IRC log for #htc-linux on 20130218

00:31.13*** join/#htc-linux detule (~detule@unaffiliated/d3tul3)
01:04.42*** join/#htc-linux surge (~surge@pool-98-118-154-23.bflony.fios.verizon.net)
01:22.26*** join/#htc-linux RickyIV (RickyIV@151.77.62.70)
01:27.20*** part/#htc-linux RickyIV (RickyIV@151.77.62.70)
03:22.21*** part/#htc-linux Cotulla (~myfakemai@109.205.253.11)
03:31.57*** join/#htc-linux llstarks (~eric@pool-98-113-39-171.nycmny.fios.verizon.net)
04:21.00jonpryit is not happy when I turn the mmu on
04:25.56detuleyou got cotulla's secret code?
04:26.26detuleo i see he pushed some stuff out
04:26.45jonprynot so secret anymore
04:27.57detulegood man ..... damn he was keeping busy
04:29.07jonpryyeah lots of code
04:29.12jonpryapparently it works
04:30.18detuleneed to pass phys_offset via config
04:30.54jonpryyeah
04:31.13jonpryalso it needs to work on v3
04:32.00detulewhere does it die _SW_MMU_START ?
04:32.54jonpry<PROTECTED>
04:32.55jonpry<PROTECTED>
04:32.55jonpry<PROTECTED>
04:36.04detuleha /*  Insert trippy space effects.  */
04:36.29jonprymy guess is that SSR_BIT_TLB is different on v3
04:37.50detulewhere did he get those defines from
04:38.04jonpryreverse engineering i think
04:38.11jonprythat one should be easy to find
04:39.14jonprythe firmware should do the same thing where it hits the tlb register a few times and then does something to syscfg
04:40.03detuleunsets the tlb register?
04:40.07jonprysomehow i zonked by q6.b02 on the touchpad lol
04:40.25jonpryit programs them with something before turning mmu on
04:50.23jonpryhrm
04:50.44jonpry<PROTECTED>
04:50.44jonpry<PROTECTED>
04:50.44jonpry<PROTECTED>
05:47.00*** join/#htc-linux kiozen (~kiozen@ppp-93-104-66-136.dynamic.mnet-online.de)
06:48.23*** join/#htc-linux lamikr (lamikr@nat/nokia/x-fmmyflzttgxempte)
07:00.50*** join/#htc-linux ElFinLazz (~elfinlazz@182.215.84.22)
07:06.18*** join/#htc-linux ElFinLazz (~elfinlazz@182.215.84.22)
07:10.08*** join/#htc-linux kiozen (~kiozen@p578a42db.dip0.t-ipconnect.de)
08:03.44*** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info)
08:43.44*** join/#htc-linux eR^zeRa` (~zzeratul@ip-94-112-113-32.net.upcbroadband.cz)
08:56.46*** join/#htc-linux gauner1986 (~Adium@p578be66c.dip0.t-ipconnect.de)
08:56.54*** part/#htc-linux gauner1986 (~Adium@p578be66c.dip0.t-ipconnect.de)
09:04.11*** join/#htc-linux eR^Z` (~zzeratul@ip-94-112-113-32.net.upcbroadband.cz)
11:39.14*** join/#htc-linux Cotulla (~myfakemai@109.205.253.11)
11:51.46Cotullaheyeeaya
12:04.54*** join/#htc-linux Bry8Star{T (~Bry8Star-@gateway/tor-sasl/bry8star-2)
12:17.37*** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi)
12:37.06*** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi)
12:44.24*** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi)
12:52.42*** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi)
13:32.09*** join/#htc-linux helicopter88 (~helicopte@host118-94-dynamic.22-79-r.retail.telecomitalia.it)
13:35.33*** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi)
13:43.07*** join/#htc-linux ntg (ntg@141.0.137.20)
13:54.48d3tul3i guess it's zero after all
14:11.02Cotullaheya detule
14:11.16CotullaI reached start_kernel()
14:11.19Cotulla:P
14:17.14d3tul3nice man i saw your commit...over my head, but when i have some time i would like to sit down and figure it out
14:17.23d3tul3like the tlb miss logic
14:17.33Cotullaok
14:17.39Cotullabut it's not final variant
15:17.58jonpryhey Cotulla
15:24.08Cotullahey
15:26.28jonpryit freaks out when i turn mmu on
15:29.15Cotullahow
15:29.25jonprydunno
15:29.34jonprycan't trace after syscfg=r0
15:29.49Cotullahow u trace?
