu, and hopefully a screenshot |
14:19.57 | Bushman | do you mean that it works with haret or not? |
14:20.16 | pichurri | I ran in haret: pd 0x1686a000 768000 and dump mmc |
14:20.32 | pichurri | yes haret starts the kernel I compiled and boots, |
14:21.02 | pichurri | but the screen is scrambled. |
14:21.12 | Bushman | and what happen when haret finished loading and the hernel started to output to the screen? |
14:21.22 | Bushman | right away? |
14:21.41 | pichurri | I did a video: http://www.youtube.com/watch?v=Siai2KspElE |
14:22.48 | pichurri | brb |
14:24.02 | Bushman | tell me, does it stay like this (like in the end of the video) or anything changes later? |
14:26.38 | *** join/#htc-linux myxor (n=myxor@pdbn-4d08976a.pool.mediaWays.net) |
14:28.36 | *** join/#htc-linux AstainHellbring (n=AstainHe@unaffiliated/astainhellbring) |
14:28.36 | pichurri | at least once, I think the hole boot process went through |
14:28.39 | pichurri | as in http://www.youtube.com/watch?v=_MrWIxDMCsE |
14:29.24 | pichurri | but the raphael/diamond kernel has screen calibration during the kernel boot process |
14:29.29 | pichurri | I taped 5 times |
14:30.34 | pichurri | and then screen went black, except for, what I suppose is the the famous moving dot, after 1-2 secs, screen went all black |
14:31.35 | pichurri | So, I guess android booted correctly but with a scrambled screen |
14:31.36 | Bushman | looks like it has wrong resolution |
14:31.43 | pichurri | yeah |
14:31.44 | Bushman | on your video |
14:32.26 | pichurri | but after that one, I compiled a kernel and changed the resolution, but still, scrambled screen |
14:33.34 | Bushman | what was the one on the video? |
14:33.41 | Bushman | res? |
14:34.20 | pichurri | any ideas? been looking into drivers/video/msm/msm_fb.c drivers/video/msm/mddi_client_toshiba.c |
14:35.27 | pichurri | that was a first try, a vanilla raphael/diamond version |
14:35.33 | pichurri | from netripper |
14:36.09 | pichurri | http://www.netripper.com/raphael/20081204-01_raph_diam_android_v0.8.exe |
14:36.17 | pichurri | brb |
14:36.54 | Bushman | tell me what was the resolution in the kernel that was shown on the video |
14:37.10 | Bushman | your video :http://pl.youtube.com/watch?v=Siai2KspElE |
14:37.15 | Bushman | * http://pl.youtube.com/watch?v=Siai2KspElE |
14:40.55 | pichurri | 640x480 |
14:43.38 | Bushman | and AFTER you changed it and compiled what was the resolution? |
14:45.53 | pichurri | 800x480 |
14:53.23 | *** join/#htc-linux letestdun00b (n=aaabraca@62-244-191-249.cust.exponential-e.net) |
14:54.27 | *** part/#htc-linux letestdun00b (n=aaabraca@62-244-191-249.cust.exponential-e.net) |
14:56.26 | pichurri | any ideas? |
14:56.57 | pichurri | in the meantime I thought about dumping data and post it on the wiki, thats why I asked about haret.... |
15:22.10 | *** join/#htc-linux kilian__ (n=kilian@92.66.94.81) |
15:47.51 | NetRipper | pichurri, nice to see that the HD display works on the kernel |
15:48.29 | NetRipper | pichurri, you compiled the kernel yourself? |
15:49.59 | NetRipper | pichurri, please look through the dmesg to see which panel it detects.. somewhere during boot there's a line like this: "[ 4.624821] mddi_pmdh: publish: mddi_c_d263_0000 |
15:50.12 | NetRipper | i'd like to know that mddi_c_XXXX_XXXX |
15:50.12 | *** join/#htc-linux dcordes (n=dcordes_@unaffiliated/dcordes) |
15:50.13 | NetRipper | :) |
15:50.17 | NetRipper | hid ChanServ |
15:50.18 | pichurri | hey netripper!!! |
15:50.19 | NetRipper | uh |
15:50.21 | NetRipper | hi dcordes |
15:50.24 | dcordes | hi all |
15:50.25 | pichurri | read a lot bout you |
15:50.37 | pichurri | I mean in the thread of the diamond/raphael |
15:50.39 | pichurri | :) |
15:50.43 | NetRipper | yer ;) |
15:50.46 | pichurri | yep, I compiled the kernel |
15:51.00 | NetRipper | ok can you check which panel it detects? |
15:51.47 | dcordes | pichurri, you've got raph100 (gsm) ? |
15:52.01 | NetRipper | dcordes, touch hd |
15:52.17 | dcordes | pichurri, you know the blackstone linux thread? |
15:52.33 | dcordes | pichurri, did you boot linux successfully? |
15:53.25 | NetRipper | 15:21:42 < pichurri> I did a video: http://www.youtube.com/watch?v=Siai2KspElE |
15:53.27 | dcordes | NetRipper, do you have the initrd where you can test-paint yellow dots for ts? |
15:53.30 | pichurri | [ 4.654241] mddi_pmdh: publish: mddi_c_4ca3_0000 |
15:53.32 | dcordes | good |
15:53.39 | NetRipper | pichurri, ok |
15:53.59 | dcordes | pichurri, is this video produced with the raphael kernel built from htc-msm-2.6.25 with raphael_defconfig? |
15:54.16 | pichurri | yes linux boots correctly I used angstrom from diamond/raphael and got shell via ssh |
15:54.30 | dcordes | bon |
15:54.30 | NetRipper | pichurri, same panel as raph800 |
15:54.35 | pichurri | toshiba |
15:54.42 | dcordes | pichurri, do me a favor and boot this zImage http://linuxtogo.org/~lgorris/builds/kernel/xperia/zImage-x1-02.bin |
15:55.12 | pichurri | testing... |
15:55.16 | dcordes | thanks |
15:55.59 | dcordes | at least the white bar at the bottom should disappear. |
15:56.00 | NetRipper | pichurri, at the top of mddi_client_toshiba.c there's a function toshiba_enable_panel_raph800 which sets width and height, you can change that to correct res and see if that helps.. (even better, you could try commenting out the assignments to width and height - perhaps it detects it correctly |
15:56.16 | dcordes | seems to me msm_fb.c needs some geometry tweak for hd |
15:56.32 | NetRipper | i gtg again, am still at work ;) |
15:56.37 | dcordes | :P |
15:57.13 | dcordes | the blackstone really is one fine devicer |
15:57.15 | dcordes | device |
15:57.28 | dcordes | Xmoo123 (vor 15 Stunden) Anzeigen Ausblenden |
15:57.28 | dcordes | 0 Schwacher Kommentar Guter Kommentar |
15:57.28 | dcordes | Als Spam markiert |
15:57.28 | dcordes | Antworten |
15:57.29 | dcordes | Copy my zImage from the Xperia, This will help to get fullscreen and maayb the GUI to work! |
15:57.29 | dcordes | lol |
15:57.29 | pichurri | same thing |
15:57.39 | pichurri | did you change the resolution to 800x480? |
15:57.40 | dcordes | he talks about the 02 build |
15:57.43 | dcordes | exactly |
15:58.22 | pichurri | also tried the xperia, from http://www.xmoo.org/X1/linux/Android%200.8/zImage |
15:58.25 | pichurri | but no good |
15:58.36 | *** join/#htc-linux cr2 (n=konversa@crpl6.physik.uni-wuppertal.de) |
15:58.44 | dcordes | this is a copy of one of the zImages at same url as above |
15:58.49 | pichurri | ahhh |
15:59.02 | dcordes | initially it was just the raphael kernel image, then I did the panel tweak |
15:59.05 | pichurri | also I got some dumps from haret |
15:59.17 | pichurri | I tried pd 0x1686a000 0xBB800 |
15:59.21 | dcordes | as I said it seems to me you need to change geometry in msm_fb.c |
15:59.44 | pichurri | couldn't it be a sync problem |
15:59.51 | pichurri | ? |
15:59.51 | dcordes | was hoping you've got the same panel as x1 |
16:00.03 | dcordes | hm no clue |
16:00.14 | dcordes | maybe you need more size for the framebuffer? |
16:00.37 | cr2 | hi |
16:00.45 | cr2 | kiozen: here ? |
16:01.09 | kiozen | yes |
16:01.15 | pichurri | would a mmu dump from haret help? |
16:01.23 | dcordes | I have one already |
16:01.32 | pichurri | from the blackstone? |
16:01.35 | dcordes | http://linuxtogo.org/~lgorris/misc/blackstone-mmu.txt |
16:01.36 | cr2 | dcordes: blackstone spl found ? |
16:01.42 | dcordes | cr2, no but I have the x1 spl |
16:01.57 | cr2 | kiozen: it seems that the polish guy will be out of business soon :) |
16:02.06 | dcordes | pichurri, can you provide cr2 with the blackstone spl? |
16:02.23 | kiozen | cr2: don't think so |
16:02.47 | cr2 | blackstone has mtype ? |
16:02.48 | pichurri | I thought the spl didn't work with the blackstone... |
16:02.52 | kiozen | gpsmapedit has another focus than mkgmap |
16:03.09 | cr2 | kiozen: time to put the last nail after christmas |
16:03.32 | kiozen | routing? |
16:03.36 | cr2 | kiozen: gpsmapedit != cgpsmapper |
16:03.38 | cr2 | yes |
16:03.56 | cr2 | routing is already working |
16:04.06 | cr2 | with mkgmap |
16:04.06 | kiozen | but what is it worth in GT? it will never calc the same rout as the device |
16:04.10 | dcordes | cr2, yea http://www.arm.linux.org.uk/developer/machines/?action=list&id=1951 |
16:04.21 | cr2 | dcordes: ok |
16:04.21 | dcordes | cr2, do we need to update haret? |
16:04.43 | cr2 | dcordes: yes, mach-types.h |
16:05.05 | cr2 | kiozen: it's about marketing |
16:05.40 | dcordes | cr2, let's add the shift and notify Kevin2 |
