| 00:17.06 | *** part/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
| 01:07.15 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
| 02:06.13 | jonpry | crud |
| 03:24.22 | *** join/#htc-linux simonsimons34 (~simonsimo@50.2.8.96) |
| 03:24.49 | *** join/#htc-linux simonsimons34 (~simonsimo@unaffiliated/simonsimons34) |
| 04:11.27 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
| 04:30.30 | *** join/#htc-linux GPFerror (~GPF@75-171-45-229.phnx.qwest.net) |
| 04:30.48 | *** part/#htc-linux GPFerror (~GPF@75-171-45-229.phnx.qwest.net) |
| 04:44.03 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
| 05:05.59 | jonpry | hey detule |
| 05:08.20 | detule | hey jonpry |
| 05:08.54 | detule | got the kernel to compile? |
| 05:08.57 | jonpry | i've almost got this thing to build a V4 image |
| 05:09.26 | jonpry | it built for V2 anyways |
| 05:10.00 | detule | nice how are you building it |
| 05:10.08 | detule | plan is to load it via pil right? |
| 05:10.21 | jonpry | yeah |
| 05:10.26 | jonpry | i patched up head.S |
| 05:10.42 | jonpry | so all it does is write a bunch of memory 0xbbbbaaaa |
| 05:11.23 | jonpry | pil is blowing my mind |
| 05:11.40 | jonpry | i'm tempted to try and load it via /dev/mem |
| 05:11.51 | jonpry | maybe use pil to launch it |
| 05:11.56 | detule | how do you jump to it |
| 05:11.57 | detule | oh |
| 05:12.32 | jonpry | the linker script is still all messed up. so only one of the elf sections can actually be loaded |
| 05:14.00 | detule | dev mem could work...you can write one word at a time via that devmem binary in busybox |
| 05:14.20 | jonpry | no dd? |
| 05:14.48 | detule | er....don't think you can dd though i could be wrong |
| 05:15.54 | jonpry | something has to be done to pil anyways |
| 05:16.40 | detule | what config are you using for this, this comet stuff? |
| 05:16.59 | jonpry | i'm about to push a new one |
| 05:17.45 | detule | why V2 image? |
| 05:18.13 | jonpry | i just pushed v4 stuff |
| 05:18.23 | jonpry | stupid kconfig wouldn't build for v4 |
| 05:18.54 | jonpry | compiling for v4 is way slower for some reason |
| 05:19.02 | jonpry | and this doesn't look good |
| 05:19.03 | jonpry | -rwxr-xr-x 1 jon jon 37568357 Feb 10 00:16 vmlinux |
| 05:19.38 | detule | damn i guess explains the slower |
| 05:23.01 | detule | this . = ALIGN(_PAGE_SIZE); is not needed? |
| 05:23.25 | detule | or it doesn't matter at c* |
| 05:23.44 | jonpry | that address has 1GB alignment |
| 05:29.47 | jonpry | hmm |
| 05:29.53 | jonpry | V2 isn't any smaller |
| 05:31.11 | detule | do we need any kind of compression in config |
| 05:32.06 | jonpry | i don't think so |
| 05:32.25 | jonpry | still need all this memory to be able to run it |
| 05:32.52 | detule | write a decompression in pil :) |
| 05:33.56 | detule | perhaps it's a toolchain issue |
| 05:34.02 | jonpry | ah |
| 05:34.03 | detule | some strip flag |
| 05:34.06 | jonpry | gotta run strip on it |
| 05:34.08 | jonpry | 3MB |
| 05:34.47 | detule | so you'll feed this zImage or whatever it is to PIL? |
| 05:34.52 | detule | and hope for the best? |
| 05:35.25 | jonpry | i think it requires a program |
| 05:35.39 | jonpry | to open up the elf, find the .init.text segment |
| 05:35.48 | jonpry | load to 0x8da0 |
| 05:36.03 | jonpry | then touch some pil thing |
| 05:36.20 | jonpry | and then hope for the best |
| 05:38.07 | jonpry | build is like make ARCH=hexagon v4_defconfig |
| 05:38.10 | jonpry | make ARCH=hexagon CROSS_COMPILE=/opt/gnu/bin/hexagon- |
| 05:38.26 | jonpry | then hexagon-strip vmlinux |
| 05:39.09 | detule | great thanks |
| 05:39.34 | jonpry | i'll be amazed if it ever jumps into head.S |
| 05:40.07 | jonpry | between the fake compilers and the ports to virtual machines |
| 05:42.27 | detule | pil_q6v4_init_image looks like what we need just to program start_addr |
| 05:42.41 | detule | then pil_q6v4_reset |
| 05:43.00 | jonpry | yeah some sort of interface to that code would work |
| 05:43.02 | detule | programs boot address from start_addr, and powers up |
| 05:43.10 | detule | by bringing regulators online |
| 05:48.11 | jonpry | pil_q6v4_reset seems to do most of it |
| 05:48.52 | detule | that n4 is running 3.4 right |
| 05:49.05 | jonpry | yeah |
| 05:49.56 | jonpry | i disabled SND_SOC in config and it appears that android never tries to operate the PIL or anything |
| 05:51.09 | jonpry | might be a whole different animal on 8960 though |
| 05:51.26 | jonpry | with q6v4 based modem and apps dsp |
| 05:54.38 | jonpry | gotta sleep. cya |
| 05:54.54 | detule | later |
| 06:09.39 | detule | jonpry, we should be able to use load_image from peripheral-loader.c it loads from userspace (just need a matching firmware name to device name in /lib/firmware/) unpacks elf header, calls init_image, writes to physical memory and calls reset |
| 07:37.36 | *** join/#htc-linux Funklord (~cow@84-55-99-121.customers.ownit.se) |
| 07:49.49 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
| 08:32.14 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
| 08:51.37 | *** join/#htc-linux eR^zeRa` (~zzeratul@ip-84-42-202-42.net.upcbroadband.cz) |
| 09:14.31 | *** join/#htc-linux Rajko (~rajkosto@wan.rajkonet.info) |
| 09:17.07 | *** join/#htc-linux helicopter88 (~helicopte@host167-115-dynamic.23-79-r.retail.telecomitalia.it) |
| 10:28.08 | *** join/#htc-linux arif-ali (~arif-ali@94-192-24-56.zone6.bethere.co.uk) |
| 10:44.01 | *** join/#htc-linux BabelO (~fcr@unaffiliated/babelo) |