15:29.52Cotullahow do u trace?
15:30.32CotullaI am reached setup_arch()
15:30.49jonpryi load @ 0x4700 0000
15:30.55d3tul3is it possible your low level debugger is causing some exceptions and failing mmu with it's writes to whatever address
15:30.59jonprywrite to memory @ 0x4710 0000
15:31.17Cotullau sure
15:31.25CotullaPHYS_OFFSET is right?
15:31.30jonpryyes
15:31.49Cotullaso how u write to memory?
15:32.54jonpryhttp://pastebin.com/PZfEpPX5
15:32.59jonpryit works right before mmu on
15:33.42jonpryi hard coded the identity map too
15:33.59jonpry0x4700 0000 -> 0x4700 0000 16MB
15:34.16Cotulladckill can't work after MMU on
15:35.47jonpryit may have worked
15:35.49jonprywtf
15:35.56jonpryi added that after it wasn't working
15:36.29Cotullamaybe different TLB format
15:36.41Cotullabut disassemble and check addresses once more maybe
15:37.03Cotullahow do u map  0x4700 0000?
15:37.58jonpry<PROTECTED>
15:37.58jonpry<PROTECTED>
15:38.33Cotullatry TLB_UC
15:39.07jonpryhmm
15:40.53CotullaI made ram console
15:40.56Cotullafor debug purposes
15:42.05Cotulla"Hello from Hexagon Linux"
15:42.05Cotulla:P
15:42.17jonpry:p
15:42.30jonpryi have a program to ELF the qImage
15:42.55Cotulla:O
15:43.08Cotullaconvert it to PE before maybe also
15:43.12CotullaELF -> PE -> RAW -> ELF
15:43.13Cotulla:D
15:44.17Cotullabtw note 3.4.8 is very important inside NB419-1
15:44.34Cotullatake care about it when u will code
15:54.06jonprymade it to start_kernel
15:55.29Cotulla?
15:56.32jonpryshould i commit?
15:59.35Cotullawhat do u change?
15:59.49jonprynothing should affect you
15:59.58Cotullahm
16:00.01jonpryadded touchpad defines and wrote the ELF creator
16:00.07Cotullaok
16:00.15CotullaI will push ramconsole soon I think
16:00.25jonpryis it real ram console?
16:00.40jonpryhow did you do the tick timer?
16:00.44Cotullanot real android/linux one
16:00.45Cotullabut compatible
16:01.08CotullaI am in setup_arch()
16:02.00jonpryprintk works?
16:02.04Cotullano
16:02.08Cotullamy_out() works :)
16:04.31jonprysmp_start_cpus() lol
16:06.38Cotullahalts before for now
16:07.03Cotullamaybe remove
16:07.04CotullaCONFIG_SMP
16:07.32jonpryyeah
16:07.41jonprydebugging 6 threads at one time is going to suck
16:10.53Cotullaright
16:12.37Cotullasomehow it halts
16:14.50Cotullawhere was ur problem btrw?
16:15.39jonpryi'm not sure
16:16.00jonpryi think it was the linker script mostly
16:16.20jonprybut by the time i figured that out so many things had changed :p. like dckill
16:16.43Cotulla?
16:16.58Cotullait halts in
16:17.04Cotullapcycle_freq_mhz = 600;
16:17.06Cotulla<PROTECTED>
16:17.18Cotullaaddress is C0351294
16:20.13*** join/#htc-linux rajkosto (~rajkosto@cable-94-189-237-31.dynamic.sbb.rs)
16:20.18jonpryis unsigned long 64-bit on heagon?
16:20.41Cotullachecking that
16:20.49Cotullaaddress is not aligned
16:21.04*** join/#htc-linux Rajko (~rajkosto@wan.rajkonet.info)
16:21.50*** join/#htc-linux kiozen (~kiozen@ppp-93-104-66-136.dynamic.mnet-online.de)
16:24.07jonpryit appears to be  memd (r29 + #0) = r17:16 }
16:24.27Cotullano
16:24.37Cotullamemw (r24) = R1 for me
16:24.44Cotullabut there is a call before
16:25.00Cotullar29 is stack pointer
16:25.10Cotullait looks like presave registers
16:25.13Cotullaat the start up
16:25.16Cotullaof function
16:25.26jonpryic
16:25.29Cotullawell
16:31.07Cotullagotcha
16:31.19jonpryboom
16:31.43Cotullaneed addr2line
16:32.36Cotullahm
16:32.36Cotullamemcpy?