16:05.46 | kiozen | no, a source of endless user complains "Why doesen't GT calc the same route as my Garmin?" |
16:06.27 | cr2 | kiozen: ok, but isn't mapsource calculating the different route too ? afaik |
16:06.47 | *** join/#htc-linux metter (n=metter@213-88.1-85.cust.bluewin.ch) |
16:06.48 | kiozen | yes, that is why I know the complains :) |
16:06.59 | cr2 | A* is heuristics anyway |
16:07.03 | kiozen | but it does a 90% job |
16:07.17 | kiozen | we don't know the weights |
16:07.22 | cr2 | so it's not an optimal route anyway |
16:07.48 | cr2 | time=distance/speed |
16:07.53 | kiozen | imho this stuff is more important for navit and roadmap |
16:08.42 | kiozen | the hard beaten garmin user will set waypoints and let the device calc the route :) |
16:08.43 | cr2 | yes, but all suck desktop apps support route planning |
16:08.51 | cr2 | s/suck/such/ |
16:08.57 | kiozen | lol |
16:09.10 | cr2 | :) |
16:09.24 | kiozen | let's see if someone comes up with it |
16:09.25 | *** join/#htc-linux GPFerror (n=gpferror@cpe-76-187-41-132.tx.res.rr.com) |
16:09.50 | kiozen | imho M needs attention badly after xmas |
16:10.10 | cr2 | compile a new qtextended image |
16:10.35 | cr2 | and we need to extract your spl somehow |
16:10.43 | cr2 | direct off the flash. |
16:10.47 | kiozen | yes, that and rewrite M to be portable to wince |
16:11.22 | cr2 | maybe. so it runs faster on linux :) |
16:11.56 | cr2 | i've tried the BabelO .exe, it works, but is really slow |
16:12.05 | *** join/#htc-linux infernixx (n=nix@unaffiliated/infernix) |
16:12.05 | cr2 | probably because of the debugging |
16:13.09 | cr2 | dcordes: an updated haret will be nice, so it detects all the phones known to us |
16:15.20 | dcordes | cr2, we figured the X on msm_fb problem. |
16:15.52 | dcordes | msm_fb_update needs to be called in order to refresh |
16:16.42 | dcordes | now I wonder what is the way to make it update automatically when required |
16:16.49 | *** join/#htc-linux tsdogs (n=tsdogs@net70-17.metalit.net) |
16:17.13 | dcordes | I'm about to make camera gpio call the function like with the polaris keyboard so I can test |
16:17.48 | kiozen | bbl |
16:17.57 | dcordes | msm_update_screen() |
16:23.37 | *** join/#htc-linux LanceHaig (i=c2d8ed19@foresight/member/lhaig) |
16:23.45 | *** part/#htc-linux LanceHaig (i=c2d8ed19@foresight/member/lhaig) |
16:23.46 | dcordes | pichurri, what's at 0x1686a000 ? |
16:26.56 | cr2 | dcordes: a vsync gpio irq ? |
16:27.36 | pichurri | the framebuffer memory, no? |
16:28.03 | cr2 | pichurri: print "%x" VRAM |
16:28.12 | pichurri | on haret? |
16:28.32 | cr2 | yes |
16:28.50 | *** join/#htc-linux imfloflo (i=5bcfd11d@gateway/web/ajax/mibbit.com/x-ab601ac1fb487f46) |
16:29.44 | pichurri | cr2: 16044800 |
16:29.52 | *** join/#htc-linux miknix (n=miknix@gentoo/developer/miknix) |
16:30.00 | cr2 | kiozen: http://www.sueddeutsche.de/wissen/293/451009/text/print.html |
16:30.26 | cr2 | pichurri: ok, then it's your framebuffer address in wince. |
16:30.56 | pichurri | how can I take a screenshot dumping the memory... |
16:31.08 | pichurri | cr2: and would that be of any use...? |
16:31.33 | imfloflo | with another apn |
16:32.26 | cr2 | pichurri: pwf screensho VRAM xsize*ysize*16 |
16:36.54 | pichurri | cr2: it created a file, screensho, what should I do with it... |
16:37.05 | BabelO_ | cr2: hi, i already generate an OSM map with routing for test :) |
16:37.36 | cr2 | BabelO_: with which program ? |
16:37.51 | cr2 | pichurri: you've asked how to create a screenshot |
16:38.35 | pichurri | cr2: gimp doesn't open it thou... |
16:38.56 | BabelO_ | cr2: mkgmap and also tried osm2mp then mkgmap from mp file :) |
16:39.12 | BabelO_ | cr2: there is mkgmap --net in beta version |
16:39.26 | cr2 | pichurri: it's raw hexdump, you may convert it into .pnm of sometrhing. |
16:40.00 | cr2 | BabelO_: yes, i've seen. but how do you test the result ? do you have a garmin device ? |
16:40.19 | dcordes | ./arch/arm/mach-msm/board-htcraphael.c:#define MSM_FB_BASE MSM_PMEM_GPU1_BASE + MSM_PMEM_GPU1_SIZE |
16:40.35 | BabelO_ | cr2: no garming device i just lot result file in gpsmapedit to see if i see arrow on the street |
16:40.48 | BabelO_ | cr2: if you want i can generate a town for you ? |
16:42.26 | dcordes | pichurri, what's your actual fb base now? |
16:43.04 | pichurri | dcordes: sorry, base? |
16:43.20 | dcordes | raphael is 1666a000 |
16:43.42 | dcordes | framebuffer memory address |
16:44.04 | pichurri | dcordes: apparently 16044800. On haret, print "%x" VRAM gives that address |
16:44.33 | dcordes | let's put it in the board.c then and try |
16:45.44 | pichurri | dcordes: in the kernel I don't have a board.c... |
16:46.59 | dcordes | arch/arm/mach-msm/board-htcraphael.c |
16:47.29 | dcordes | 2621440 is enough for blackstone fb size ? |
16:47.33 | dcordes | 0x280000 |
16:47.43 | dcordes | it seems ok on the x1 |
16:48.49 | *** join/#htc-linux pleemans (n=toi@d54C2AAB7.access.telenet.be) |
16:50.26 | pichurri | dcordes: 0x280000 should go in : "#define MDDI_CLIENT_CORE_BASE"? |
16:50.52 | pichurri | dcordes, sorry, in MSM_FB_SIZE |
16:51.28 | *** join/#htc-linux Dubbeld00 (n=opera@82-168-116-233.ip.telfort.nl) |
16:52.34 | dcordes | yea |
16:54.05 | pichurri | dcordes: I ended up with: #define MSM_FB_BASE 0x16044800. #define MSM_FB_SIZE 0x280000. |
16:54.13 | pichurri | dcordes: compiling and testing |
16:54.32 | dcordes | just compiled with same values |
16:54.48 | dcordes | did you also put the resolution in the mddi client ? |
16:58.08 | pichurri | dcordes: yes, I have full screen. should the penguin go to the right or left? |
16:58.28 | dcordes | left iirc |
16:58.40 | pichurri | dcordes, I have most of it on right top, |
16:58.43 | dcordes | did you just replace 640 with 800 in the toshiba client? |
16:58.51 | pichurri | yep |
16:58.57 | pichurri | but also... |
16:59.03 | dcordes | that is also in the zImage-x1-02.bin |
16:59.43 | pichurri | on drivers/input/touchscreen/msm_ts.c |
16:59.46 | pichurri | +#define MSM_TS_ABS_Y_MAX 799 |
16:59.56 | pichurri | -#define MSM_TS_ABS_Y_MAX 639 |
17:00.03 | dcordes | ah :) |
17:00.22 | pichurri | also in drivers/input/touchscreen/msm_vkeyb.c |
17:00.28 | cr2 | msm_ts ? does the irq work now ? |
17:00.33 | pichurri | -#define MSM_VKEYB_LCD_HEIGHT 640 |
17:00.33 | pichurri | +#define MSM_VKEYB_LCD_HEIGHT 800 |
17:00.38 | dcordes | ok |
17:00.51 | pichurri | and in drivers/video/msm/mddi_client_toshiba.c |
17:00.57 | cr2 | why does the ts driver need to know the display size ?! |
17:01.05 | pichurri | - panel->height = 640; |
17:01.06 | pichurri | + panel->height = 800; |
17:01.06 | cr2 | sounds braindead to me |
17:01.13 | dcordes | oh come on. not that harsh |
17:01.21 | cr2 | :) |
17:01.21 | *** join/#htc-linux Kuma (n=John@f054236114.adsl.alicedsl.de) |
17:01.31 | dcordes | at least there is a working ts driver, no? |
17:01.42 | pichurri | I guess |
17:01.46 | cr2 | no idea |
17:02.04 | pichurri | I think the kernel I use, does a touchscreen calibration |
17:02.28 | pichurri | and I've seen the amazing moving dot when android starts... |
17:02.31 | dcordes | it expects you to tap top left and top right corner |
17:02.35 | dcordes | NetRipper, correct? |
17:02.41 | pichurri | just 2 taps? |
17:02.52 | dcordes | not sure druidu made the driver |
17:03.09 | dcordes | also the msm console hack and the on screen keyboard. he's also on the xda forum |
17:03.20 | pichurri | after some taps, screen went black and something moved on the center |
17:03.29 | pichurri | after that, all black... |
17:04.13 | cr2 | suckers ;) |
17:04.18 | cr2 | dcordes: monatliche Grundgebühr34,95 € (ab dem 7. Monat 44,95 Euro) |
17:05.03 | cr2 | 42.4125€ per month |
17:06.17 | cr2 | Datenflatrate ohne Volumenbegrenzung |
17:06.54 | dcordes | sounds ok. do you also get call minutes and text messages? |
17:08.19 | cr2 | no |
17:08.41 | cr2 | but they pay back 420€ in 2 months |
17:08.48 | *** part/#htc-linux BabelO_ (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
17:09.39 | dcordes | shouldn't the calibration be done in userspace? |
17:10.05 | dcordes | also, is there no on screen keyboard in android? |
17:10.17 | dcordes | would also be useful for the vogue users |
17:10.26 | pichurri | THERE IS |
17:10.29 | pichurri | sorry, |
17:10.39 | pichurri | there is a keyboard |
17:11.22 | dcordes | I don't mean the kernel keyboard |
17:13.04 | *** join/#htc-linux LunohoD_ (n=alex@e180079204.adsl.alicedsl.de) |
17:13.30 | *** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
17:19.05 | *** join/#htc-linux Bally3 (n=chatzill@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