| 10:49.44 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
| 11:51.47 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi) |
| 12:24.51 | *** join/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
| 12:26.18 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi) |
| 12:34.41 | Cotulla | hello |
| 12:41.22 | Cotulla | jonpry ping |
| 12:55.45 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
| 13:08.10 | jonpry | Cotulla, pong |
| 13:09.11 | Cotulla | how is it? |
| 13:09.13 | Cotulla | got it working? |
| 13:09.22 | jonpry | i got it to build |
| 13:09.40 | Cotulla | good |
| 13:09.47 | Cotulla | run? |
| 13:10.03 | jonpry | gotta mess with the PIL |
| 13:11.02 | Cotulla | hehe |
| 13:11.06 | Cotulla | I got it working |
| 13:11.12 | Cotulla | qMAGLDR runs with SW MMU |
| 13:11.14 | jonpry | or i could overwrite stuff in /lib/firmware |
| 13:11.18 | jonpry | orly |
| 13:11.22 | Cotulla | yes |
| 13:11.32 | Cotulla | I used same format like in HVM |
| 13:11.42 | Cotulla | but assumed that L2 is always 4096 bytes aligned however |
| 13:11.58 | Cotulla | it's with 6 standalone pagetables, no kernel one yet |
| 13:12.36 | jonpry | 4MB per table? |
| 13:12.45 | Cotulla | yes |
| 13:12.59 | Cotulla | 4K table |
| 13:13.06 | Cotulla | describes 4M |
| 13:13.31 | jonpry | the hole on 8x60 isn't big enough for the page tables |
| 13:13.42 | Cotulla | why |
| 13:13.47 | Cotulla | primary is 4K |
| 13:15.26 | jonpry | isn't l2 1024 pages? |
| 13:16.20 | Cotulla | it can describe 1024 pages |
| 13:16.24 | Cotulla | but size is 1024 * 4 |
| 13:16.27 | Cotulla | 4K too |
| 13:16.38 | Cotulla | each L2 table can describe 4M virtual space |
| 13:16.48 | jonpry | 4GB of VA |
| 13:17.05 | jonpry | <PROTECTED> |
| 13:17.16 | jonpry | = 4MB per thread |
| 13:17.21 | jonpry | = 24MB |
| 13:17.48 | Cotulla | well |
| 13:17.55 | Cotulla | where do u see 4GB VA used? :D |
| 13:18.07 | jonpry | i don't |
| 13:18.11 | Cotulla | so it's much more smaller |
| 13:18.18 | Cotulla | it's common, how ARM system works too |
| 13:18.24 | jonpry | hmm |
| 13:18.33 | jonpry | i thought the pagetables were complete |
| 13:18.36 | Cotulla | no |
| 13:18.42 | Cotulla | when u need a new entry |
| 13:18.46 | Cotulla | u just allocate L2 page table |
| 13:18.55 | jonpry | so access is just like PA = PTE[VA/SZ_4K] |
| 13:18.56 | Cotulla | and fill pointer inside L1 table |
| 13:19.25 | Cotulla | it's not necessary to have all L2 page entries filled |
| 13:19.29 | jonpry | ic |
| 13:19.30 | Cotulla | *L1 |
| 13:19.40 | Cotulla | they actually added when u need mapping |
| 13:20.24 | jonpry | good work |
| 13:20.53 | jonpry | i guess hexagon linux is getting close |
| 13:20.58 | jonpry | *closer |
| 13:21.03 | Cotulla | yea |
| 13:21.07 | Cotulla | maybe will run today |
| 13:21.10 | Cotulla | who knows haha |
| 13:21.31 | jonpry | i think we need a bootloader of sorts |
| 13:21.38 | jonpry | maybe built into head.S |
| 13:21.48 | Cotulla | sorts? |
| 13:22.04 | jonpry | like .text has to be linked at 0xc000 0000 |
| 13:22.16 | jonpry | but loaded at totally different address |
| 13:22.24 | Cotulla | VA is 0xC0000000 |
| 13:22.29 | jonpry | yeah |
| 13:22.35 | Cotulla | PA øû different |
| 13:22.37 | Cotulla | PA is different |
| 13:22.51 | jonpry | the thing with a real bootloader is that it loads the kernel at the wrong address |
| 13:23.06 | jonpry | and then uses atags or something similar to tell head.S what happened |
| 13:23.21 | Cotulla | hm |
| 13:23.23 | Cotulla | atags |
| 13:23.29 | Cotulla | wait |
| 13:23.32 | Cotulla | why address is wrong? |
| 13:23.36 | jonpry | PA |
| 13:23.44 | Cotulla | this kernel just is not designed to be run by native |
| 13:23.45 | jonpry | no mmu on load |
| 13:23.47 | Cotulla | only inside HVM |
| 13:23.54 | Cotulla | wait |
| 13:23.57 | Cotulla | head.S |
| 13:24.01 | Cotulla | adds a special mapping |
| 13:24.04 | Cotulla | to switch mode |
| 13:24.24 | jonpry | the head.S is easier |
| 13:24.34 | jonpry | the problem is PIL |
| 13:24.41 | jonpry | it reads ELF |
| 13:24.44 | Cotulla | yes |
| 13:24.48 | jonpry | and is trying to load crap to VA |
| 13:25.11 | jonpry | i was thinking of modifying the ELF |
| 13:25.23 | jonpry | VA->PA after compile |
| 13:26.46 | Cotulla | it should be special ELF actually |
| 13:26.49 | Cotulla | not program |
| 13:26.54 | Cotulla | but with physical regions |
| 13:27.14 | Cotulla | there are two catalogs inside ELF |
| 13:27.19 | Cotulla | program sections |
| 13:27.20 | Cotulla | and |
| 13:27.24 | Cotulla | physical sections |
| 13:28.05 | jonpry | http://pastebin.com/xEjkiECW |
| 13:28.50 | Cotulla | dump original ELF maybe |
| 13:28.52 | Cotulla | also |
| 13:29.00 | Cotulla | or make flat kernel |
| 13:29.08 | jonpry | 0x8da0 stuff and IMEM |
| 13:29.09 | Cotulla | it should be the best |
| 13:29.18 | Cotulla | cuz same experiene like with ARM |
| 13:29.24 | Cotulla | *experience |
| 13:29.42 | jonpry | i think i can make this work |
| 13:29.55 | Cotulla | ok |
| 13:35.45 | jonpry | did you see my head.S? |
| 13:36.25 | Cotulla | no |
| 13:36.34 | Cotulla | I didn't join yet to github was busy with code :P |
| 13:36.47 | Cotulla | around 400 assembler lines :O |
| 13:37.49 | jonpry | https://github.com/detule/linux-hexagon/commit/12a26cdf01b04fbebfc59f20eba1e59fde114d08 |
| 13:38.05 | jonpry | i can't even tell if that will run |
| 13:38.15 | jonpry | my assembler sucks |