16:32.40Cotullaline 515
16:32.45Cotulla:O
16:37.08Cotullawtf
16:40.06Cotullalol it inside ram_console_write
16:40.08Cotullamemcpy
16:40.09Cotullawtf
16:42.34Cotullashit
16:42.43CotullaI need make own memcpy
16:43.49jonprywhy is that?
16:45.33Cotullanote 3.4.8 inside NB419-1
16:45.40Cotulladid u read it?
16:46.03Cotullaso it worked twice
16:46.13Cotullathen address started to be not 4bytes aligned
16:46.22Cotullaand it choosed another flow in memcpy
16:46.25Cotulla-> FAIL
16:47.08jonprylol
16:47.19Cotullatake care about it
16:47.25Cotullait's my second time
16:47.28Cotullaalready
16:47.29Cotulla:D
16:47.34jonprywhy not set everything cached and then dcflush
16:47.41Cotulladon't like it
16:47.51Cotullafuckl yea
16:47.52Cotullaworking
16:47.56Cotullait reaches
16:48.00Cotullait reaches
16:48.12jonpryhow are device drivers going to work when the compiler generates code breaking 3.4.8?
16:48.27Cotullaiowrite32() ioread32()
16:48.41Cotullanot sure about writel() and readl()
16:48.46Cotullacheck if they mapped too
16:49.08Cotullait reaches
16:49.11Cotullasetup_arch_memory()
16:49.14Cotulla^^
16:49.24jonpryvery cool
16:49.51Cotullaha
16:49.53Cotullait died
16:49.59Cotullacuz it wanted to access C100 0000
16:50.00Cotulla:D
16:50.14Cotullawell I should push I think
16:50.26jonpryyeah lemme see if i can get that to work
16:56.46Cotullahm should I do git pull
16:57.55jonpryyou can try to pull before commit
16:57.58CotullaI am not sure about start up seq
16:58.00Cotullabut ok
16:58.06jonpryand it will either work or not
16:58.14Cotullaone conflict
16:58.16Cotulla^^
16:58.45Cotullawell u added code
16:59.09jonpryto where?
16:59.14Cotullahead.S
16:59.16Cotullalet me port
17:00.34Cotullaur test code is "active"
17:00.44Cotullawill do #if 0 too
17:01.08jonprynothing
17:02.01jonpryoh shit
17:02.09jonpryi commented that out but didn't save the file
17:02.15Cotulla:P
17:02.36Cotullapushed
17:03.35Cotullathere are few #errors
17:04.42Cotullacompiling
17:05.47jonpry0xF800 0000 is virtual?
17:05.51Cotullayes
17:07.24Cotullano idewa if it is good
17:07.58*** join/#htc-linux lamikr (lamikr@nat/nokia/x-dgcgbizlyvqhirxr)
17:10.57Cotullasetup_arch_memory
17:11.24jonpryi think it's running
17:11.36Cotulladump ramconsole
17:11.39jonprybut i have no way to dump ramconsole on webos yet
17:11.41Cotullau can copy from android driver
17:11.48Cotulladump function I think
17:17.22jonpryHello from Hexagon Linux 1
17:17.22jonpryHello from Hexagon Linux 2
17:17.22jonprysetup_arch() 2
17:17.22jonpryERROR: ELR=C00171D0 ADDR=C1000000 LR=C000B838
17:17.30Cotullaright
17:17.32Cotullasame
17:17.33Cotulla:)
17:17.39Cotullanice
17:17.45CotullaI am looking to arch/cris now
17:18.08jonprywhy is it trying to touch memory there?