17:20.13 | Bally3 | boo! |
17:20.46 | tcccp | eeeek! |
17:22.48 | dcordes | pichurri, you still have the LSD colors? |
17:23.03 | pichurri | yep |
17:23.18 | dcordes | pichurri, it is the same problem on the raphael800 and xperia x1 |
17:23.20 | pichurri | wait |
17:23.26 | pichurri | I think the colors are fine |
17:23.45 | *** join/#htc-linux oktopouce (n=oktopouc@3.69.73-86.rev.gaoland.net) |
17:23.50 | pichurri | does this devices have refresh rates and sync? |
17:24.03 | pichurri | I think it is more that kind of problem... |
17:24.14 | dcordes | grep msm_fb.c |
17:24.21 | dcordes | drivers/vidoe/msm/ |
17:24.26 | dcordes | vidö |
17:24.35 | Bally3 | dcordes.. to run the haret/initrd setup, do you have to have an ext3 partition set up first? |
17:25.07 | dcordes | no the initrd is for the sort of situation where you have no such partition or ability to access a loopfile or such |
17:26.15 | Bally3 | thanks matey.. did you check the changes rolk made last night? dwaradzyne posted up a zimage with the changes.. we seem to be making some progress.. |
17:26.33 | Bally3 | could really use your input on it tho |
17:26.50 | pichurri | dcordes, have to leave...but I'll be coming back in 6-7 hours I think... |
17:27.38 | dcordes | pichurri, see you |
17:28.07 | dcordes | tmzt_, ping |
17:32.13 | Bally3 | dcordes.. dzo put 0800000 whatevver as his RAMSIZE we have 0780000 - does it make a difference? |
17:32.35 | Bally3 | a few more zeroes but I I'm not in it at the mo to check :P |
17:43.17 | dcordes | what is #define MDDI_CLIENT_CORE_BASE 0x108000 ? |
17:49.37 | tmzt_ | dcordes: hey |
17:49.50 | tmzt_ | maejrep: did you get that message from NetRipper ? |
17:50.38 | dcordes | tmzt_, need your help |
17:51.33 | dcordes | tmzt_, how can we call msm_update_screen automatically in order to make X work? |
17:51.44 | dcordes | liek it is done in http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=blob;f=arch/arm/mach-msm/vogue-ts.c;hb=refs/heads/htc-msm |
17:52.10 | dcordes | there it's used to update the screen when the on screen keyboard is called |
17:52.21 | dcordes | but it also makes X draw as a side effect |
17:56.44 | tmzt_ | I guess put back the virtual VBI timer from the early dzo msm_fb |
17:57.30 | dcordes | do you know where the early dzo msm_fb is? |
17:57.49 | tmzt_ | 2.6.24? |
17:58.58 | dcordes | sorry I mean where can I get it? |
17:59.12 | *** part/#htc-linux oktopouce (n=oktopouc@3.69.73-86.rev.gaoland.net) |
18:02.00 | *** part/#htc-linux Dubbeld00 (n=opera@82-168-116-233.ip.telfort.nl) |
18:08.56 | tmzt_ | dcordes: is the 2.6.24 tree still on ltg? |
18:09.19 | dcordes | tmzt_, htc-msm ? |
18:10.07 | tmzt_ | yes |
18:10.45 | dcordes | yea in that branch dwaradzyn used X on the polaris |
18:10.56 | dcordes | pressing the cam button to refresh |
18:11.03 | dcordes | http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=blob;f=arch/arm/mach-msm/vogue-ts.c;hb=refs/heads/htc-msm |
18:11.25 | tmzt_ | I mean as in a timer, an irq that is triggered 60 times a second or whatever |
18:11.39 | tmzt_ | does pressing the camera button refresh X? |
18:11.43 | dcordes | yes |
18:11.50 | dcordes | because it calls msm_update_screen |
18:12.17 | tmzt_ | we need to find someone who knows how to write a ddx then |
18:12.33 | dcordes | ddx? |
18:13.16 | tmzt_ | display dependent x / dix is display independent x (or x means something else) |
18:13.30 | tmzt_ | basically the driver part of the X server |
18:13.58 | dcordes | you want to call msm_update_screen from X ? |
18:14.14 | tmzt_ | you can't call a kernel function from X |
18:14.30 | tmzt_ | you need a ioctl or syscall or netlink or something |
18:14.33 | dcordes | how do you want to make the msm_fb refresh then? |
18:14.54 | dcordes | so ioctl mediates between kernel and userspace? |
18:15.16 | tmzt_ | anything that calls msm_update_screen should work, setting the yoffset (and double buffering), using PAN |
18:15.45 | tmzt_ | or we could just add an UPDATE_SCREEN for our driver, but it would be better to be compatible with google |
18:16.26 | tmzt_ | ioctl is an operation on a device, called through a file descriptor (fd) which has been used to open that device |
18:19.27 | tmzt_ | we could also probably use MSM_BLIT if we can figure out the semantics for calling it |
18:20.29 | tmzt_ | we know it takes a source region and destination region, but we are not using offscreen pixmaps (which would be implemented using the extra framebuffer memory) |
18:21.09 | tmzt_ | we want to draw directly into the framebuffer like the fbdev driver normally does (using memory pixmaps if neccessary in the display independent part of X) |
18:22.26 | dcordes | tmzt_, ok. for now I should confirm on kaiser what dwaradzyn did on polaris |
18:22.32 | tmzt_ | right |
18:22.58 | tmzt_ | you can also take out the onscreen keyboard part and just make camera refresh the display (call msm_fb_update) |
18:23.06 | dcordes | I would like to make a button call the function |
18:23.14 | dcordes | right |
18:23.49 | dcordes | board-kaiser-keypad.c would be a good place? |
18:24.01 | dcordes | I added "extern void msm_update_screen(void);" |
18:24.47 | dcordes | {KEY_CAMERA, KEY_CAMERA, EGPIO_6_5_KEY_CAM1, 0, "Camera 1"}, |
18:27.43 | dcordes | tmzt_, we could also have pendown call it |
18:29.31 | cr2 | dcordes: btw, what about changing from CPQ/itsy to NetChip/rndis gadget on 720xA ? |
18:29.37 | cr2 | is the msm usb rndis-capable ? |
18:30.01 | dcordes | cr2, feel free to commit the patch I guess |
18:30.08 | dcordes | do you have access? |
18:30.19 | dcordes | yes it is |
18:30.48 | *** join/#htc-linux woodson (n=CDP@c-76-101-90-149.hsd1.fl.comcast.net) |
18:30.58 | cr2 | no i don't have the access. |
18:31.00 | cr2 | somebody should check it first too |
18:31.12 | cr2 | but imho it's a reasonable change |
18:31.38 | cr2 | the ids are 0x0525 / 0xa4a2 |
18:34.10 | dcordes | in http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=blob;f=arch/arm/mach-msm/tsc2003.c;hb=refs/heads/htc-msm how can I make pendown call msm_upadte_screen ? |
18:39.41 | *** join/#htc-linux Bally3 (n=chatzill@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
18:41.33 | NetRipper | 18:02:31 < dcordes> it expects you to tap top left and top right corner |
18:41.33 | NetRipper | 18:02:36 < dcordes> NetRipper, correct? |
18:41.37 | NetRipper | top left, lower right |
18:42.14 | *** join/#htc-linux rolk (n=rolk@ip5457417f.direct-adsl.nl) |
18:43.29 | *** join/#htc-linux earthman (n=chatzill@diamond.la.net.ua) |
18:43.32 | NetRipper | and yes - calibration should eventually be done in user space.. currently, the first 3 taps provide calibration.. to make the driver returns calibrated results, as android 0.8 didn't have a calibration screen |
18:45.01 | Bally3 | hi rolk |
18:45.25 | rolk | Hi |
18:45.48 | *** join/#htc-linux dwaradzyn (n=dwaradzy@chello089079197022.chello.pl) |
18:45.58 | Bally3 | did you give any thoughts to last nights doings? |
18:46.11 | Bally3 | lo dwaradzyne :) |
18:46.26 | dwaradzyn | hi there |
18:46.32 | rolk | Gi D, |
18:46.34 | *** join/#htc-linux bertramt (n=chatzill@63.246.89.17) |
18:46.39 | rolk | Hi D. |
18:47.03 | dwaradzyn | i read again what dzo said, and he was talking about different two values |
18:47.24 | dwaradzyn | the zimage i posted is not really good for testing |
18:48.22 | rolk | I've booted the kernel and initrd D. posted a few more times, and each time the SD card is successfully detected. 3-4 timeouts, no more. File operations (modest size, no stress testing) seem to work fine. |
18:48.23 | cr2 | NetRipper: do you have usb working ? |
18:48.31 | dwaradzyn | i think we need a test scenario and put some experiments with different max_block_count i max_block_size |
18:48.51 | Bally3 | again we seem to have moved forward after the sd card detection? |
18:48.55 | rolk | I retrieved the config file from that kernel, and am rebuilding from the patched LTG sources. |
18:49.12 | dwaradzyn | rolk - it was going fine from months. we need stress tests |
18:49.32 | cr2 | guys, i'll suggest the sd clock documenting project |
18:49.38 | Bally3 | can you post up the config file please rolk? |
18:50.12 | rolk | Yes, that should be no problem. |
18:50.17 | cr2 | please dump the a0 and a4 clock regs for sd0 and sd1, and put them somewhere in wiki |
18:50.27 | cr2 | naming the device and the (SD or MMC) card that was used |
18:50.38 | dwaradzyn | also i did not change clock to 66mhz as dzo suggested |
18:52.53 | *** join/#htc-linux Rogro82 (n=rogro82@s5591104d.adsl.wanadoo.nl) |