| 13:39.07 | Cotulla | should work in theory |
| 13:39.12 | Cotulla | but is it PA? |
| 13:39.21 | jonpry | yeah |
| 13:39.28 | jonpry | that is the entry point |
| 13:39.50 | jonpry | hmm. i think memw is supposed to be after 1: |
| 13:41.39 | Cotulla | yea |
| 13:47.00 | Cotulla | head.S should be replaced almost fully I think |
| 13:48.16 | Cotulla | interesting that |
| 13:48.33 | Cotulla | instruction packet with stack space allocate fails if stack is uncached |
| 13:48.46 | Cotulla | it didn't work until I changed caching type |
| 13:49.53 | jonpry | is it required to packet the instructions? |
| 13:49.59 | Cotulla | not really |
| 13:50.03 | Cotulla | but it's speed up |
| 13:50.12 | Cotulla | my current tlbmiss code without packets yet |
| 13:50.13 | jonpry | assembler does it automatically? |
| 13:50.39 | Cotulla | no |
| 13:50.47 | Cotulla | u need manually do that by { } |
| 13:50.56 | Cotulla | and it's not really easy |
| 13:51.36 | jonpry | just one instruction per packet without { }? |
| 13:52.13 | Cotulla | yes |
| 13:52.21 | Cotulla | but it's not about used soace |
| 13:52.22 | Cotulla | space |
| 13:52.25 | Cotulla | about speed |
| 13:52.34 | jonpry | yeah VLIW |
| 13:52.41 | jonpry | quad issue |
| 13:52.55 | Cotulla | not all instructions can be issues at the same time |
| 13:53.08 | Cotulla | like two memw() looks not possible |
| 13:53.30 | jonpry | v4 has more possibilities |
| 13:53.36 | Cotulla | maybe |
| 13:53.44 | Cotulla | but I am proud V2 user :P |
| 13:54.16 | Cotulla | okay |
| 13:54.20 | Cotulla | next step |
| 13:54.24 | jonpry | probably aim for portability |
| 13:54.52 | Cotulla | I should clone git I think and start work under kernel |
| 13:55.18 | Cotulla | I am going to make copy of comet_defconfig to htcleo_defconfig |
| 13:55.29 | Cotulla | as well we need |
| 13:55.34 | Cotulla | some CONFIG_ for V2 and V4 |
| 13:55.41 | jonpry | i made one for v4 |
| 13:55.47 | jonpry | you can use comet if you want |
| 13:56.20 | jonpry | there is comet_defconfig and v4_defconfig now |
| 13:56.52 | jonpry | CONFIG_HEXAGON_ARCH_V2 |
| 13:56.53 | jonpry | CONFIG_HEXAGON_ARCH_V4 |
| 13:57.13 | jonpry | it sets up all the compiler options and everything |
| 13:58.41 | Cotulla | okay |
| 14:01.08 | Cotulla | hm is it cat |
| 14:01.21 | Cotulla | or cattyoctopus ? :O |
| 14:01.53 | Cotulla | https://github.com/detule/linux-hexagon ? |
| 14:03.10 | jonpry | git@github.com:detule/linux-hexagon.git |
| 14:03.17 | jonpry | for commit access |
| 14:03.19 | Cotulla | why so plain name |
| 14:03.29 | Cotulla | :D |
| 14:03.44 | Cotulla | hexy-linux much better :P |
| 14:03.55 | jonpry | hex-core-android? |
| 14:04.02 | jonpry | lol |
| 14:04.03 | Cotulla | joke, actually it's confusing is it official or not |
| 14:04.32 | Cotulla | like linux-hexagon@vger.kernel.org |
| 14:05.00 | jonpry | linux-real-hexagon |
| 14:05.17 | Cotulla | kinda yes |
| 14:06.52 | *** join/#htc-linux stroughtonsmith (~steven@86-42-150-7-dynamic.b-ras1.bbh.dublin.eircom.net) |
| 14:07.40 | jonpry | i hope that qcom rage quits |
| 14:08.02 | Cotulla | well |
| 14:08.13 | Cotulla | seems my mail is gone |
| 14:08.18 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
| 14:08.18 | Cotulla | maybe they paid to remove it :D |
| 14:08.23 | Cotulla | lol |
| 14:08.38 | Cotulla | maybe I need register at mail list before send mail? |
| 14:09.14 | jonpry | did you send it to linux-hexagon or just linux ? |
| 14:09.22 | Cotulla | linux-kernel |
| 14:10.26 | Cotulla | detule |
| 14:10.28 | Cotulla | hi |
| 14:13.11 | jonpry | so i guess i need some deal to turn elf into those stupid mdt and b00 things |
| 14:13.26 | Cotulla | shit I am bad with GIT still :( |
| 14:13.52 | Cotulla | but I am still afraid how this cattypus looks to me :O |
| 14:16.33 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
| 14:17.01 | Cotulla | ok I gotta boot linux pc |
| 14:17.03 | Cotulla | and git clone |
| 14:17.05 | Cotulla | and compile |
| 14:17.08 | Cotulla | *try to compile |
| 14:17.41 | Cotulla | btw cover next time "added entry point testing code" by ARCH_V4 |
| 14:20.54 | jonpry | k |
| 14:29.41 | detule | hi folks *waking up* |
| 14:29.46 | detule | nice job Cotulla |
| 14:30.26 | detule | figured out this business of enabling exceptions while handling tlb miss? |
| 14:31.42 | Cotulla | no |
| 14:31.48 | Cotulla | all done inside one tlbmiss |
| 14:31.51 | Cotulla | without nested exceptions |
| 14:32.03 | Cotulla | we discussed that with jonpry above |
| 14:32.16 | Cotulla | it's not possible to use nested exceptions to fetch L2 value |
| 14:34.23 | detule | you made a github account? |
| 14:45.50 | Cotulla | yea |
| 14:46.16 | Cotulla | Cotulla one |
| 14:46.19 | Cotulla | strange |
| 14:46.23 | Cotulla | :D |
| 14:48.29 | detule | you should have commit acc to that repo |
| 14:51.26 | Cotulla | okay |
| 14:51.30 | Cotulla | I hope I won't mess anything |
| 14:51.34 | Cotulla | I am still rather bad with GTI |
| 14:52.27 | detule | just no git push force :p everything else is reversible |
| 14:53.07 | *** join/#htc-linux Marc (~quassel@phpbb/modifications/marc1706) |
| 14:56.04 | Cotulla | okay |
| 14:56.23 | *** join/#htc-linux [Kosh] (~zz@109.205.253.11) |
| 14:56.28 | [Kosh] | hi guys |
| 14:56.42 | detule | jonpry, can't we just memblock remove a larger hole? |
| 14:57.04 | jonpry | or fixup |
| 14:57.44 | [Kosh] | how change user name in Linux? |
| 14:57.52 | [Kosh] | or PC name |
