irclog2html for #htc-linux on 20070316

02:18.48*** join/#htc-linux Hawk||- (n=Hawk@p549AE753.dip.t-dialin.net)
03:04.23*** join/#htc-linux dullard_ (n=jim@adsl-static-1-30.uklinux.net)
04:06.10*** join/#htc-linux Ralith (n=ralith@soggy202.drizzle.com)
05:02.38*** join/#htc-linux RoEn_PC (n=roen@p54A64CB4.dip.t-dialin.net)
06:10.06*** join/#htc-linux _law_ (n=law@mail.stiftadmont.at)
06:48.56*** join/#htc-linux Hawk|- (n=Hawk@p549ae753.dip.t-dialin.net)
06:51.34*** join/#htc-linux pH5 (n=ph5@e178201107.adsl.alicedsl.de)
07:14.45*** join/#htc-linux Hawk||- (n=Hawk@p549AE753.dip.t-dialin.net)
07:56.35*** join/#htc-linux Hawk|- (n=Hawk@p549AE753.dip.t-dialin.net)
08:00.26*** join/#htc-linux dullard_ (n=jim@adsl-static-1-30.uklinux.net)
09:03.31*** join/#htc-linux TeringTuby (n=maarten@55-100-dyn.ipact.nl)
09:51.18*** join/#htc-linux cyrill62 (n=cyrill62@ble59-5-82-233-205-36.fbx.proxad.net)
09:59.39*** join/#htc-linux TBBle (n=tbble@202.55.155.91)
10:14.41*** join/#htc-linux garyr (n=garyr@gateway.ash.thebunker.net)
10:36.01*** join/#htc-linux cr2 (n=konversa@crpl22.physik.uni-wuppertal.de)
10:36.29*** join/#htc-linux cr2 (n=konversa@crpl22.physik.uni-wuppertal.de)
10:45.17TBBleDoes anyone here own http://userpage.fu-berlin.de/~zabel/magician/ ?
10:46.53pH5that'd be me.
10:51.14*** join/#htc-linux psokolovsky (n=psokolov@82.193.98.7)
11:10.28TBBlepH5: Well, I have to say it rocks. Thankyou very much.
11:12.27TBBleJust one question, is it supposed to be that a non-root user can't write to /tmp or in fact save most settings? Or is that a consequence of the way it's set up? I'm pretty much brand new to GPE.
11:13.34garyrHas anyone got linux running in any shape or form on the hermes yet?
11:16.14pH5TBBle: glad to hear that it works for you. But if /tmp (or /var/tmp, rather) doesn't have +t mode set, it's a bug.
11:18.12pH5garyr: last I heard was that cr2 was close to boot the kernel. unfortunately the combined set of low-level hackers who own a hermes and have a lot of free time is ...very small.
11:19.36*** part/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk)
11:22.13garyrBummer
11:26.15*** join/#htc-linux Fossi (n=Fossi@e176105192.adsl.alicedsl.de)
11:31.46cr2hi pH5
11:32.04pH5hey cr2
11:32.08cr2garyr: i've read in the logs that Kevin2 has bought a hermes
11:32.40cr2if we'll join forces then we'll be able to boot linux on hermes.
11:32.52pH5cr2: that's what I just got from the p0 neo's gsm modem:
11:32.57pH5AT-Command Interateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterateraterater
11:33.12*** join/#htc-linux tudenbart (n=willi@xdsl-213-196-221-171.netcologne.de)
11:33.37cr2pH5: you have one ?
11:33.58pH5yup. only I can't get the modem to work :/
11:34.20cr2hmm.
11:41.37TBBlepH5: Bizarre. It didn't seem to have +t set ('touch me' in /var/tmp produced 'permission denied') but I proceeded to reboot the phone (knocked the back cover off) log in as root, and +t was set on /var/tmp. Log in as the user again, and 'touch me' worked.
11:42.37TBBleAnd things that weren't working before (one of the keybaord layouts, gomunicator) work now. (Although gomunicator can't open /dev/ttyS1)
11:44.00pH5TBBle: maybe there is some but in the first boot process, I didn't spend too much time on the userspace parts.
11:45.06TBBleFair 'nuff. I'm thoroughly impressed that such a thing is possible at all. The phone I wanted to upgrade to (SE P990) apparently failed certification in Australia, so now I'm taking the opportunity to learn about Linux-based PDA phones.