18:56.43 | dwaradzyn | cr2: how do we do that? do we put some printks in sd init code? |
18:57.33 | *** join/#htc-linux Bally3 (n=chatzill@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
18:58.06 | Bally3 | damn died big time :S |
18:58.14 | *** join/#htc-linux Dubbeld00 (n=opera@82-168-116-233.ip.telfort.nl) |
18:58.40 | Bally3 | hey dubbeld00 |
18:59.07 | Dubbeld00 | hello Bally3 |
18:59.52 | dwaradzyn | cr2: or can it be done using haret only? |
19:02.21 | dcordes | dwaradzyn, hi, can you help with my question? |
19:02.34 | dwaradzyn | dcordes: yes? |
19:03.12 | dcordes | I'd like to call msm_update_screen in a loop or on ts press |
19:03.14 | dcordes | pendown |
19:04.17 | *** join/#htc-linux miknix (n=miknix@gentoo/developer/miknix) |
19:06.07 | Bally3 | hi rogro :) |
19:06.36 | cr2 | dwaradzyn: actually, we may just dump the whole clock area, and create an archive |
19:06.37 | dwaradzyn | dcordes: does kaiser use vogue-ts.c? |
19:06.59 | cr2 | dwaradzyn: yes, just use 'pd' on the 0xa86* MSM_CLK area |
19:07.12 | *** join/#htc-linux Bally3 (n=chatzill@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
19:07.24 | dwaradzyn | cr2: so it is haret then |
19:07.45 | cr2 | yes |
19:08.10 | cr2 | then you don't write weird stuff into clock registers |
19:08.11 | cr2 | and let wince do it :) |
19:08.26 | cr2 | the 7200 values are not appropriate for 720xA |
19:08.43 | Bally3 | polaris is 7200? |
19:09.15 | dwaradzyn | ok i'll try, i only used haret cosole once and it was with your guidance :) |
19:09.26 | dwaradzyn | Bally3: yes its 7200 |
19:09.29 | cr2 | and my 720xA spl research shows different values than actually used by wince. |
19:09.30 | cr2 | but at least it's the documented and clearly undestodd code with bitmask group manipulation |
19:10.10 | Bally3 | can you post up the config file we're using dwaradzyn please |
19:10.10 | *** join/#htc-linux thejigsaw (n=thejigsa@77.29.75.249) |
19:10.18 | cr2 | i.e. there is not better reference anyway |
19:10.24 | dcordes | dwaradzyn, nope. it uses tsc2003.c |
19:10.30 | cr2 | unless somebody will leak the msm .pdf docs ;) |
19:10.32 | thejigsaw | hi to all |
19:11.21 | thejigsaw | i saw that polaris developement team is gathering today at 8 o clock |
19:11.30 | thejigsaw | i wanted to help but i have a little work |
19:11.44 | thejigsaw | so i come to express my support and count on me for future |
19:13.00 | Dubbeld00 | great :) |
19:13.15 | thejigsaw | i install ubuntu |
19:13.27 | thejigsaw | so from tomorow i will help too as much as i can |
19:14.00 | Dubbeld00 | btw, little question, how can you checkout the git using a commit hash (so I want to checkout a specific commit)? |
19:14.22 | thejigsaw | any news on the progress |
19:14.23 | thejigsaw | <PROTECTED> |
19:14.24 | *** join/#htc-linux BabelO (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
19:14.36 | dwaradzyn | git-reset --hard <commit id> |
19:14.52 | Dubbeld00 | tnx dwaradzyn (sorry to lazy to read the man page ;)) |
19:14.58 | dcordes | cr2, what about looking at the clock dlls with a rather big machine? |
19:15.16 | dcordes | cr2, that's how ginge got the 7200 mmc clock. |
19:15.52 | cr2 | dcordes: they contain some convoluted code, referencing other dlls |
19:15.58 | cr2 | there is no direct hw access there |
19:16.17 | dcordes | how did ginge get the mmc clocks out of them then? |
19:16.27 | cr2 | at least i don't see it, and nobody proved the opposite |
19:16.50 | cr2 | don't know, we need to ask him directly |
19:17.09 | cr2 | keep also in mind that 7200 != 720xA |
19:17.23 | dcordes | I asked him how I can try it on my own. but I still didn't buy a copy of the disassembly program and didn't come to learn asm |
19:18.08 | dcordes | that's why I said 7200 |
19:18.42 | cr2 | the msm_clk problem is certainly the wrong task to start with |
19:19.16 | cr2 | if it was easy, i've already done it :) |
19:19.28 | dcordes | what about "<cr2> guys, i'll suggest the sd clock documenting project" ? |
19:19.33 | cr2 | like gpio assignment |
19:19.58 | cr2 | dcordes: dump your a0 and a4 regs in wince |
19:20.39 | cr2 | or just the 0x200 bytes from the clock area |
19:20.52 | dwaradzyn | dcordes: try putting msm_update_screen() in tsc2003.c in tsts_thread just after the line pendown = 1 |
19:21.08 | dcordes | ok |
19:26.13 | Bally3 | dwaradzyn: didnt that vm link? |
19:26.25 | Bally3 | dwaradzyn: didnt get that vm link? |
19:27.51 | dwaradzyn | bally, i did not upload it yet. i'll compress that and open ftp for you in a while |
19:28.15 | Bally3 | thanks |
19:29.31 | Bally3 | sohow come you didnt make 66mhz change dwaradzyn |
19:32.51 | dwaradzyn | Bally3: its a simple change in code - i'll do that later |
19:33.23 | Bally3 | so where we at? |
19:35.16 | dwaradzyn | cr2: so according to kaiser memory map there are 12 clock values unser MSM_CLK. size for one value is 32 bits? |
19:38.59 | dwaradzyn | cr2: my dumps are here http://pastebin.com/m2f50f25d, could you take a look and tell it is what we need? |
19:40.57 | rolk | The config file I retrieved from the working zImage is here: http://rapidshare.com/files/172786952/confi |
19:41.11 | rolk | http://rapidshare.com/files/172786952/config |
19:41.33 | rolk | Sorry for the typo. |
19:42.33 | Bushman | i've fell asleep -.- |
19:42.35 | Bushman | ._<' |
19:42.40 | Bushman | * >_<' |
19:43.14 | rolk | D. I recompiled the LTG+full_keyboard.diff patch with this config file, but got a different zImage. I've been looking for an easy way to disassemble the zImage from you and the one I just compiled. I'd like to understand the difference. |
19:46.05 | dwaradzyn | rolk: post link to git commit you checked out - i want to be sure we are talking about the same sources |
19:47.55 | rolk | D. The branch is htc-msm, commit 1082670ccae30e6592311cc9dee93ba978382a25. |
19:49.12 | Rogro82 | hiya |
19:49.14 | cr2 | dwaradzyn: documented values. the other registers are used for other clocks that we don't know about |
19:49.14 | cr2 | dwaradzyn: do your a0 and a4 values match kernel code ? |
19:49.36 | cr2 | dwaradzyn: yes, 1 clock reg is 4byte=32bit |
19:50.03 | dwaradzyn | cr2: where in kernel i can check a0 and a4 values? |
19:50.23 | rolk | D. That .config and that code base + patch now seems to give me a kernel that more or less reliably detects the SD card. |
19:50.28 | cr2 | dwaradzyn: you can just dump the 0x200 block with 'pd 0xa8600000 0x200' |
19:50.36 | cr2 | in the SD related driver |
19:50.51 | cr2 | a0 and a4 clock registers are SD0 related |
19:51.01 | cr2 | the a8 and ac are for the SD1 |
19:51.12 | Rogro82 | i dont now what im doing wrong but somehow my zImage tries to mount the partition on the sd card before it gets initialized... did i miss something? |
19:51.16 | cr2 | aka wifi |
19:51.35 | rolk | Its odd, as the differences between the .config I used yesterday and the one I retrieved from your kernel seem all 'unrelated' to SD. Except perhaps the general config switch 'CONFIG_BLK_DEV'. |
19:51.39 | dwaradzyn | Rogro82: add rootdelay=5 to default.txt to beat that |
19:51.48 | Rogro82 | k thnx |
19:52.31 | dwaradzyn | cr2: reading the code... |
19:52.37 | dcordes | Rogro82, add a rootwait to cmdline |
19:53.42 | cr2 | dwaradzyn: the current clock code is a hack, and therefore a fscking mess :) |
19:54.09 | dwaradzyn | dcordes: you always correct me that way :) and as always i ask you: what is the difference between rootdelay and rootwait ? |
19:54.45 | dwaradzyn | cr2: damn it, but i still want to try |
19:54.47 | cr2 | rootwait is something new |
19:56.27 | Bally3 | how do you log whats happening on the screen when you run haret? I had an interesting screen tellig me that mmc0 was found but it disappeared too fast |
19:57.04 | Bally3 | I'm sure using dwaradzyn's zimage the sd card error isnt reported |
19:57.10 | cr2 | Bally3: if the see the linux messages, haret is already over |
19:57.34 | Bally3 | ok.. so how do I get it from linux? the thing crashes at that point so reboot |
19:57.50 | cr2 | as in "completely dead and overwritten" |
19:57.50 | dwaradzyn | Bally3: check linux command: dmesg |
19:57.56 | cr2 | like wince |
19:57.59 | Bally3 | but dont see a log file anywhere |
19:58.18 | dwaradzyn | in that case it is not logged anywhere... |
19:58.29 | Bally3 | where dwaradzyn? I cant get into linux |
19:58.38 | cr2 | dmesg dumps an internal kernel buffer |
19:59.00 | Bally3 | can you place that in your default file to generate it? |
19:59.22 | Bally3 | I cant get at a console to type anything |
19:59.39 | Bally3 | black screen and freeze only |