| 14:57.56 | jonpry | i'm trying to convert the vmlinux to something that can be stuck in /firmware |
| 14:58.09 | [Kosh] | what is there? MBN? |
| 14:58.20 | jonpry | and then i guess umount /firmware , mount loop /fimware |
| 14:58.27 | jonpry | echo "get" > pil ? |
| 14:58.47 | jonpry | mdt |
| 14:58.51 | jonpry | +bin |
| 14:58.53 | [Kosh] | can u just map region, copy binary and start Q6 |
| 14:59.23 | jonpry | starting q6 is the trouble |
| 14:59.30 | [Kosh] | why |
| 14:59.33 | jonpry | it will work better if PIL is left alone |
| 14:59.47 | jonpry | cause 8960 needs q6v4 pil to operate the modem |
| 15:00.41 | [Kosh] | oh |
| 15:00.45 | [Kosh] | yes |
| 15:00.59 | jonpry | but that may not work anyways |
| 15:01.05 | jonpry | since modem needs trusted loader |
| 15:01.06 | [Kosh] | should be |
| 15:01.14 | [Kosh] | call to RPM |
| 15:01.16 | [Kosh] | to run them |
| 15:01.51 | jonpry | <PROTECTED> |
| 15:01.51 | jonpry | <PROTECTED> |
| 15:02.08 | jonpry | i think that is digital signature |
| 15:02.16 | [Kosh] | maybe |
| 15:02.21 | [Kosh] | it can be a problem yes] |
| 15:05.11 | [Kosh] | ok git cloner |
| 15:05.18 | [Kosh] | speed is good |
| 15:05.20 | [Kosh] | ;) |
| 15:11.45 | [Kosh] | done |
| 15:12.21 | [Kosh] | ya commit ba9459b9124911b795713b420edce3716af6d28b |
| 15:13.03 | detule | i guess kvm is the latest blow up on lkml |
| 15:13.17 | detule | linus on the receiving end of "you're so full of shit" |
| 15:13.27 | [Kosh] | ? |
| 15:13.40 | [Kosh] | why u hardcoded 0x8da00000?] |
| 15:14.11 | detule | that's where q6 firmware is currently being loaded |
| 15:14.23 | [Kosh] | well for me address is totally different :) |
| 15:15.03 | detule | *on msm8960 and apq8064 or whatever is in the n4 |
| 15:15.10 | [Kosh] | yes |
| 15:15.14 | [Kosh] | but I need another address |
| 15:15.21 | jonpry | yeah needs to be a config option |
| 15:15.25 | [Kosh] | it was PAGE_OFFSET + LOAD_ADDRESS |
| 15:15.25 | jonpry | didn't get to it |
| 15:15.31 | [Kosh] | we should look how it's done in ARM |
| 15:15.34 | jonpry | don't do that |
| 15:15.48 | jonpry | PAGE_OFFSET+LOAD_ADDRESS causes all kind of problems |
| 15:16.20 | jonpry | just CONFIG_PHY_BASE |
| 15:16.23 | jonpry | and be done |
| 15:16.32 | jonpry | *PHYS |
| 15:16.34 | [Kosh] | #define PAGE_OFFSET_AC(0xc0000000, UL) |
| 15:16.40 | [Kosh] | it's kernel VA start |
| 15:16.52 | jonpry | which is no good for init.text |
| 15:17.49 | [Kosh] | why |
| 15:17.52 | [Kosh] | for ARM it's . = PAGE_OFFSET + TEXT_OFFSET; |
| 15:19.14 | [Kosh] | and it's textofs-y:= 0x00008000 |
| 15:19.23 | [Kosh] | maybe we can make it |
| 15:19.32 | [Kosh] | no |
| 15:19.36 | [Kosh] | or yes |
| 15:19.50 | jonpry | i already got it to do "the right thing" |
| 15:19.53 | [Kosh] | space for ATAGs |
| 15:20.00 | [Kosh] | wait |
| 15:20.02 | jonpry | just define 0x8da0 for your layout |
| 15:20.08 | [Kosh] | constant can not be right thing :) |
| 15:20.16 | jonpry | no it needs defined |
| 15:20.25 | jonpry | but the resulting ELF is good |
| 15:20.26 | [Kosh] | wait |
| 15:20.37 | [Kosh] | 0x8da00000 is wrong |
| 15:20.39 | [Kosh] | it's PA |
| 15:20.43 | [Kosh] | and we need VA |
| 15:20.46 | jonpry | init.text is PA |
| 15:20.48 | [Kosh] | 0xC0000000+ there |
| 15:20.49 | jonpry | mmu off |
| 15:21.17 | jonpry | readelf -l vmlinux |
| 15:22.12 | [Kosh] | I am looking |
| 15:22.41 | jonpry | i am fairly good at the trampoline stuff |
| 15:23.04 | jonpry | lot's of kexec experience |
| 15:23.25 | [Kosh] | well |
| 15:24.28 | [Kosh] | it doesn't used for ARM |
| 15:24.36 | jonpry | so like init.text is at PA |
| 15:24.37 | [Kosh] | they just in the start up code use offsets |
| 15:24.49 | [Kosh] | between PA and VA to access _few_ variables |
| 15:24.50 | jonpry | then first order of business is to setup an idmap |
| 15:24.56 | [Kosh] | I think the best way to do same |
| 15:25.02 | jonpry | so VA=PA for 0x8da0 |
| 15:25.03 | [Kosh] | cuz ARM is well tested |
| 15:25.11 | jonpry | then mmu on |
| 15:25.16 | [Kosh] | oh |
| 15:25.18 | [Kosh] | shit |
| 15:25.25 | [Kosh] | I forgot one thing ofcourse |
| 15:25.37 | [Kosh] | unlike ARM very hard to make position independent code |
| 15:25.43 | jonpry | then setup 0xc000 to kernels PA |
| 15:25.46 | [Kosh] | in that case it can be better to use |
| 15:25.49 | [Kosh] | that section maybe |
| 15:26.04 | [Kosh] | so it should be CONFIG_INIT_OFFSET maybe |
| 15:26.08 | jonpry | sure |
| 15:27.00 | [Kosh] | as well "core system" will be mapped to FFF00000 I think |
| 15:27.09 | [Kosh] | we need to add a special section for that into LDS |
| 15:27.26 | [Kosh] | well I am going to compile it now |
| 15:28.30 | Cotulla | well let me try commit something |
| 15:28.34 | Cotulla | like htcleo_defconfig |
| 15:30.21 | Cotulla | hm how to do that |
| 15:30.36 | detule | you have it commited locally? |
| 15:30.47 | jonpry | git add file; git commit -m "added bs"; git push |
| 15:30.48 | Cotulla | not yet. I think I need to add key |
| 15:30.56 | Cotulla | afaik |
| 15:31.06 | detule | right you need your ssh key on github |
| 15:31.23 | detule | then what jonpry said |
| 15:33.25 | [Kosh] | passphrase too short: have 3 bytes, need > 4 |
| 15:33.28 | [Kosh] | lol |
| 15:36.10 | [Kosh] | Hi Cotulla! You've successfully authenticated, but GitHub does not provide shell access. |
| 15:36.11 | [Kosh] | lol |
| 15:36.17 | [Kosh] | looks like end of level in Mario |
| 15:36.19 | [Kosh] | :O |