17:18.24Cotullanot sure why
17:18.26Cotullabut we need rewrite it
17:18.36Cotullasetup_arch_memory()
17:18.43Cotullacurrent implementation totally bad
17:19.00CotullaI am looking how it's done for another arch
17:20.23*** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net)
17:20.43Cotullafor example
17:20.44Cotulla<PROTECTED>
17:20.44Cotulla<PROTECTED>
17:22.53Cotullafuck who designed that shit
17:25.13Cotullamaybe make it like ARM
17:25.30jonpryi think thats a little complicated
17:25.36Cotullayea bad idea
17:25.39jonprywe don't need highmem support or anything
17:25.40Cotullajust hardcode it there
17:25.44Cotullaand done
17:25.48Cotulla:D
17:25.57Cotullait's already complicated with such variables names
17:26.17jonpryand PTE support
17:26.45CotullaPTE is not connected here
17:27.47jonpryit thinks it can use printk
17:28.10Cotullalook inside
17:28.24Cotullabut I prefer for now use my my_out()
17:28.33Cotullacuz it's OS independent
17:28.37Cotullaas much as possible
17:31.30*** join/#htc-linux rajkosto (~rajkosto@cable-94-189-237-31.dynamic.sbb.rs)
17:32.17*** join/#htc-linux Rajko (~rajkosto@wan.rajkonet.info)
17:34.42*** join/#htc-linux bitrot (~rajkosto@wan.rajkonet.info)
17:35.55jonprylooks like it needs a mem= command line parameter
17:36.08*** join/#htc-linux ntg (ntg@141.0.137.20)
17:39.29Cotullayea
17:39.34Cotullabut we are going to remove it
17:39.44jonpryremove mem= ?
17:40.12jonpryi added it to config_cmdline
17:40.20Cotullayes
17:40.25jonprynew problem
17:40.26jonpryERROR: ELR=C002A0C4 ADDR=00000000 LR=C0024710
17:40.37Cotullahow
17:40.46jonpryedit .config
17:40.50Cotullalol
17:40.57jonpryCONFIG_CMDLINE="mem=16M"
17:40.57CotullaI am wokring under setup_arch_memor() now
17:41.02CotullaI think it's useless to try with old one
17:41.10Cotullait assumes no PHYS_OFFSET for example
17:41.19jonpryit needs lots of work
17:41.24Cotullakinda
17:41.29jonprybut all platforms support mem=
17:41.41Cotullaour platform doesn't support true command line
17:41.46Cotulla:P
17:41.51jonprynot yet
17:41.54Cotullaanyway it's not a problem]
17:42.00CotullaI assume
17:42.03Cotullaone memory space
17:42.30jonprythe null pointer is in free_bootmem
17:42.37jonpryprobably a bug_on
17:42.53Cotullaok
17:43.23Cotullau can tell me ram start and size
17:43.30CotullaI can include it
17:43.42jonprystart is 0x4700 0000
17:43.45jonpry16MB
17:44.08jonprystart is just PHYS_OFFSET
17:44.25Cotullayes I hope
17:44.56jonpryi'll have to hack the WebOS kernel to get more ram. hopefully that is enough for now
17:45.19Cotullaok
17:45.22Cotullawait
17:45.28Cotullaoriginal design wants
17:45.32Cotulla+ 4 for DMA
17:45.38Cotullabut it sucks like I said before
17:45.50jonpryneed that for networking
17:46.01Cotullaso can u increase it up to 20?
17:46.24jonpryanything over 16 might as well be 500
17:46.32Cotullaheh
17:53.15jonprylolz
17:53.30jonprythe arguments at free bootmem
17:53.37jonpry0x375000 0xFFC8B000
17:53.40jonpryphys and size
17:54.10jonpryi thinks PFN_PHYS is broke
17:58.01CotullaI suggest u to do something else cuz I am rewritting it
17:58.59jonpryyeah i'm just messing around
17:59.24jonprythe whole thing is fubar
17:59.39jonprythinks all memory is at zero or some crap
18:00.23Cotullarigtht
18:11.43*** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info)
18:12.21d3tul3whoa
18:12.53d3tul3just saying something meaningless so that the log records i was present at these momentous times
18:13.19Cotullawhat/
18:13.22CotullaI don't get u
18:13.33Cotulla:(
18:15.37jonprythis port is a little different from our usual endeavors because all the really nasty parts of the kernel need to be rewritten.
18:15.48Cotullanot really little
18:15.57Cotullait just another thing
18:16.04Cotullakinda innovative development
18:16.05Cotullalol
18:16.10jonprywhole different animal?
18:16.14Cotullayea
18:16.31CotullaI can't count porting driver from one minor version to another one minor version  to be cool
18:16.46Cotullacuz no backward compatible
18:16.50Cotullanothing
18:17.24jonprythey did make a compiler that sort of works
18:17.32d3tul3jonpry, you're membooting these guys?