11:45.11pH5TBBle: about /dev/ttyS1, you can either add your user to the dialout group in /etc/groups or just change ttyS1 owner or permissions so that your user account can r/w it.
11:45.44pH5"some but" above should read "some bug", obviously :)
11:46.38TBBleYeah, I'm just gonna do the former.
11:46.44TBBleO_O No man command.
11:47.11TBBleHmm. If I'm reading the status page right, the phone can call people, but you can't use the built-in microphone so you'd need a headset?
11:48.20pH5TBBle: no, the built-in microphone works, it is directly connected to the gsm chip.
11:48.46TBBleOh. But the userspace code can't access it then?
11:51.33pH5TBBle: not yet. my local version of the sound driver already can capture noise from the microphone, but for some reason the sound quality is very low.
11:53.57TBBleOK, cool. Hmm, the System Information doodad seems to close immediately.
11:56.02TBBleHmm. I guess to network this thing I need a usb ethernet dongle that has one of those small USB connectors on it... Can it be done via a link to a laptop or anything?
11:58.14pH5TBBle: the magician has no host usb port, so a simple usb cable can be used to connect it to the laptop via g_ether (on the pda) and usbnet/cdc_ether modules (on the laptop).
12:00.15*** join/#htc-linux sadeness_ (n=LamersIn@88.87.88.27)
12:00.39TBBleOh, neat. Off to google I go. ^_^
12:17.50*** join/#htc-linux g3gg0_ (n=g3gg0@ppp-62-245-162-191.dynamic.mnet-online.de)
12:37.20TBBleHmm. This usbnet/g_ether stuff is interesting, thanks for the pointer. ^_^
13:20.51*** join/#htc-linux RoEn_PC (n=roen@p54A64CB4.dip.t-dialin.net)
13:48.15TBBleHmm. g_ether loaded on PDA, usbet and cdc_ether loaded on laptop... Would it be because I'm using a different cable (ie one that came with my USB multi card reader)?
13:55.49TBBleHmm. Lots of 'device descriptor read/64, error -62' when I plug the cable in... I'll have to see if I can score the cable to go with this phone.
15:33.16Kevin2Hi cr2
15:33.38Kevin2I tried booting the hermes.  I put in writes to the framebuffer during the compress / early init stages.
15:34.11Kevin2The phone always seems to come out of uncompress.  However, it seems to fail sometimes before going into setting up the memory pages.
15:35.09Kevin2The fact that it seems to fail randomly would indicate we aren't shutting down some service correctly before booting.
15:35.22Kevin2What that is - I don't know.
15:36.19Kevin2I'm shutting down dma (but it doesn't look like it is in use).  I'm shutting down the usb host controller (which is definitely in use).  And I'm clearing the irq controller (which oddly has a FIQ defined).
15:38.23garyrSounds cool
15:38.23*** join/#htc-linux jeanseb (n=jeanseb@gazypan.dyndns.org)
15:39.08pofKevin2: have you had a look at how SSPL for hermes does it ?
15:39.24pofcr2: glad to see you here again :-)
15:44.48Kevin2pof: Where can I find the source for SSPL?
15:45.39pofKevin2: is not availale, but if you ask Des I think he will kindly provide it
15:46.43pofsend him a PM in xda-dev
15:47.06cr2pof: hi. what is SSPL ?
15:47.15cr2hi Kevin2
15:48.16pofcr2: it's a modified 1.09 SPL for hermes which allows you to bypass signatures when flashing code in hermes
15:48.35cr2pof: i've looked at the hermflasher source, it uses 'long' in the crc32 routines, which is not good. will ceratinly fail on a 64bit machine.
15:48.36cr2ok.
15:48.44pofhave a look here: http://forum.xda-developers.com/showthread.php?t=293651
15:49.36pofor here for technical details: http://forum.xda-developers.com/showthread.php?t=291050  (from comment #21 on...)
15:50.38cr2ok. btw, will it run on the older phones ? imho, not, because of the 'retour' stuff.
15:51.40pofwhat do u mean ?
15:51.51pofthe 'retuoR' stuff is just for radio bootloader & at command interpreter
15:52.31cr2yes, but it is not available on universal afaik.
15:52.44cr2hehe. "main idea is clearly described in readme and code itself is small enough to disassemble for details"
15:55.02pofcr2: i haven't checked universal's bootloader, but I think the 'retuoR' stuff is present in universal's radio too... but anyway.. have no relation with SSPL at all.. (?)