| 15:36.29 | jonpry | not ssh:// |
| 15:36.35 | jonpry | edit .git/config |
| 15:37.15 | jonpry | ok i finished the vmlinux post processor |
| 15:37.31 | detule | damnit kosh=cotula |
| 15:38.17 | [Kosh] | lol detule? |
| 15:38.24 | [Kosh] | u are soooo fast |
| 15:38.26 | [Kosh] | :D |
| 15:38.44 | [Kosh] | this is from Linux PC |
| 15:38.50 | detule | hey it took you a while to figure out who d3tul3 is |
| 15:39.01 | [Kosh] | lol I was 95% sure |
| 15:39.03 | [Kosh] | just checked |
| 15:39.04 | [Kosh] | :P |
| 15:39.04 | detule | and those two nicks have much more resemblance |
| 15:39.19 | [Kosh] | yea correlation is very close |
| 15:39.36 | [Kosh] | 2 e - > 2 3 |
| 15:39.41 | [Kosh] | len is 6 |
| 15:39.43 | [Kosh] | so |
| 15:39.45 | detule | i was surprised that someone else is jumping in on the project with "you should do this or that" |
| 15:39.45 | [Kosh] | 2 / 6 |
| 15:40.04 | [Kosh] | so |
| 15:40.07 | [Kosh] | 100 / 6 = 16.6 |
| 15:40.12 | [Kosh] | 2 * 16.6 = 32.2 |
| 15:40.19 | [Kosh] | 100 - 32.2 = 67.8 |
| 15:40.29 | [Kosh] | but we need take into account that e and 3 looks rather simular |
| 15:40.32 | [Kosh] | but dunno how to measure that |
| 15:40.49 | [Kosh] | so lets say it's 70% correlation |
| 15:40.50 | jonpry | it's l33t speak |
| 15:40.51 | [Kosh] | :P |
| 15:42.20 | Cotulla | yea |
| 15:42.26 | Cotulla | it's my favorite troll style |
| 15:42.32 | Cotulla | just enter and start to give advises to ppl |
| 15:42.33 | Cotulla | :D |
| 15:42.39 | Cotulla | from air |
| 15:43.29 | detule | i think [Kosh] 's english is better...that threw me off to |
| 15:43.30 | detule | o |
| 15:46.10 | jonpry | should i push this pos to the kernel? |
| 15:51.30 | [Kosh] | pos? |
| 15:51.44 | langpoo | hi guys |
| 15:51.59 | langpoo | and ladies |
| 15:52.26 | jonpry | piece of shit |
| 15:52.46 | jonpry | this horrible program to turn vmlinux into PIL images |
| 15:53.01 | langpoo | Maybe. |
| 15:53.13 | jonpry | i don't really want to make a whole other repo for it |
| 15:53.23 | langpoo | I think I gotta use qImage. |
| 15:53.52 | jonpry | in theory i could wrap a qImage in ELF |
| 15:54.11 | jonpry | but it's like hard to tell what the linker is doing |
| 15:54.44 | detule | maybe start another branch for now just so we can have the commit up there |
| 15:55.09 | *** join/#htc-linux gauner1986 (~Adium@hmbg-5f760146.pool.mediaWays.net) |
| 15:55.29 | gauner1986 | hey |
| 15:55.33 | langpoo | fatal: remote error: |
| 15:55.33 | langpoo | <PROTECTED> |
| 15:55.33 | langpoo | <PROTECTED> |
| 15:55.35 | langpoo | hm |
| 15:55.39 | langpoo | I need change to SSH? |
| 15:55.59 | detule | in .git/config |
| 15:56.03 | detule | change the repo address |
| 15:56.04 | detule | to the one the told you above |
| 15:56.11 | detule | git@ * |
| 15:56.31 | Cotulla | yesterday? |
| 15:56.38 | langpoo | git@github.com:detule/linux-hexagon.git |
| 15:56.40 | langpoo | this one? |
| 15:56.42 | detule | yes |
| 15:57.20 | langpoo | I guess |
| 15:57.20 | langpoo | url = git://github.com/detule/linux-hexagon.git |
| 15:57.21 | langpoo | this> |
| 15:57.22 | langpoo | ? |
| 15:57.57 | jonpry | change that |
| 15:58.05 | detule | yes right under [remote "origin"] |
| 15:59.21 | detule | congratulations |
| 16:00.16 | langpoo | fuck yea |
| 16:00.17 | langpoo | 11 |
| 16:02.04 | langpoo | Shit. I forgot to put a dot at the end of commit sentence. . . |
| 16:02.08 | langpoo | :O |
| 16:02.52 | langpoo | time to compile |
| 16:16.03 | jonpry | it thought about running my firmware |
| 16:16.11 | jonpry | RPM fail |
| 16:16.41 | langpoo | is it possible to ignore changes inside one specific file for git? |
| 16:16.49 | langpoo | kinda my Makefile with hack |
| 16:16.56 | langpoo | and nobody gotta change it more |
| 16:17.24 | jonpry | there is something called git ignore |
| 16:17.55 | Cotulla | but I should able to commit still |
| 16:18.29 | langpoo | compiling |
| 16:19.12 | langpoo | oh I need CONFIG_HEXAGON_ARCH_V2=y |
| 16:21.55 | Cotulla | lol compiled |
| 16:21.57 | Cotulla | strange |
| 16:22.17 | Cotulla | my aura is positive today I think |
| 16:22.46 | *** join/#htc-linux detule (~detule@pool-96-234-132-77.bltmmd.east.verizon.net) |
| 16:23.27 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
| 16:24.40 | detule | Cotulla, add "Makefile" to your .gitignore file in the root of your repo, you will also want to add ".gitignore" to .gitignore |
| 16:24.59 | Cotulla | okay |
| 16:26.05 | langpoo | hm still |
| 16:26.06 | langpoo | #modified: .gitignore |
| 16:26.06 | langpoo | #modified: Makefile |
| 16:26.22 | langpoo | nvm |
| 16:26.43 | langpoo | no it is |
| 16:29.29 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
| 16:30.11 | jonpry | detule, i have q6 images |
| 16:31.08 | langpoo | why arch/hexagon/kernel/vmlinux.lds is untracked now hm] |
| 16:31.09 | langpoo | ? |
| 16:31.26 | jonpry | normal |
| 16:31.34 | langpoo | ok |
| 16:31.37 | jonpry | it is generated from vmlinux.lds.S |
| 16:31.44 | langpoo | oh |
| 16:32.21 | langpoo | I will add CONFIG_MACH_HTCLEO ok? |
| 16:32.33 | langpoo | or how we are going to separate parts? |
| 16:32.42 | langpoo | LEO/N4 and etc |
| 16:32.55 | jonpry | v4/v2? |
| 16:33.04 | jonpry | is there really machine specific stuff? |
| 16:33.06 | langpoo | yes also possible |
| 16:33.14 | langpoo | yes ofcourse |
| 16:33.44 | jonpry | like n4/sgs3? |
| 16:33.54 | langpoo | like memory mapping |
| 16:33.59 | langpoo | for example |