18:17.38d3tul3oh wait no
18:17.41Cotulla1<<(30 - 22)
18:17.47Cotullawtf that guy means
18:17.49d3tul3you've membooted a kernel with a hacked loader
18:18.03Cotullait's actually 0x100
18:18.05jonpryit's 1gb / 4MB
18:18.17CotullaC00 + 100 = D00
18:18.21jonpryduh, you can't read :p
18:18.36Cotullaor no
18:18.40Cotulla22 is 4M
18:18.42Cotulla20 is 1M
18:18.59jonpryyeah it's the number of l2 tables required for kernel segment
18:19.00Cotullabut is it ever worked
18:19.26Cotullahis comment wrong
18:19.27jonpryd3tul3, yeah i membooted a kernel with new loader
18:19.30Cotullaabout PAGE_OFFSET >> 22
18:19.38Cotullaor not
18:19.41Cotullanot really
18:19.44Cotulla87%
18:19.52Cotullashift to C000
18:19.52Cotullathen
18:19.55Cotullahe adds 100
18:20.01Cotullaactually 100 * 4 = 400
18:20.02Cotullaso
18:20.05Cotullano
18:20.08d3tul3lol
18:20.40Cotullaso
18:20.55CotullaC000 0000 is 3G right?
18:21.01jonpryyes
18:21.17Cotullafull table is 0x1000
18:21.26Cotulla0x400 entries
18:21.32Cotullahe shifted to + 0x300
18:21.36Cotullathen
18:21.42Cotulla1 << (30 - 22)
18:21.45Cotullait's 0x100
18:21.46Cotulla1 << 8
18:21.52Cotullaso he shifted 0x400
18:21.54Cotullaup to the end
18:21.57Cotullaof address space
18:22.06Cotullaand he started to map it down
18:22.21Cotullabut other part says that HVM is there ? :O
18:22.33Cotullaoh
18:22.37Cotullait's different variables
18:22.39Cotullaf them
18:22.47Cotullasegtable_end is really end
18:22.58Cotullahe could just add 0x400 to it
18:22.59Cotullapff
18:23.34Cotullathen he adjust setable
18:23.35Cotullato the end
18:24.23Cotullaclear now
18:24.25Cotullait's between
18:24.30Cotullakernel end and VMALLOC
18:24.31Cotulla:D
18:26.09jonpry16MB entries?
18:29.01Cotulla4J
18:29.03Cotulla4M
18:29.25Cotullain my case it should be
18:29.42ntgCotulla: write me a uefi bootloader for my tf101 so i can use winrt on it
18:29.43Cotulla16M C+ 4M C + 4M UC
18:29.48ntgi know you're that bored
18:29.55Cotullano
18:30.16Cotulla:P
18:30.18ntgthere's cake
18:30.26Cotullait wont' work good anyway
18:30.56CotullaI will finish my LEORT port
18:31.23Cotullagotta away for some time
18:31.38Cotullawill come back in 1 hour
18:31.42CotullaI think
19:01.58*** join/#htc-linux eR^zeRa` (~zzeratul@ip-94-112-113-32.net.upcbroadband.cz)
19:04.28Cotullaback
19:06.34*** join/#htc-linux arif-ali (~arif-ali@31.106.3.248)
19:10.41*** join/#htc-linux ntg (ntg@141.0.137.20)
19:36.47*** join/#htc-linux langpoo (~zz@109.205.253.11)
20:07.35jonpryi'm surprised this touchpad hasn't caught on fire yet
21:13.42*** join/#htc-linux BabelO (~fcr@unaffiliated/babelo)
21:18.59*** join/#htc-linux lamikr (lamikr@nat/nokia/x-ekluvbuimsoulons)
21:34.48*** join/#htc-linux ALoGeNo (~alogeno@243.Red-217-125-20.staticIP.rima-tde.net)
21:34.48*** join/#htc-linux ALoGeNo (~alogeno@unaffiliated/alogeno)
22:41.15*** join/#htc-linux arif-ali (~arif-ali@81.5.168.97)
23:35.09*** join/#htc-linux detule (~detule@unaffiliated/d3tul3)
23:46.10*** join/#htc-linux eR^zeRa` (~zzeratul@ip-94-112-113-32.net.upcbroadband.cz)
23:51.35*** join/#htc-linux lamikr (lamikr@nat/nokia/x-efnyusqgwejeivwd)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.