20:00.11 | *** part/#htc-linux Dubbeld00 (n=opera@82-168-116-233.ip.telfort.nl) |
20:00.55 | dcordes | dwaradzyn, it will just wait until everything is ready |
20:01.11 | dcordes | I think it mounts when the device is up |
20:01.19 | Bally3 | what did you get from your zimage dwaradzyn? it looked like it had got past the sd init |
20:01.30 | dwaradzyn | bally: one way is to connect to device by telnet/ssh and run "dmesg | tail -f". if you have luck there will be something reported before device reboots |
20:01.49 | dcordes | dwaradzyn, msm_update_screen works |
20:02.05 | dwaradzyn | dcordes: ok - just looked it up an now i know the difference :) |
20:02.06 | Bally3 | ok .. thanks... will try that when I get your set up I guess |
20:02.23 | dwaradzyn | dcordes: that great |
20:03.16 | *** join/#htc-linux RgR (n=irc@195-241-90-16.ip.telfort.nl) |
20:03.23 | rolk | Anyone an idea why the experimental 64 bit memory and io configuration setting is enabled in the kernel config? |
20:03.53 | dwaradzyn | rolk: i have no idea, but it was there all the time AFAIR |
20:04.21 | dwaradzyn | rolk: didn't have guts to change that :P |
20:04.34 | dcordes | rolk, compare with msm_defconfig |
20:07.03 | rolk | D. have you checked the commit? |
20:07.40 | dwaradzyn | Bally3: about your logging question. the only way to connect through network to polaris now is to connect it to 3g and do a reverse ssh. there was an angstrom initrd that had scripts for doing that |
20:08.45 | dwaradzyn | rolk: give me few mins, i needed to shut down vmware with linux for a while |
20:09.02 | Rogro82 | :-S still failed to mount sd card |
20:09.23 | Bally3 | thanks dwaradzyn .. may be worth looking at... Im sure there was something there |
20:09.40 | dwaradzyn | Rogro82: what are your symptoms? |
20:10.40 | Rogro82 | its trying to mount /dev/block/mmcblk0p1... that fails then the polling is triggered and the mmc is initialized :S |
20:11.12 | dwaradzyn | Rogro82: did it fall back to shell? |
20:11.35 | Rogro82 | it hangs |
20:12.00 | Bally3 | you loading angstrom Rog? |
20:12.11 | Rogro82 | no android |
20:12.29 | dwaradzyn | Rogro82: could you pastebin last few messsages you have onscreen ? |
20:12.41 | Bally3 | black screen? |
20:12.55 | Rogro82 | the last few are the ones that show the initialization of the mmc card you want those or the failing ones? |
20:13.11 | cr2 | rolk: wtf 64bit access on a 32bit cpu ? |
20:13.43 | dwaradzyn | Rogro82: pastebin both |
20:13.44 | cr2 | afair even alpha used only 48bit |
20:13.45 | Rogro82 | mount: mounting /dev/block/mmcblk0p1 on /sdcard failed: No such device or address |
20:13.56 | Rogro82 | Failed to mount SD Card |
20:14.14 | cr2 | ~ping NetRipper |
20:14.15 | apt | pong NetRipper |
20:14.17 | dwaradzyn | Rogro82: use www.pastebin.com for that |
20:18.08 | *** join/#htc-linux dcordes_ (n=dcordes_@unaffiliated/dcordes) |
20:18.55 | Rogro82 | http://pastebin.com/d37b08fda |
20:20.19 | dcordes_ | rootdelay=5 ? |
20:20.20 | dwaradzyn | Rogro82: if you put those messages in order they were shown on screen then it is clearly the case that sd needs more time to init |
20:20.25 | NetRipper | cr2, hi |
20:20.38 | Rogro82 | yeah already tried rootdelay |
20:20.48 | cr2 | NetRipper: do you have working usb ? |
20:21.07 | Rogro82 | also rootwait |
20:21.15 | dwaradzyn | Rogro82: can you pastebin your default.txt? |
20:21.25 | NetRipper | cr2, yes |
20:21.30 | *** join/#htc-linux kimhoon (n=kimhoon@s559116c1.adsl.wanadoo.nl) |
20:21.33 | NetRipper | cr2, i have a few dmesg's on my open dir |
20:21.51 | cr2 | NetRipper: can you switch to netchip/rndis ids ? |
20:22.32 | NetRipper | what are the id's? and how does netchip/rndis work (wouldn't the phone need dhcpd running/) |
20:22.38 | NetRipper | /=? |
20:23.02 | cr2 | 0x0525 / 0xa4a2 |
20:23.34 | cr2 | no, it's standard cdc_ether (with rndis extension) like implemented on pxa |
20:23.43 | Rogro82 | http://pastebin.com/d477c0ab1 |
20:24.37 | cr2 | the cpq/itsy is a very antiquated way of doing cdc_ether |
20:24.41 | NetRipper | cr2, i think rndis is not implemented |
20:24.53 | dwaradzyn | Rogro82: i see it now. rootdelay and rootwait won't work for you as there is no root=/dev/... specified |
20:25.07 | cr2 | then use 0x0525 / 0xa4a1 |
20:25.16 | dwaradzyn | Rogro82: you seem to be trying to boot android 1.0 from vogue, right? |
20:25.30 | Rogro82 | already tried root=/dev. |
20:25.33 | dcordes_ | dwaradzyn, to have an ugly but working hack, how would I run msm_update_screen() in a loop and what interval should I use? or do you know a better way? |
20:25.52 | Bushman | is there any mips linux related channel? |
20:27.02 | NetRipper | cr2, for my info, what's wrong with 0x049F/0x505A? is that the antique way you mentioned? |
20:27.06 | dwaradzyn | dcordes_: i think polaris does have a infinite refresh loop, you can even enable fps counter in vogue_ts |
20:27.14 | Bushman | nevermind, found one |
20:27.32 | tmzt_ | cr2: any idea on msm_fb to make it refresh in X? |
20:28.02 | dwaradzyn | dcordes_: it was about 60 fps all the time i checked it last times. you can safely use 20ms interval |
20:28.09 | cr2 | NetRipper: i think it's the protocol used on zaurus sl5500, and there are some quirks |
20:28.17 | Rogro82 | k gotta run |
20:28.19 | NetRipper | cr2, aha |
20:28.28 | cr2 | NetRipper: netchip is the standard cdc_ether way |
20:28.30 | NetRipper | cr2, ok i'll test the 0x0525 / 0xa4a2 |
20:28.37 | cr2 | ok |
20:28.54 | dcordes_ | dwaradzyn, can you give me an 'example' ? :) |
20:29.51 | cr2 | tmzt_: i don't get it why the fb driver does not autorefresh. seems as a really awkward way of doing things to me |
20:29.57 | dwaradzyn | Rogro82: the mounting in androig 1.0 is done in /init script. it is inside initrd.gz. in order to make it work you should edit this script and change "sleep 1" line to "sleep 5" |
20:30.40 | dcordes_ | obviously android rendering mechanism sends ioctls or syscalls or whatever to achieve the update? |
20:31.12 | NetRipper | cr2, someone mentioned the fbdev mmaps the framebuffer.. which wouldn't include dma'ing the mddi.. couldn't that explain why X isn't refreshed? |
20:31.20 | cr2 | then android is a fucked up piece of software ;) |
20:31.21 | dcordes_ | and msm_fb.c is ignoring anything but android gui |
20:31.56 | dwaradzyn | dcordes_: try doing something similiar to tsts_thread with infinite loop and wait inside |
20:32.06 | cr2 | NetRipper: but why should the app developer take care of the hw refresh ? |
20:32.28 | cr2 | i don't understand what you can win by controllig it in software |
20:32.34 | dcordes_ | dwaradzyn, should this really stay in the touchscreen function? |
20:32.36 | NetRipper | cr2, performance reasons maybe? and it doesn't take care of hw refresh.. it would just update the fb dev memory |
20:32.58 | dwaradzyn | dcordes_: i have no idea... |
20:33.05 | cr2 | NetRipper: i suspect that there is some evil DRM conspiracy behind it. |
20:33.35 | dcordes_ | dwaradzyn, we should put something that works in msm_fb.c and enable it with a config MSM_FB_NON-ANDROID |
20:33.51 | NetRipper | cr2, that's a pretty wild thought |
20:34.00 | dcordes_ | or rather the other way around, excluded it for android configs and use it in all the htc msm kernels |
20:34.22 | cr2 | NetRipper: they name DRM as one of the resons for developing MDDI |
20:34.31 | cr2 | NetRipper: they=qualcomm |
20:34.40 | dcordes_ | would a loop calling the refresh drain any performance or is the operation tiny enough? |
20:34.48 | dwaradzyn | dcordes_: second solution seems better to me |
20:35.08 | dwaradzyn | dcordes_: 60fps takes 0.5% of cpu AFAIR |
20:35.56 | dwaradzyn | dcordes_: but i think it uses dma, so there could be some performance drain not visible in top |
20:36.19 | dcordes_ | ok. if I want to put the 20ms loop in msm_fb.c must I do an extra function or do I just put it in the prsent msm_update_screen ? |
20:36.20 | NetRipper | cr2, but, then it must be some sort of hardware limit? as the drivers so far are fully open source.. in which case we'd be able to remove certain checks |
20:37.14 | dcordes_ | cam driver isn't |
20:37.38 | NetRipper | yes but we're not using cam driver yet |
20:37.41 | NetRipper | nor wifi |
20:37.44 | dwaradzyn | dcordes_: i never did something like that before. i was planning to do it this weekend. have not started yet |
20:37.51 | dcordes_ | wifi only has the firmware |
20:38.08 | cr2 | NetRipper: they are so opensource, that the people can't thge palette issue for more that a month now |
20:38.53 | NetRipper | cr2, that's also a matter of experience and available time.. i mean the palette issue is just cmda.. it's not that big a community |