15:55.17pofSSPL is only available for hermes & Trinity at the moment
15:56.33cr2yes, it was just something i wanted to tell you. retuoR is certainly not available in my universal.
15:58.21cr2Kevin2: then more reading of the samsung manual is needed. i'll look at compiling the gpio list first.
15:58.58pofcr2: i don't have it here to check now, but you have to put 1second delay between each letter for it to work
15:59.08pofcheck here http://pof.eslack.org/tmp/flasher.c
15:59.55Kevin2cr2: I scanned through the manual and couldn't find anything that looked like dma.
16:00.05pofcr2: function exitRadioBL()
16:01.09Kevin2It is always possible something is twiddling the memory directly.  It is also possible that the linux fb code isn't correct and Linux is booting okay.  (Though the random halts would make this less likely.)  I also think I need a different cpu-flush cache for hermes in haret, but I don't think that would be the problem.
16:06.44cr2Kevin2: is there a chance to use an IR console ? i'm not that familiar with the s3c34xx capabilities, is it possible for an external device (ATI fb) to access the main DRAM directly ?
16:07.21cr2pof:ok, thanks.
16:08.36Kevin2cr2: Not sure on ir console - it would be helpful.  Does the hermes have an external fb chip?  If so, then yes, I think it is possible for it to alter memory directly - though I'm not sure we've seen that before.
16:14.58cr2Kevin2: hermes does not use s3c24xx-fb. the framebuffer RAM is in the external ATI video chip.
16:21.03cr2Kevin2: what was the latest status on dumping wince MODULES from virtual RAM ?
16:31.49Kevin2cr2: I haven't looked at the modules stuff recently.
16:31.54cr2Kevin2: there is atihwtbl0.txt in \windows. let's hope that w100fb driver is compatible.
16:31.58cr2[SD]
16:31.58cr2"SdClockRate"=dword:312500              // SD Card Initial Clock Rate (Hz)
16:31.59cr2"SdCardDetectMethod"=dword:1            // SD Card Detection Method is initialized to polling
16:31.59cr2[CAMERA]
16:31.59cr2"CameraFlashGpioPin"=dword:0x11         // GPIO17
16:32.29cr2[ASIC]
16:32.30cr2"ExtMemCfg"=dword:5
16:32.30cr2"ExtMemEnable"=dword:1
16:32.30cr2"CoreVoltage"=dword:0x00015000          //1.5V
16:32.30cr2"ChipPhysAddr"=dword:0x10000000
16:33.27cr2so the video capture is on ATI ? like on blueangel and  himalaya.
16:33.43cr2hmm. and the microSD ??
16:43.55cr2http://wiki.xda-developers.com/index.php?pagename=Hermes_Video
16:45.23*** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be)
16:55.10*** join/#htc-linux drbombay[lap] (n=drbombay@pdpc/supporter/active/drbombay)
16:55.57cr2Kevin2: is it the GPIO ? -> 0x56000000   1   I/O Port
16:58.10Kevin2cr2: the gpios are at 0x56000000
17:00.02cr2ok. thanks.
17:00.27cr2the ati driver initializes S3_SPI port. very interesting.
17:02.03cr2and the 0x40000 S3_ clock. do you think it's the touchscreen ? AFAIR hermes uses tsc2046.
17:03.00cr2the WDT and RTC are also referenced.
17:05.33cr2found the S3_ gpio setup.
17:07.09cr2but it should be easier just to trace the LCD powerup/powerdown sequence with haret.
17:12.33cr2uh. some memory locations at the RAM start are hardcoded. +0x9370 and +0x9374.
17:13.51Kevin2That is an interesting possibility - maybe the device isn't locking up - maybe the fb is..
17:19.31cr2Kevin2: did you put the cpu-independent functions from gnu-haret to haret ? like the LCD and backlight controls ?
17:20.23cr2the s3_spi is setup completely with gpios from ati driver.
17:23.32Kevin2cr2: Some stuff was ported, but not all.  I don't remember the lcd/backlight stuff.
17:23.40cr2ok.
17:23.45cr2+0x935c
17:23.55cr2can you dump these 3 addresses ?
17:25.19cr2it seems that 935c is the flashlight status (the flashlight is on the s3_ gpio).
17:25.49cr2i need to learn the gpio layout.