| 16:34.16 | jonpry | yes that is going to be fun |
| 16:34.40 | jonpry | mach types are fine with me |
| 16:34.47 | Cotulla | ok |
| 16:34.53 | Cotulla | lets add CONFIG_MACH_* |
| 16:35.35 | detule | so it's time to see if PIL will play nice with these? |
| 16:35.37 | Cotulla | I think better to put new "real" code to separate folder |
| 16:35.54 | Cotulla | "coresys" for example |
| 16:36.01 | Cotulla | under arch\hexagon\coresys |
| 16:36.10 | jonpry | not mm? |
| 16:36.19 | Cotulla | to separate our files |
| 16:36.22 | jonpry | detule, trusted PIL freaked out on me |
| 16:36.26 | Cotulla | from mainline/QCT/CA |
| 16:36.49 | jonpry | i'm not picky |
| 16:37.02 | Cotulla | ok |
| 16:37.13 | jonpry | at some point i think it should be a patch set against mainline though |
| 16:37.27 | jonpry | try and push that upstream :p |
| 16:37.41 | Cotulla | yea why not |
| 16:37.54 | Cotulla | then we can move working code to MM |
| 16:37.58 | Cotulla | it's not hard |
| 16:38.03 | jonpry | sure |
| 16:41.57 | jonpry | it claims to have booted dsp |
| 16:43.05 | jonpry | so now how do i inspect /dev/mem? |
| 16:43.18 | Cotulla | ioremap? |
| 16:43.28 | jonpry | from userland |
| 16:43.38 | Cotulla | hm no idea |
| 16:47.05 | detule | jonpry, use busybox devmem |
| 16:47.52 | jonpry | i only have toolbox |
| 16:48.53 | detule | you have that rooted? |
| 16:49.06 | jonpry | yeah |
| 16:50.07 | detule | this is what you want https://github.com/CyanogenMod/android_external_busybox/blob/gingerbread/miscutils/devmem.c |
| 16:51.01 | Cotulla | heh |
| 16:51.03 | Cotulla | hehe |
| 16:51.07 | Cotulla | u may need dckill |
| 16:51.42 | detule | need busybox binary? |
| 16:52.24 | jonpry | i found something on play |
| 16:54.02 | jonpry | hmm |
| 16:54.11 | jonpry | the code is in memory |
| 16:54.36 | jonpry | doesn't appear to have run though |
| 16:54.42 | detule | but it didn't clobber? |
| 16:54.46 | Cotulla | maybe no dckill |
| 16:54.47 | jonpry | no |
| 16:56.56 | jonpry | dckill? |
| 16:58.09 | Cotulla | cache clear kinda |
| 16:58.16 | Cotulla | data cache |
| 16:58.45 | jonpry | ah |
| 16:58.56 | jonpry | i think the problem is with my elf stuff |
| 16:59.07 | Cotulla | maybe compiled wrong |
| 16:59.22 | Cotulla | relocations and etc |
| 16:59.34 | jonpry | somehow i can't disassemble it |
| 16:59.43 | Cotulla | use objdump |
| 16:59.50 | jonpry | i tried |
| 16:59.55 | jonpry | but it looks nothing like head.S |
| 16:59.57 | Cotulla | -D ? |
| 17:01.36 | jonpry | http://pastebin.com/KQGyqbcG |
| 17:05.19 | detule | maybe stripped too agressively> |
| 17:05.26 | jonpry | appears to be endian |
| 17:05.39 | jonpry | <PROTECTED> |
| 17:05.39 | jonpry | <PROTECTED> |
| 17:06.33 | Cotulla | yeah some shit |
| 17:07.22 | jonpry | is qMAGLDR littlehexagon? |
| 17:10.45 | jonpry | Cotulla, can you send me a qMAGLDR binary? |
| 17:11.12 | Cotulla | yes it's little |
| 17:11.28 | Cotulla | DSP.bin: file format elf32-littleqdsp6 |
| 17:12.56 | jonpry | typo |
| 17:13.02 | Cotulla | ? |
| 17:13.20 | jonpry | r25.L = #0x0000 |
| 17:13.20 | jonpry | r26.H = #0x8eff |
| 17:13.21 | jonpry | dur |
| 17:13.37 | Cotulla | moll |
| 17:16.22 | *** join/#htc-linux eR^Z` (~zzeratul@ip-84-42-202-42.net.upcbroadband.cz) |
| 17:17.17 | jonpry | hmm |
| 17:17.32 | jonpry | still no boot |
| 17:19.01 | Cotulla | added dckill? |
| 17:19.27 | jonpry | seems like PIL should have what it needs |
| 17:19.40 | jonpry | and /dev/mem is uncached i think |
| 17:20.31 | *** join/#htc-linux arif-ali (~arif-ali@94-192-24-56.zone6.bethere.co.uk) |
| 17:21.24 | Cotulla | but in Q6 it should be cached |
| 17:21.47 | jonpry | yes |
| 17:22.03 | jonpry | thats why i set it to write 16MB |
| 17:23.31 | Cotulla | yea should not |
| 17:32.51 | Cotulla | try to add that at the begining |
| 17:32.52 | Cotulla | http://pastebin.com/fqRWQueF |
| 17:40.50 | Cotulla | working under vm_ops.S implementation |
| 17:43.22 | jonpry | how do you do forward pc relative jump? |
| 17:43.46 | fakker | galaxy nexus is broken |
| 17:44.50 | Cotulla | earth fakker is broken |
| 17:45.00 | Cotulla | jonpry looks like it's not really possible |
| 17:45.05 | Cotulla | u can load pc value to register |
| 17:45.13 | Cotulla | and then add something |
| 17:45.18 | Cotulla | but hard to calc exactly value |
| 17:45.21 | Cotulla | in asm |
| 17:45.30 | fakker | screen is broken - do i wait or do i buy something now? |
| 17:45.41 | Cotulla | wait |
| 17:45.47 | jonpry | Cotulla, jump 1f ? |
| 17:45.59 | Cotulla | it's relative |
| 17:46.01 | Cotulla | oh yes |
| 17:46.09 | Cotulla | 1: |
| 17:46.10 | Cotulla | then |
| 17:50.00 | jonpry | hmm |
| 17:50.02 | jonpry | still nothing |
| 17:50.13 | fakker | wait for what? |
| 17:50.14 | Cotulla | did u add my code? |
| 17:50.20 | jonpry | yes |
| 17:50.24 | Cotulla | fakker for Windows Phone 9 |
| 17:50.33 | Cotulla | hm |
| 17:50.39 | Cotulla | is there old values? |
| 17:50.42 | fakker | stupid russian |
| 17:50.46 | Cotulla | :D |
| 17:50.59 | Cotulla | u didn't overwrite code? |
| 17:51.37 | jonpry | something is wrong with the disassembly |
| 17:55.08 | Cotulla | ? |
| 17:55.18 | Cotulla | are u sure u are compiling for V4? |
| 17:55.40 | jonpry | yes |
| 17:57.26 | jonpry | http://pastebin.com/ky414nwh |
| 17:58.32 | Cotulla | there is also something called chicken |
| 17:59.18 | Cotulla | it inited by 00 20 20 00 value |
| 18:00.26 | Cotulla | p0 = cmp.gtu (r25, r24) |