20:38.56 | Bally3 | rog di you get a black screen freeze? |
20:39.38 | cr2 | it's just good luk that the g1 driver works on raph100 as is |
20:40.10 | cr2 | NetRipper: blackstone too, afaik |
20:40.15 | dwaradzyn | Bally3: rogro needs to modify initrd.gz as i did |
20:40.31 | NetRipper | cr2, blackstone and raph800 have the same panel.. |
20:40.44 | NetRipper | cr2, at least, same panel id is detected |
20:41.08 | cr2 | ok, but it does not really mean anything |
20:41.15 | Bally3 | so did either of you get a frozen black screen or di you get to a prompt? |
20:41.21 | cr2 | it's not just some $random 64bit uuid |
20:41.30 | NetRipper | cr2, using your id's "unsupported MDLM descriptors" |
20:41.45 | cr2 | a4a1 ? |
20:41.55 | NetRipper | <PROTECTED> |
20:41.55 | NetRipper | <PROTECTED> |
20:41.58 | NetRipper | a4a2 |
20:42.01 | cr2 | i think unsupported MDLM descriptors are rndis |
20:42.10 | NetRipper | it doesn't create a usb0 |
20:42.26 | NetRipper | want me to try a4a1? |
20:42.36 | cr2 | lsusb ? |
20:42.36 | cr2 | yes |
20:43.09 | cr2 | i'll check the pxa source |
20:43.17 | NetRipper | ok |
20:43.46 | NetRipper | cr2, do you have any hints on where to start looking to solve mmc problem? |
20:44.19 | dcordes_ | NetRipper, I would test trout-mmc |
20:44.26 | NetRipper | i have, no luck |
20:44.40 | NetRipper | although i can't fully boot trout, i inited the trout-mmc from my halibut board |
20:44.51 | dcordes_ | you built board-trout-mmc* ? |
20:44.54 | NetRipper | yes |
20:45.03 | NetRipper | and inited it |
20:45.11 | NetRipper | does nothing different |
20:45.31 | dwaradzyn | rolk: i checked your commit id. with which zimage built by me did you compare the one built by you? |
20:48.12 | cr2 | */ |
20:48.12 | cr2 | #define CDC_VENDOR_NUM0x0525/* NetChip */ |
20:48.12 | cr2 | #define CDC_PRODUCT_NUM0xa4a1/* Linux-USB Ethernet Gadget */ |
20:48.30 | cr2 | NetRipper: it's from ether.c gadget source |
20:48.42 | NetRipper | cr2, same message, unsupported mdlm descriptors |
20:48.54 | cr2 | NetRipper: we need to fix the broken clk_ api |
20:49.08 | cr2 | NetRipper: what do you have on the host side ? |
20:49.38 | NetRipper | cr2, host side, mmc or usb? :p |
20:49.46 | cr2 | usb |
20:50.07 | NetRipper | everything as usual |
20:50.11 | NetRipper | no strange messages |
20:50.29 | rolk | dwaradzyn: That would be the one you posted on the 'Polaris' thread on www.xda-developers.com |
20:50.48 | NetRipper | let me try again with the subclass suggested by lavender.t |
20:50.57 | dwaradzyn | rolk: which post# ? |
20:51.02 | *** join/#htc-linux AstainHellbring (n=AstainHe@unaffiliated/astainhellbring) |
20:51.05 | cr2 | ok |
20:51.41 | cr2 | NetRipper: it should be netchip/a4a1. so something else is wrong too. |
20:51.43 | NetRipper | cr2, clk api is proc_comm? |
20:52.03 | cr2 | goes through proc_comm only on g1 |
20:52.12 | rolk | I think it is post #51. |
20:52.17 | cr2 | we need to have direct access |
20:52.51 | dcordes_ | cr2, doesn't dzo perform pm operations using proc_comm ? |
20:53.17 | NetRipper | cr2, ah ok, the kaiser/vogue way? but why do we need the clk api? aren't all clocks initialized by winmo already? i'd like to get it working first - before we add support of switching sd cards (which implies clock changes) |
20:54.04 | dcordes_ | I think you need the mmc clocks either way |
20:54.17 | cr2 | NetRipper: i think wince switches the clock on demand. |
20:54.18 | cr2 | it saves power |
20:54.31 | cr2 | and we have complete mess in the current code |
20:54.35 | NetRipper | ok but since i boot from sd, i'd say sd is pretty active |
20:54.37 | NetRipper | :p |
20:55.02 | dcordes_ | I don't see how the clocks can be mess when ginge has them from the dll |
20:55.24 | cr2 | dcordes_: i'd like to hear the full version first |
20:55.45 | cr2 | because the clock settings between wince and spl differ too. |
20:55.58 | rolk | Anyone here that has an idea of the 'loops per jiffie' value on Polaris? |
20:56.38 | dwaradzyn | rolk: ok i see it clearly now. the patch from post 51 is broken. it just adds full_keyboard.c and does not contain changes to other files (as well as .config). |
20:56.59 | cr2 | NetRipper: btw, do the smem smd channels work ? |
20:57.16 | NetRipper | cr2, only receiving, not send |
20:57.42 | cr2 | dcordes_: yes, dzo uses proc_comm. it needs bettwe docs though. |
20:57.57 | cr2 | NetRipper: so gps works ? |
20:58.01 | cr2 | NetRipper: it's strange that send does not work. |
20:58.17 | NetRipper | cr2, regarding usb, with 0x00 as subclass i don't get the MDLM error anymore, but i don't get usb0 either.. |
20:58.19 | dcordes_ | cr2, unfortunately he didn't use the wiki |
20:58.39 | cr2 | ok |
20:58.40 | NetRipper | cr2, looking at the usb function ether code it's really limited, so maybe it is the minimum |
20:58.57 | dcordes_ | I also find the broken AT send on raph100 strange. on the x1 it works with the trout code both ways |
20:59.02 | NetRipper | CE: hpet increasing min_delta_ns to 15000 nsec |
20:59.03 | cr2 | ok, so leave it as is. |
20:59.11 | cr2 | seems weird to me though-. |
20:59.17 | dwaradzyn | rolk: i see that this sd code is completely time dependent and based on magic timer values. thats a mess. i have no idea what jiffies mean |
20:59.27 | *** join/#htc-linux BabelO (n=fcr@lun34-2-82-238-28-28.fbx.proxad.net) |
20:59.46 | cr2 | NetRipper: what about gps ? |
20:59.56 | NetRipper | cr2, haven't tried gps, i can do that in a bit |
20:59.59 | NetRipper | let me revert usb |
21:00.15 | cr2 | ok |
21:01.17 | NetRipper | i flashed elite RC2 rom but it can be annoying.. sometimes it hangs during boot (known issue) but means i sometimes have to soft reset a few times before it loads again :p |
21:01.27 | cr2 | dcordes_: dzo uses the 3 parameter g1 proc_comm proto |
21:01.53 | cr2 | dcordes_: but wince has more variations |
21:02.06 | cr2 | hehe |
21:02.24 | dcordes_ | yea he explained the 3 parameter thing when I asked him about the battery. btw bat should be easy |
21:02.42 | dcordes_ | I can forward the info from the mail exchange about it if somebody wants to try |
21:02.48 | dcordes_ | it didn't work for me on the kaiser |
21:03.13 | rolk | I'm trying to delay the printk messages using the boot_delay parameter. But for that , I need a loops-per-jiffie value that isn't too far from reality. |
21:03.53 | NetRipper | crap forgot to init gps before loading haret |
21:03.53 | NetRipper | :P |
21:04.24 | cr2 | dcordes_: i've documented the battery in wiki pretty well |
21:04.39 | cr2 | NetRipper: it's an AT command to be sent ? |
21:04.47 | cr2 | hm. it does not work now ;) |
21:04.55 | dcordes_ | cr2, yes I looked the regs up but didn't know where to put |
21:05.20 | dcordes_ | and hat several other problems as always when I destroy code |
21:05.33 | cr2 | ok |
21:05.40 | NetRipper | rolk, msleep()? :) |
21:05.56 | NetRipper | (once irq is init'ed :P) |
21:06.40 | NetRipper | cr2, what's an AT command? to enable gps? but i can't send from linux |
21:07.37 | rolk | Ok. On my Polaris, lpj=1277952. |
21:07.53 | dcordes_ | cr2, I'm currently trying to make msm72* work with the frameworkd witht he moko devs |
21:08.19 | rolk | dwaradzyn: Good work with the full keyboard, by the way! |
21:08.54 | dwaradzyn | rolk: i'm editing that post and there will be correct diff |
21:09.01 | NetRipper | cr2, locks up during boot when gps is enabled |
21:09.13 | NetRipper | freezes |
21:09.34 | cr2 | strange |
21:09.38 | dcordes_ | arm9 watchdog |
21:10.01 | dcordes_ | or the gps channel writes somewhere it destroys something important? |
21:10.34 | NetRipper | wouldn't arm9 watchdog reset the device completely, like soft reset? |
21:10.58 | cr2 | it may do it. if it gots the command |
21:11.12 | cr2 | NetRipper: the ts irq still not show up ? |
21:11.46 | NetRipper | cr2, no.. although i've seen 'touch of jobo' comment it out in his .diff.. not sure why.. maybe he saw it fire |
21:12.05 | NetRipper | or maybe he was just playing with it |
21:12.22 | cr2 | ok |
21:12.24 | tmzt_ | dcordes_: pen down makes sense for your onscreen keyboard, but an X apllication shouldn't need input to refresh the screen |
21:12.33 | cr2 | there are 2 irqs actually. and they are seen in wince |
21:13.00 | NetRipper | cr2, i havent actually bothered with irq anymore |
21:13.14 | dcordes_ | tmzt_, yea I already figured that cause I need to pendown permanently. idk hwo to make the loop |
21:13.43 | cr2 | ok |
21:14.14 | cr2 | NetRipper: do you have a linload with a working usb ? |
21:14.26 | NetRipper | yes |