17:26.05cr2port +0x20 from 0x56000000
17:26.39cr2bic 0x3000 ; orr 0x1000
17:29.07cr2not listed here: http://wiki.xda-developers.com/index.php?pagename=HermesGPIO
17:29.07Kevin2HaRET(1)# pd 0x3000935c 4
17:29.07Kevin23000935c | 01000001                            | ....
17:29.07Kevin2HaRET(2)# pd 0x30009370 8
17:29.07Kevin230009370 | 81000000 00000000                   | ........
17:30.27cr281
17:31.08cr2hmm. you probably can't powerdown the lcd with mainline haret.
17:35.06cr2i like ATI. OEMFlashON() function.
17:35.31cr2i2c unused.
17:36.45cr20x48000000   1   Memory control
17:37.02cr2setCSTiming changes this CPU register.
17:37.29cr2fast/slow options.
17:38.51cr2+0x93ae is also LCD setup related.
17:39.18cr2the CAM control is in external dll.
17:41.24cr2brightness is nonobviuos.
17:43.03cr2it seems that we need to look at the S3_WDT setup.
17:46.12*** join/#htc-linux rmoravcik (n=rmoravci@adsl-dyn244.91-127-6.t-com.sk)
17:47.11cr2the atihwtbl0.txt parser is not very interesting.
17:57.04cr2ok, backlight is with the leds.
18:26.00Kevin2Well, I've tried a bunch of things.  Boot up still fails randomly before mmu enabling.
18:30.15*** join/#htc-linux jeanseb (n=jeanseb@gazypan.dyndns.org)
18:33.42cr2the leds are controlled by cpld. not surprising.
18:38.59cr2ADC is referenced by the battery code.
18:39.19cr2and cpld...
18:39.58Kevin2cr2: Are you disassembling things or watching for changes?
18:40.07cr2Kevin2: do you have 0xb4700000 in the "dump mmu" ?
18:40.11cr2disassembling.
18:41.04cr2just basic disassembling to get an idea what subsystems are used by which hardware.
18:42.40cr2it seems i have found where is the ACin and USB detect.
18:44.04cr2they are on cpld.
18:46.44cr2for some reason 0xb4700000 is not in the static memory map.
18:47.07cr2but it's the CPLD gpio block.
18:51.14*** join/#htc-linux Hawk|- (n=Hawk@p549ae753.dip.t-dialin.net)
18:53.46cr2Kevin2: how should we proceed ?
18:55.04Kevin2In what way?  I'd like to see why Linux isn't booting.
18:55.52cr2ok, but how ?
18:56.29Kevin2I was tracing the memory configuration to see that it makes sense.
18:57.26cr2ok. can you tell me where is (virt) 0xb4700000 ?
18:59.20Kevin2You want the virtual address of phys 0xb47..  or you want the phys of that virtual address?
19:00.26cr2physical of this virtual.
19:01.20Kevin2b4700000  | 08000000 | 1MB section | D=0    AP=1
19:02.20Kevin2Which is Bank 1 of the memory controller
19:02.33Kevin2Bank 1 - 16bit, no wait      nGCSn=1,access=2,nOE=1
19:02.49cr2ok. great.
19:04.33cr2so this page is wrong, and just sees the wifi power on CPLD.
19:04.35cr2http://wiki.xda-developers.com/index.php?pagename=HermesMemoryMap
19:05.06cr2Kevin2: can you post the "dump mmu" somewhere ?
19:09.38Kevin2cr2: http://handhelds.org/~koconnor/HTCHermes/dump_mmu.txt
19:11.20cr2thanks.
19:14.47Kevin2cr2: I have different mappings then the wiki you show above.  My 0xb6b00000 is unmapped.
19:16.51cr2ok, but the main question is the wifi location now. s3c24xx does not support pcmcia ?
19:19.18cr2imho, it should be physical 0x20000000
19:20.04cr2i don't see any other sensible mapping.
19:20.58Kevin2Bank 4 doesn't look like it is setup - the memory controller settings are at their defaults.
19:21.13cr2do you have the s3c24xx manual at hand ?
19:21.21Kevin2Yes
19:21.24cr2hmm. is your wifi running ?
19:21.29Kevin2Yes
19:22.55cr2ok. can you tell me what is 0x52000140 ?
19:23.10Kevin2phys or virtual?
19:23.42cr2phys.
19:24.00cr2hehe. s3 ADC is used for isPenDown.
19:24.51cr2i'm confused now.