| 18:00.42 | Cotulla | oh it's reversed compare |
| 18:01.09 | jonpry | q6 init is simple on 8x60 |
| 18:01.37 | Cotulla | there is V3? |
| 18:01.48 | jonpry | http://pastebin.com/SCB0t8zv |
| 18:02.13 | jonpry | it looks to me like 0x40 is setting up the c stack |
| 18:02.33 | Cotulla | stack is r29 |
| 18:03.03 | jonpry | r26 is 0xc000 0000 |
| 18:03.33 | jonpry | r27 is like end of .text |
| 18:03.41 | Cotulla | maybe try to modify place where code is located |
| 18:04.02 | Cotulla | 8DA |
| 18:04.11 | jonpry | i figure RPM is messing with it |
| 18:04.29 | jonpry | like running the stock q6 images with untrusted boot doesn't work either |
| 18:04.38 | Cotulla | ? |
| 18:04.41 | Cotulla | also problem? |
| 18:04.53 | jonpry | doesn't do anything |
| 18:04.59 | jonpry | like it doesn't come out of reset |
| 18:05.01 | Cotulla | hm |
| 18:05.07 | Cotulla | so ur code maybe not executed |
| 18:05.08 | Cotulla | ? |
| 18:06.30 | jonpry | thats what i think |
| 18:06.58 | Cotulla | bad |
| 18:07.26 | jonpry | somehow you are able to boot it |
| 18:07.54 | Cotulla | ? |
| 18:08.27 | jonpry | without RPM |
| 18:09.02 | Cotulla | how? |
| 18:10.17 | jonpry | i'm not sure |
| 18:10.30 | jonpry | qMAGLDR works though |
| 18:10.49 | Cotulla | it's QSD8250B |
| 18:10.56 | Cotulla | no RPC there at all |
| 18:11.03 | Cotulla | RPM |
| 18:11.15 | jonpry | oh |
| 18:11.23 | Cotulla | sadly otherwise rMAGLDR could be cool to have :) |
| 18:12.23 | Cotulla | it can also use TZ |
| 18:17.47 | Cotulla | yes I think PIL is using TZ services not RPM |
| 18:19.03 | jonpry | SCM? |
| 18:19.13 | Cotulla | what is it? |
| 18:19.32 | jonpry | some kind of ROM code i think |
| 18:20.28 | Cotulla | hm |
| 18:20.40 | jonpry | i think this PIL stuff is just all wrong |
| 18:20.42 | Cotulla | Secure Channel Manager seems |
| 18:20.51 | jonpry | stale code |
| 18:26.05 | Cotulla | why |
| 18:26.18 | Cotulla | it's working every boot] |
| 18:26.39 | jonpry | working using the trusted code path |
| 18:26.53 | jonpry | CA has much different stuff now |
| 18:27.43 | detule | maybe i should've tried it |
| 18:27.52 | detule | with my spanking new kernel |
| 18:28.38 | jonpry | should i push this new head.S? |
| 18:29.02 | Cotulla | with define? |
| 18:29.04 | *** join/#htc-linux dsdsfds (511253ef@gateway/web/freenode/ip.81.18.83.239) |
| 18:29.12 | Cotulla | with #ifdef |
| 18:29.20 | jonpry | you don't need ssr=0? |
| 18:29.40 | Cotulla | need |
| 18:29.55 | Cotulla | but that can be done separetely |
| 18:30.03 | Cotulla | like two include files |
| 18:30.06 | Cotulla | v2 v4 |
| 18:31.17 | jonpry | i pushed something |
| 18:31.37 | jonpry | detule, so like if you want to try this, build the vmlinux |
| 18:32.01 | jonpry | then g++ arch/hexagon/tools/splitter.cpp |
| 18:32.04 | jonpry | ./a.out |
| 18:32.10 | *** join/#htc-linux arif-ali (~arif-ali@94-192-24-56.zone6.bethere.co.uk) |
| 18:32.15 | jonpry | then you have q6.mdt and b00 |
| 18:32.21 | jonpry | umm |
| 18:32.42 | Cotulla | hm |
| 18:32.46 | jonpry | then i like modified the kernel so it has no sound support and no q6 support except pil |
| 18:32.49 | Cotulla | do u have rw access |
| 18:33.05 | Cotulla | to that area jonpry? |
| 18:33.09 | Cotulla | or only ro |
| 18:33.17 | jonpry | for firmware? |
| 18:33.23 | Cotulla | for QDSP6 memory area |
| 18:33.31 | jonpry | rw |
| 18:33.37 | jonpry | i read back /dev/mem |
| 18:33.41 | jonpry | code is loaded |
| 18:33.47 | Cotulla | not via PIL |
| 18:33.52 | Cotulla | but via usual access? |
| 18:33.58 | jonpry | i believe so |
| 18:34.19 | Cotulla | hm |
| 18:34.32 | Cotulla | can u access all registers? |
| 18:35.04 | jonpry | i think so |
| 18:35.36 | jonpry | detule, and then i patched pil-q6v4 with something like if (0){//pas_supported(pdata->pas_id) > 0) { |
| 18:37.01 | jonpry | i think the nontrusted pil support was developed for mss |
| 18:37.30 | jonpry | new terminology here. modem q6 = mss, apps q6 = lpass |
| 18:37.50 | detule | low power audio sub-sys |
| 18:37.56 | detule | modem sub-sys |
| 18:38.03 | jonpry | seems like my PIL is doing modem stuff |
| 18:38.24 | jonpry | so the reset vector is all fucked, and the clocks, the whole thing is wrong |
| 18:38.48 | Cotulla | yes mss and lpass |
| 18:38.59 | detule | how do you ask the kernel to bring q6 online |
| 18:39.15 | jonpry | <PROTECTED> |
| 18:39.22 | Cotulla | jonpry afaik pil has load and run as separate functions |
| 18:39.23 | jonpry | echo "put" > q6 to turn off |
| 18:39.26 | jonpry | get to turn on |
| 18:39.36 | detule | thanks |
| 18:39.54 | detule | i'll try it tonight...gotta get some work done this afternoon and you two aren't helping |
| 18:40.09 | detule | by being all constructive and what not |
| 18:40.55 | jonpry | i think the trick is to get the stock firmware to work with that trusted crap turned off |
| 18:42.42 | Cotulla | I think it copies stuff |
| 18:42.46 | Cotulla | then checks signature |
| 18:42.48 | Cotulla | and bailout |
| 18:49.29 | Cotulla | do u have access to 0x28800000? |
| 18:50.41 | jonpry | non trusted doesn't use anything but arm |
| 18:50.43 | jonpry | and checks nothing |
| 18:50.56 | jonpry | but it tries to boot the wrong processor |
| 18:51.05 | Cotulla | why? |
| 18:51.18 | Cotulla | I mean registers at 0x28800000 |
| 18:51.32 | jonpry | the code is for booting modem |
| 18:51.41 | Cotulla | but u need lpass I think |
| 18:51.44 | jonpry | it doesn't know how to boot apps |