21:14.28 | NetRipper | its on wiki |
21:14.42 | NetRipper | latest progress entry |
21:14.42 | cr2 | NetRipper: and the small ramdisk. i don't need the nadroid bloatware :) |
21:14.47 | cr2 | ok. |
21:14.47 | NetRipper | http://wiki.xda-developers.com/index.php?pagename=RaphaelLinux#Progress |
21:14.52 | NetRipper | it uses angstrom |
21:14.59 | NetRipper | gives you quite some commands to play with |
21:15.15 | NetRipper | credits to dcordes;p |
21:16.21 | NetRipper | i need more tea, brb |
21:18.36 | dcordes_ | crap I need a new laptop |
21:19.32 | dwaradzyn | rolk: check post 51 again |
21:20.41 | rolk | Ok. |
21:25.09 | dcordes_ | NetRipper, did you try the x11 initrd? |
21:25.45 | tmzt_ | gone? |
21:26.19 | dcordes_ | http://members.home.nl/j0b0/boot-linux-x11.mp4 |
21:27.14 | dwaradzyn | nice quality video :P |
21:27.38 | rolk | dwaradzyn: what is the purpose of the full_keyboard2.diff? It worked already quite good without that patch. I don;t doubt that it is needed, can you elaborate? |
21:28.22 | dcordes_ | dwaradzyn, and nice quality vga display |
21:29.19 | dcordes_ | dwaradzyn, it's the raph100 in the video. It seems refresh works there |
21:29.49 | dwaradzyn | full_keyboard.diff did nothing :( if you apply full_keyboard2.diff and compile you should see completely different keyboard (not the green one) |
21:30.53 | NetRipper | dcordes_, yes, i had a black screen right away.. no cursor.. haven't tried more than once though |
21:31.22 | dcordes_ | try touching where the on screen keyboard is |
21:31.29 | dcordes_ | so you get the refresh |
21:31.49 | NetRipper | it works then? |
21:32.04 | dcordes_ | I'm trying to figure |
21:32.21 | NetRipper | it should update the whole screen, whenever you click anywhere |
21:32.57 | dwaradzyn | dcordes_: when i tried older openomo build back in september i had exatly the same screen output as in this video (qvga of course). |
21:33.50 | NetRipper | i need to apply that msm_fb patch so that onscreen keyboard doesn't flicker anymore in android |
21:34.36 | dwaradzyn | NetRipper: i was trying to fight that flicker issue. is there a patch for it? |
21:34.52 | NetRipper | someone put a diff on diamond linux thread |
21:35.05 | NetRipper | dwaradzyn, but im working on raphael/diamond (not to get things confused) |
21:35.17 | rolk | dwaradzyn: The build completes, and the zImage is still slightly different. Perhaps this can be explained by the fact that I use an arm-none-linux-gnueabi- toolchain that is used in our company. |
21:35.24 | NetRipper | and oi, where'd cr2 go :p |
21:35.40 | NetRipper | who gave him permission to leave ;) |
21:36.14 | rolk | I'd like to break the zImage open, and have a look at the disassembly, but that will have to wait. |
21:36.28 | dwaradzyn | NetRipper: is there a video showing flicker on raph/diam? i want to make sure that its the same issue |
21:37.37 | NetRipper | dwaradzyn, dont think so |
21:38.04 | dwaradzyn | rolk: i use arm-none-linux-gnueabi-, version arm2008-q1 |
21:42.28 | rolk | Ok. I don't know how to check the too |
21:42.33 | rolk | lchain version. |
21:43.11 | *** join/#htc-linux OpnSrc1 (i=harisha1@gateway/tor/x-574f750592c329b2) |
21:43.34 | OpnSrc1 | hello |
21:44.00 | dwaradzyn | rolk: arm-none-linux-gnueabi-gcc --version |
21:44.04 | OpnSrc1 | hi Bally, Dwaradzyn |
21:45.01 | dwaradzyn | rolk: mine is (Sourcery G++ Lite 2008q1-126) 4.2.3 |
21:45.11 | dwaradzyn | OpnSrc1: hi, whats up |
21:45.26 | rolk | arm-none-linux-gnueabi-gcc (CodeSourcery Sourcery G++ Lite 2007q3-51) 4.2.1 |
21:45.26 | rolk | C |
21:45.46 | rolk | It seems like a slightly older toolchain. |
21:45.54 | Bally3 | did you get any feedback about the x64 config line rolk? |
21:46.16 | rolk | trying that now. I think it can be safely disabled. |
21:46.28 | dwaradzyn | Bally3: we should check if other config have that enabled |
21:46.47 | Bally3 | checking vogue one now |
21:47.00 | OpnSrc1 | I've been following xda-developers for a while, I'm upto date. |
21:48.00 | OpnSrc1 | Applying git-diff-2.6.25-170908.diff patch to git://git.android.com/kernel.git results in error. |
21:48.16 | Bally3 | config_RESOURCES_64BIT=y |
21:48.31 | OpnSrc1 | I've built zImages before off of linuxtogo.org repo. |
21:48.41 | NetRipper | OpnSrc1, seeing as it's a patch from 17-09-08, i'd say you need a git from about that date ;) |
21:48.44 | Bally3 | thats it |
21:49.04 | rolk | I'm going to focus on getting the usb device controller on Polaris to work, I want a virtual Ethernet connection. |
21:49.27 | NetRipper | OpnSrc1, and you may need to select the correct branch |
21:49.30 | Bally3 | so that you can debug rolk? |
21:49.31 | OpnSrc1 | Ok, |
21:49.45 | rolk | Correct bally3 |
21:50.02 | OpnSrc1 | I've checked out android-msm-2.6.25 |
21:50.06 | *** join/#htc-linux cr2 (n=konversa@crpl6.physik.uni-wuppertal.de) |
21:50.13 | Bally3 | makes sense.. the errors we get are hard to trace |
21:50.37 | cr2 | NetRipper: back |
21:50.49 | NetRipper | cr2, welcome ;) |
21:50.53 | NetRipper | cr2, tried the linload? |
21:51.17 | cr2 | NetRipper: i don't undertstand why i2c and mmc show their virtual addresses in /proc/iomem |
21:51.21 | Bally3 | can someone post up the polaris config file please?? |
21:51.28 | cr2 | yes, even ssh is working :) |
21:51.43 | cr2 | but there is a lot of google code to be removed :) |
21:52.28 | cr2 | NetRipper: did you change the SD detect gpio ? |
21:52.43 | cr2 | btw, i can't boot off the SD |
21:52.44 | NetRipper | cr2, no, i just manually set it to return 1 |
21:52.51 | NetRipper | cr2, you can't? strange |
21:53.02 | Bally3 | CONFIG_MMC_MSM700XA=y |
21:53.05 | NetRipper | cr2, what happens when you do? |
21:53.22 | cr2 | the keyboard does not show up, and the whole thing hangs after the power sink STUB messages |
21:53.41 | NetRipper | cr2, same as some people on forum are reporting |
21:53.55 | NetRipper | cr2, do you still run stock rom? |
21:54.32 | cr2 | NetRipper: i think much google crap should be removed, and the virt address issue to be fixed first. |
21:54.32 | cr2 | yes |
21:55.02 | Bally3 | whats MTYP? |
21:55.07 | Bally3 | MTYPE even |
21:55.29 | NetRipper | cr2, i've compared it to how other drivers use the phys/virt mem and it works exactly the same |
21:55.44 | dwaradzyn | OpnSrc1: checkout according to post #40 on polaris thread, everything should go cleanly |
21:56.41 | cr2 | NetRipper: cat /proc/iomem on the desktop machine. it's all physical |
21:57.32 | cr2 | dcordes_: can you cat /proc/iomem and pastebin the output ?= |
21:57.36 | Bally3 | thanks dwaradzyn just noticed the confog file :) |
21:57.46 | dwaradzyn | kaiser people: you have tried to enable usb on kaiser, don't you? i dont think its possible to enable usb host on kasier/polaris, but can it work at least as usb client? |
21:57.57 | NetRipper | cr2, i know.. i just havent been able to find the reason |
21:58.02 | cr2 | dwaradzyn: can you pastebin the /proc/iomem ? |
21:58.18 | OpnSrc1 | Dwaradzyn: ok, post #40, let me try that. |
21:58.23 | dwaradzyn | cr2: no problem |
21:58.40 | OpnSrc1 | let me blow away this folder.. & start over again. |
21:58.45 | cr2 | NetRipper: i donb't see an obvoius reason too, but it's wrong to have virt addresses in /proc/iomem |
21:59.41 | tmzt_ | does mddi use internal sram for display memory? |
22:00.19 | Bally3 | CONFIG_MMC_MSM7X00A=y - why is this still set in the polaris config rolk? |
22:00.35 | NetRipper | cr2, i think dcordes did that earlier, kaiser also showed virtual addresses |
22:02.31 | cr2 | NetRipper: then it's a kernel bug |
22:03.48 | dcordes_ | dwaradzyn, I need to find a good solution to call msm_update_screen() first |
22:03.48 | Bally3 | anyone? |
22:04.28 | dcordes_ | dwaradzyn, I think ginge looked at kaiser usb |
22:04.45 | dcordes_ | usb host on the kaiser would be ace |
22:04.48 | rolk | Bally3: It causes the msm_sdcc.c to be included in the kernel build. |
22:05.20 | dwaradzyn | dcordes_: i will try to code periodic update and try that on polaris. i'll tell you tomorrow if i succeeded |
22:05.27 | Bally3 | but someone mentioned our device is not 7200A does that not matter? |
22:05.39 | tmzt_ | I assume when google renamed there files/config options they decided to make it MSM7X00A instead of halibut or 7200 or whatever it was |
22:06.35 | cr2 | good night |
22:06.44 | Bally3 | ok..so thats not an issue then |
22:06.47 | dwaradzyn | cr2: http://pastebin.com/m67e7300d |
22:06.52 | dwaradzyn | too late :P |
22:06.57 | Bally3 | lol |
22:07.33 | dwaradzyn | rolk: can you see prv messages i sent you earlier? |