19:24.53Kevin2HaRET(1)# pd 0x52000140 4
19:24.53Kevin252000140 | 00000000                            | ....
19:25.20cr2Kevin2: in the manual. it should be some UDC register, but which one ?
19:25.33Kevin2FUNC_ADDR_REG Function address register 0x140(L) / 0x143(B)
19:25.46cr2it's always referenced directly.
19:25.50cr2interesting.
19:26.11Kevin2FUNCTION ADDRESS REGISTER (FUNC_ADDR_REG)
19:26.11Kevin2This register maintains the USB device controller address assigned by the host. The Micro Controller Unit (MCU)
19:26.11Kevin2writes the value received through a SET_ADDRESS descriptor to this register. This address is used for the next
19:26.11Kevin2token.
19:26.26cr2why does the touchscreen care about it
19:29.57cr2ooh. sdhc_sc2410 has some ati stuff in it. bad.
19:31.23*** join/#htc-linux Hawk||- (n=Hawk@p549AE753.dip.t-dialin.net)
19:31.52cr2it seems that in spite of the 2410 name, the ATI SD controller is used ;-)
19:39.38Kevin2Hrmm.  Bank 4 might be in use.  It is set from 8bit to 16bit accesses.  However, it still has the default Tacc=14 cycles.
19:40.42cr2imho it is wifi.
19:41.17cr2the touchscreen is not tsc2046, sorry. one puzzle solved. it's something on the S3 ADC port.
19:41.51cr2then what is on the spi ?? keyboard ?
19:42.05cr2what why the LCD driver controls the spi ?
19:43.05cr2the next puzzle: which SD controller is used. >50% probability that it is ATI.
19:43.38cr2PIC16LF72
19:46.02Kevin2Maybe they are both in use?
19:46.26Kevin2SPCON0 shows s3's spi0 interface is on.
19:48.43cr2yes, i think all clocks are on.
19:48.46Kevin2cr2: I'm pretty sure we can get memory tracing on the s3c cpus like we have on the pxa cpus by fiddling with the memory page mappings.  It will be a bitch to program, but it should help find these things out pretty quickly.
19:49.12cr2i'm looking at the SDHC now, will look at the keyboard after it.
19:50.42cr2hmm, S3_SD is also referenced.
19:51.28cr2SD uses S3_DMA. big surprise ;)
19:53.27cr2don't see direct ATI refs. they may go through ahioem.dll
19:54.27cr2"sandisk sd512" is hardcoded into the driver :)
19:56.57cr2"card0 detected", "card1 detected". it seems the driver support(ed) 2 SD cards. on a prototype board ?
20:10.05*** join/#htc-linux Hawk|- (n=Hawk@p549AE753.dip.t-dialin.net)
20:12.56*** join/#htc-linux crashmatrix (n=crashmat@s5590785f.adsl.wanadoo.nl)
20:26.47*** join/#htc-linux Hawk||- (n=Hawk@p549ae753.dip.t-dialin.net)
20:26.58Kevin2I'll be back later.
20:30.02*** join/#htc-linux cyrill62 (n=cyrill62@44.247.6.195.dynamic.adsl.abo.nordnet.fr)
20:51.54*** join/#htc-linux dinokpir (n=dinokpir@ip24-251-17-37.ph.ph.cox.net)
21:04.44dinokpirI am a newbie to linux and am not a developer. I just want to thank you guys for working on xanadux. You guys rocks!!
21:17.06*** join/#htc-linux Hawk|- (n=Hawk@p549AE753.dip.t-dialin.net)
21:38.54*** join/#htc-linux Tjikkun (n=tjikkun@82-204-54-115.dsl.bbeyond.nl)
21:43.40*** join/#htc-linux Hawk|- (n=Hawk@p549AE753.dip.t-dialin.net)
22:00.42*** join/#htc-linux crashmatrix (n=crashmat@s5590785f.adsl.wanadoo.nl)
22:09.18*** join/#htc-linux BoSJo (n=bosjo@45.84-48-72.nextgentel.com)
22:37.14*** join/#htc-linux BoSJo (n=bosjo@45.84-48-72.nextgentel.com)
23:07.16*** part/#htc-linux rmoravcik (n=rmoravci@adsl-dyn244.91-127-6.t-com.sk)
23:35.19*** join/#htc-linux Hawk|- (n=Hawk@p549AE753.dip.t-dialin.net)

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.