| 18:51.51 | jonpry | lpass |
| 18:51.55 | Cotulla | why |
| 18:52.02 | jonpry | cause it's old code |
| 18:52.04 | Cotulla | it should be implemented there |
| 18:52.20 | jonpry | i don't know what to tell you |
| 18:52.25 | jonpry | it doesn't work |
| 18:52.41 | Cotulla | but it should load lpass to get audio |
| 18:52.48 | Cotulla | how it worked with original fw |
| 18:53.10 | jonpry | because SCM or whatever has it's own code |
| 18:53.18 | jonpry | it knows how to boot lpass |
| 18:53.25 | Cotulla | hm |
| 18:53.38 | Cotulla | so u think it loads elf from filesystem? |
| 18:53.51 | jonpry | it does |
| 18:54.02 | Cotulla | hm |
| 18:54.12 | Cotulla | but where lpsass mbn located? |
| 18:54.12 | jonpry | can we like disassemble SCM? |
| 18:54.22 | jonpry | firmware/image |
| 18:54.36 | Cotulla | inside system? |
| 18:54.40 | Cotulla | or another partition? |
| 18:54.42 | jonpry | root |
| 18:54.52 | Cotulla | hm |
| 18:55.03 | Cotulla | where is pil code? |
| 18:55.43 | detule | https://github.com/detule/linux-msm-d2/blob/msm-jb_2.5/arch/arm/mach-msm/pil-q6v4.c |
| 18:55.45 | detule | and |
| 18:55.52 | detule | https://github.com/detule/linux-msm-d2/blob/msm-jb_2.5/arch/arm/mach-msm/peripheral-loader.c |
| 18:56.37 | jonpry | i'm gonna go out for a bit |
| 18:56.58 | jonpry | personally i think codeaurora pil or trying to extract the SCM rom code is the way to go |
| 18:57.19 | Cotulla | pil-q6v4.c looks good |
| 18:58.19 | Cotulla | how is board file is called? |
| 18:59.07 | detule | <PROTECTED> |
| 18:59.24 | Cotulla | #define MSM_LPASS_QDSP6SS_PHYS0x28800000 |
| 18:59.32 | Cotulla | all right it should work . . . |
| 18:59.40 | Cotulla | if it has access to those registers |
| 18:59.58 | Cotulla | https://github.com/detule/linux-msm-d2/blob/msm-jb_2.5/arch/arm/mach-msm/devices-8960.c#L1321 |
| 19:00.38 | jonpry | i think that isn't the problem |
| 19:00.49 | jonpry | like register layout it wrong |
| 19:01.02 | Cotulla | so why it doesn't work? |
| 19:01.13 | Cotulla | did u try with pil-q6v4.c?> |
| 19:01.20 | jonpry | #define QDSP6SS_RST_EVB0x0 |
| 19:01.20 | jonpry | #define QDSP6SS_RESET0x04 |
| 19:01.26 | jonpry | i think that is wrong |
| 19:01.43 | jonpry | #define MSS_RESET(MSM_CLK_CTL_BASE + 0x2C64) |
| 19:01.47 | jonpry | doesn't even make sense |
| 19:01.55 | *** join/#htc-linux tehtrk (~The@rrcs-24-173-220-30.sw.biz.rr.com) |
| 19:01.55 | jonpry | #define MSS_S_HCLK_CTL(MSM_CLK_CTL_BASE + 0x2C70) |
| 19:01.56 | jonpry | #define MSS_SLP_CLK_CTL(MSM_CLK_CTL_BASE + 0x2C60) |
| 19:01.56 | Cotulla | no it's right |
| 19:01.59 | jonpry | moar wtf |
| 19:02.03 | jonpry | that MSS clocks |
| 19:02.06 | jonpry | no LPASS |
| 19:02.11 | Cotulla | the first two once |
| 19:02.12 | Cotulla | ones |
| 19:02.15 | jonpry | ah |
| 19:02.24 | Cotulla | but not sure about MSS_S_HCLK_CTL |
| 19:02.25 | jonpry | i though i saw 0x10 being used elseware |
| 19:02.56 | Cotulla | hm pil_q6v4_init_modem |
| 19:03.38 | Cotulla | but modem_base is 0 |
| 19:03.47 | Cotulla | so it's never called |
| 19:04.00 | Cotulla | hm or not |
| 19:05.13 | Cotulla | yers |
| 19:05.38 | Cotulla | device file has |
| 19:05.39 | Cotulla | 1)msm_8960_q6_lpass_data |
| 19:05.48 | Cotulla | 2)msm_8960_q6_mss_fw_data |
| 19:05.58 | Cotulla | 3)msm_8960_q6_mss_sw |
| 19:08.35 | langpoo | #ifdef CONFIG_HEXAGON_ARCH_V2 |
| 19:08.35 | langpoo | . = 0x10000000; |
| 19:08.35 | langpoo | #else |
| 19:08.35 | langpoo | . = 0x8da00000; |
| 19:08.35 | langpoo | #endif |
| 19:08.38 | langpoo | I will do for now |
| 19:09.06 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi) |
| 19:14.55 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-67-174.dynamic.mnet-online.de) |
| 19:16.28 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
| 19:20.15 | *** join/#htc-linux arrrghhhAWAY (~arrrghhh@unaffiliated/arrrghhh) |
| 19:23.11 | Cotulla | looking to interrupts |
| 19:26.09 | jonpry | does the image work on yours? |
| 19:28.59 | Cotulla | which image? |
| 19:35.27 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
| 19:37.28 | *** join/#htc-linux ntg (ntg@141.0.137.20) |
| 19:44.03 | *** join/#htc-linux ntg (~ntg@141.0.137.20) |
| 19:49.12 | *** join/#htc-linux jonpry (~jon@2602:306:c417:8aa0:b5c2:f85e:9b15:7ad6) |
| 20:00.46 | *** join/#htc-linux eR^Z` (~zzeratul@ip-84-42-202-42.net.upcbroadband.cz) |
| 20:03.49 | *** join/#htc-linux fakker (~fakker@cpc30-hitc6-2-0-cust607.9-2.cable.virginmedia.com) |
| 20:03.50 | *** join/#htc-linux fakker (~fakker@unaffiliated/fakker) |
| 20:19.42 | *** join/#htc-linux eR^zeRa` (~zzeratul@ip-84-42-202-42.net.upcbroadband.cz) |
| 20:31.23 | *** join/#htc-linux arif-ali (~arif-ali@94-192-24-56.zone6.bethere.co.uk) |
| 20:47.58 | *** part/#htc-linux langpoo (~zz@109.205.253.11) |
| 20:59.49 | *** join/#htc-linux AstainHellbring (AstainHell@unaffiliated/astainhellbring) |
| 21:30.03 | *** join/#htc-linux AstainHellbring (AstainHell@unaffiliated/astainhellbring) |
| 21:51.12 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
| 21:56.52 | *** join/#htc-linux helicopter88_2 (~helicopte@host11-8-dynamic.55-79-r.retail.telecomitalia.it) |
| 23:18.21 | *** part/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
| 23:29.36 | *** join/#htc-linux the-leviathan (~quassel@c-82-192-226-27.customer.ggaweb.ch) |
| 23:34.34 | *** join/#htc-linux the-leviathan (~quassel@c-82-192-226-27.customer.ggaweb.ch) |
| 23:59.15 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi) |