22:07.40 | NetRipper | dwaradzyn, same problem as we have - msm_i2c and msm_sdcc show virtual addresses |
22:08.15 | dcordes_ | dwaradzyn, hm is it complex to try the perioidic update? I just finished a X image build which I would like to try |
22:08.26 | Bally3 | how do I post mine to pastebin? |
22:09.05 | dcordes_ | either write it to the card or copy it manually or run the command remotely from gprs terminal |
22:09.23 | Bally3 | thaks dcordes |
22:10.20 | dwaradzyn | Bally3: this is how i did it: boot angstrom, cd /media/mmcblk0p1; cat /proc/iomem > iomem.txt; reboot and send the iomem.txt file from sdcard to pastebin using sdcard usb adapter :) |
22:11.30 | Bally3 | but I cant boot angstrom with the image you posted up and thats the one I want to check as Im sure i'm getting different results.. but it hangs after the trace to a black screen |
22:13.01 | *** join/#htc-linux Bally3_ (n=chatzill@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
22:13.13 | dwaradzyn | Bally3: use kernel from post 51, everything should work with that one |
22:13.27 | Bally3 | ok :) |
22:13.57 | OpnSrc1 | Dwaradzyn: Kernel on post 51 is with no patches applied (expect for the KB) right ? |
22:14.24 | Bally3 | did i just leafve or has another bally3 just joined?? :S |
22:15.20 | Bally3 | damn whoever you are change yer name :S |
22:16.14 | dwaradzyn | dcordes_: i will just copy the tsts way of creating threads and see if it does the job of refreshing. i will need to merge that into tsc2003.c later and learn how to make #ifdefs configurable in .config |
22:17.57 | dwaradzyn | OpnSrc1: yes, thats right |
22:19.02 | dcordes_ | asd |
22:19.29 | dcordes_ | oops playing with vnc |
22:20.25 | dcordes_ | dwaradzyn, what's a thread? |
22:22.22 | OpnSrc1 | hi decordes, is there any easy way to recreate the SD card issue? without relying on the failure of 1.0 boot off of SD as an indication? |
22:22.54 | dcordes_ | boot a console initrd and move large files |
22:23.38 | dcordes_ | http://linuxtogo.org/~lgorris/kaiser-bootkit/angstrom-20081127.cpio.gz |
22:24.43 | NetRipper | erm, any c expert |
22:24.43 | NetRipper | vkeyb_landscape = !!(readl(MSM_GPIO2_BASE + 0xC20) & 0x00400000); |
22:24.44 | dwaradzyn | dcordes_: an example of thread is in tsc2003.c, function tsts_thread gets daemonized and it is visible as process (ie in top command) |
22:24.51 | NetRipper | !! is the same as nothing, isn't it? |
22:24.51 | NetRipper | :) |
22:24.56 | NetRipper | as no ! at all i mean |
22:26.07 | dwaradzyn | NetRipper: i think you are right |
22:26.55 | rolk | NetRipper: not really. |
22:27.05 | rolk | !!a means !(!a) |
22:27.30 | rolk | If a is 0, then !a is 1. And !!a is 0 again. |
22:27.45 | rolk | If a is not 0 then !a is 0 and !!a is 1. |
22:28.05 | rolk | So, !!a == 0 only if a == 0 otherwise it is 1. |
22:28.17 | NetRipper | rolk, ah, so it would evaluate to 1, when the result between ( and ) could be 0x123 |
22:28.33 | rolk | !!0x123 == 1 |
22:28.36 | NetRipper | exactly |
22:28.37 | NetRipper | thanks |
22:28.51 | NetRipper | i was afraid it was a typo, but that makes sense :) |
22:31.40 | tmzt_ | it that for the orientation sensor? |
22:31.44 | tmzt_ | is |
22:32.22 | OpnSrc1 | decordes_: thanks, I've angstrom on my kaiser. how big a file would hang-up ? 20 MB file? |
22:33.00 | dcordes_ | yes |
22:33.13 | dcordes_ | and I have no clue why this is the limit |
22:33.54 | tmzt_ | block layer/caching? maybe it triggers larger reads on larger files |
22:46.59 | dwaradzyn | OpnSrc1: try running a script that is running parallel copys. it will die in seconds |
22:47.38 | rolk | Aha! I just found out something interesting. |
22:47.47 | rolk | Maybe this is old news, maybe not. |
22:48.21 | dwaradzyn | rolk: yes? |
22:48.22 | rolk | I've had success booting the kernel version from dwaradzyn and detecting the SD card. |
22:48.35 | rolk | Then, it failed 5 times in a row. |
22:48.54 | rolk | I think I know the 'variable' that influences the success factor. |
22:49.08 | dcordes_ | .. |
22:49.12 | rolk | Note: this is not a stability issue, this is the 'detection issue' |
22:49.22 | rolk | If I access the card from WinMob |
22:49.38 | rolk | prior to starting haret, and booting, it finds the card allright. |
22:50.03 | rolk | If I soft reset, and do not touch the sd card from WinMob and then start haret and boot, it fails. |
22:51.14 | dcordes_ | hm I never ran into problems with detection |
22:51.19 | rolk | So, WinMob leaves the card interface in a proper state, and linux simply does not initialize it, but uses it. |
22:51.37 | rolk | Well, just reporting my findings. |
22:51.43 | dcordes_ | but I have at least haret on the card always |
22:51.52 | rolk | Figures. |
22:52.00 | dcordes_ | so there is always access on boot |
22:52.03 | rolk | Then you access the card from WinMob. |
22:52.40 | dcordes_ | you could easily boil it down booting without the card inserted |
22:53.08 | rolk | ? |
22:56.47 | dwaradzyn | rolk: what exactly do you mean by "accessing the card"? |
22:57.26 | tmzt_ | do we have the right proc_comm's for vreg? for the mmc port |
22:57.30 | rolk | With 'file explorer' listing a directory seems to be enough. |
22:57.40 | tmzt_ | it's different in halibut and trout it seems |
23:02.27 | dwaradzyn | rolk: i always do it before launching linux, i have everything on sd, so it is accessed just before boot |
23:04.50 | rolk | OK. Same thing as with dcordes. |
23:05.18 | rolk | I have haret, default.txt, initrd and zImage all on internal flash... |
23:06.10 | rolk | I was trying to boot from an ext3 partition on the sd card, so I figured I better boot from internal flash storage. |
23:09.35 | tmzt_ | if it's not a SDHC (or they work on yours), make a small boot partition at the beginning usually works |
23:09.45 | tmzt_ | or just access the sd to power it up |
23:17.52 | Bally3 | please wait... extracting data image rm: cannot remove /sdcard/data.img: no such file or directory |
23:18.03 | Bally3 | and I now see the new keyboard |
23:18.58 | Bally3 | I never had that last message before |
23:19.27 | dwaradzyn | Bally3: the rm is not an issue, this message will be there because the init script is written that way |
23:19.27 | rolk | tmzt: what's up with that proc_comm? |
23:22.48 | Bally3 | ok |
23:23.55 | rolk | Ok. That same procedure seems to work as well with the android-msm-htc-2.6.25 code base. |
23:24.03 | *** join/#htc-linux stefan_schmidt (n=stefan@p5B034B1F.dip.t-dialin.net) |
23:25.04 | dcordes_ | dwaradzyn, did the ts calibration work for you in the magician image? |
23:25.06 | dwaradzyn | rolk: nice find |
23:25.17 | dwaradzyn | dcordes_: no it did not |
23:25.39 | dcordes_ | how did you try things in it? |
23:26.31 | dwaradzyn | i saw calibration back in september. this time openmoko started to homescreen without calibration |
23:27.14 | dcordes_ | maybe because /etc/pointercal is present |
23:27.24 | tmzt_ | rolk: I mean I don't think the card itself is being powered up. Look at the halibut and trout versions of board-*-mmc or whatever it's called. the vreg that is used has a different name and might be a different proc_comm (command to the arm9 core through the shared memory area) |
23:27.29 | dcordes_ | yea it has the magician pointercal |
23:28.34 | *** join/#htc-linux timebomb (n=tb@e176105083.adsl.alicedsl.de) |
23:37.09 | rolk | Ok. This seems to be a reliable procedure... First access the card from WinMob before booting linux. That fixes my detection problem for the moment. |
23:37.27 | rolk | I check a few more times. |
23:39.37 | rolk | s/check/checked/ |
23:39.47 | rolk | Yes. |
23:45.19 | *** join/#htc-linux Bally3 (n=chatzill@cpc1-blac6-0-0-cust618.manc.cable.ntl.com) |
23:46.10 | *** join/#htc-linux dwaradzyn (n=chatzill@chello089079197022.chello.pl) |
23:54.17 | rolk | dwaradzyn: I applied your patches to the android branch, and the full keyboard is working there as well. |
23:55.00 | dwaradzyn | rolk: ok |
23:55.14 | dwaradzyn | rolk: but it blinks? |
23:55.36 | dcordes_ | dwaradzyn, did you manage to test updater? |
23:55.54 | rolk | nope, no blinking |
23:56.26 | dwaradzyn | which updater? |
23:57.21 | dcordes_ | fb |
23:57.35 | rolk | I'm 'stress' testing the sdcard now. |
23:57.50 | rolk | Just wrote 20MB of data to partition 3 (ext3). |
23:57.59 | dwaradzyn | dcordes_: not yet |
23:58.14 | dcordes_ | rolk, I just got the mmc0: Transaction timed out error |
23:58.29 | rolk | I'm now mv-ing that 20MB from part. 3 to part. 2 (both ext3). |
23:58.41 | rolk | I don't get any messages, but it's slow. |
23:58.45 | dcordes_ | Aborting DMA operation for MMC |