| 00:35.30 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
| 00:59.57 | detule | jonpry, you got a second |
| 01:00.18 | jonpry | sure |
| 01:00.44 | detule | so my phys_offset is 80200000 |
| 01:00.53 | jonpry | yes |
| 01:00.56 | detule | this should correspond to start of ram right, (or thereabouts) |
| 01:01.10 | jonpry | 0x8000 0000 is the start of ram |
| 01:01.22 | detule | ok my page_offset is C0000000 |
| 01:01.30 | detule | this should be what, first virtual address? |
| 01:01.41 | detule | corresponding to phys_offset |
| 01:01.53 | jonpry | thats pretty much the same on all 3g split kernels |
| 01:02.01 | jonpry | its just the beginning of kernel only address space |
| 01:02.55 | jonpry | did you ever disassemble machine_kexec? |
| 01:03.02 | detule | ok so the 1G in between is user acc mem |
| 01:03.13 | jonpry | well no |
| 01:03.18 | detule | yes, and can't make sense of machine_kexec + 0x54 |
| 01:03.43 | jonpry | do arm-*-objdump -S machine_kexec.o |
| 01:04.03 | jonpry | nobody can access physical addresses |
| 01:04.10 | jonpry | especially not user programs |
| 01:04.47 | jonpry | and what usually happens is that 0xc000 0000 is virtual for 0x8200 0000 |
| 01:06.08 | detule | oh -S is nice |
| 01:06.52 | jonpry | especially with all that automatic inlining the compiler does |
| 01:09.36 | detule | ok so there's 1GB between 820* and c00 and that's kernel only....where's the rest |
| 01:10.34 | jonpry | no |
| 01:10.42 | jonpry | there is zero between those addresses |
| 01:10.48 | jonpry | they are identical |
| 01:11.06 | jonpry | just one is phys and the other virt |
| 01:11.13 | detule | i see |
| 01:12.08 | detule | http://pastebin.com/PPwEeMbE that has the traceback and the disassem |
| 01:12.15 | detule | this machine_kexec has a few printks |
| 01:12.21 | detule | none of which i see |
| 01:13.56 | jonpry | where is 0x54? |
| 01:15.22 | detule | i mean if 00000148 is machine_kexec then machine_kexec + 0x54 is what 1a2? |
| 01:15.37 | detule | er |
| 01:17.54 | jonpry | i got 19c |
| 01:18.15 | jonpry | but that is obviously wrong |
| 01:19.00 | jonpry | like r2 has the address of the fault so its likely involved |
| 01:19.09 | jonpry | and r7 could be too |
| 01:19.55 | detule | 19c maybe |
| 01:20.01 | jonpry | is this dissembly complete. i expect more code for some reason |
| 01:21.10 | jonpry | somehow they must have defined pr_crit to be a noop |
| 01:21.19 | jonpry | probably a flag you can set or something |
| 01:22.53 | detule | i just pasted the code around machine_kexec....i used pr_crit for the printk's ebcause i see an earlier printk in ram_console that's printed with pr_crit |
| 01:24.41 | jonpry | so your not getting any output from these new printks? |
| 01:24.50 | detule | none |
| 01:25.10 | jonpry | same crash? |
| 01:26.20 | detule | ere's my machine_kexec diff http://pastebin.com/VmzBpcQj |
| 01:27.06 | detule | nd the crash i just pulled http://pastebin.com/mXYCr3ri |
| 01:27.17 | jonpry | well first i would check that pr_crit actually compiles to something |
| 01:27.30 | jonpry | otherwise i think that image pointer is bad |
| 01:27.39 | jonpry | so image->head could blow is sky high |
| 01:27.40 | detule | that KEXEC: MSM Watchdog Exit - Deactivated is printed with pr_crit |
| 01:27.59 | jonpry | from the same file? |
| 01:28.17 | detule | no not the same |
| 01:28.29 | jonpry | they might be defining something before including the definition of pr_crit |
| 01:29.53 | detule | i only have only one define pr_crit and it's just printk(KERN_CRIT |
| 01:30.03 | detule | so i guess the image pointer is bad |
| 01:30.49 | jonpry | there are some other issues with your technique though |
| 01:31.29 | jonpry | like the compiler doesn't make any guarantees about the order that local assignments take place |
| 01:31.47 | detule | yeah but i was hoping to at least see that first pr_crit |
| 01:31.55 | jonpry | so there is no reason to think that page_list = image->head & PAGE_MASK; is actually going to take place before your pr_crit |
| 01:32.19 | jonpry | or that your first pr_crit is happening before all those other assignments |
| 01:33.22 | jonpry | except that disassembly can reveal the actual order |
| 01:35.41 | detule | so the only thing is ee before that pr_Crit is page_list = image->head & PAGE_MASK; |
| 01:35.58 | detule | in the objdump |
| 01:36.11 | jonpry | that is good |
| 02:18.02 | jonpry | my kexec is so unreliable that its useless |
| 02:23.48 | detule | at least now the address of the failue involves an r2 instruction |
| 02:41.13 | jonpry | stupid delay loop won't calibrate |
| 02:42.08 | jonpry | i think timers are dead |
| 02:45.08 | *** join/#htc-linux MacDrunk (~marper@201.165.148.78) |
| 02:46.36 | *** part/#htc-linux MacDrunk (~marper@201.165.148.78) |
| 03:36.52 | *** join/#htc-linux raymonddull (~raymonddu@75-134-191-7.dhcp.bycy.mi.charter.com) |
| 03:53.49 | *** join/#htc-linux jianC (4c141154@gateway/web/freenode/ip.76.20.17.84) |
| 04:00.43 | *** join/#htc-linux NeoMatrixJR (~NeoMatrix@173-18-84-218.client.mchsi.com) |
| 04:14.24 | *** join/#htc-linux raymonddull (~raymonddu@75-134-191-7.dhcp.bycy.mi.charter.com) |
| 05:28.43 | *** join/#htc-linux mes (~mes@sentry.lazo.ca) |
| 05:31.46 | *** join/#htc-linux skodde (~skodde@unaffiliated/skodde) |
| 05:38.44 | *** join/#htc-linux raymonddull (~raymonddu@75-134-191-7.dhcp.bycy.mi.charter.com) |
| 06:28.08 | *** join/#htc-linux raymonddull (~raymonddu@75-134-191-7.dhcp.bycy.mi.charter.com) |
| 06:55.47 | *** join/#htc-linux balans2 (~user@82-170-217-205.ip.telfort.nl) |
| 07:34.32 | *** join/#htc-linux GNUtoo (~gnutoo@host87-128-dynamic.7-79-r.retail.telecomitalia.it) |
| 07:41.15 | *** join/#htc-linux GNUtoo|laptop (~gnutoo@host87-128-dynamic.7-79-r.retail.telecomitalia.it) |
| 07:43.43 | *** join/#htc-linux eR^zeRa` (~zzeratul@88.103.98.168) |
| 07:59.30 | *** join/#htc-linux conantroutman (~chatzilla@cpc5-pert4-2-0-cust205.sgyl.cable.virginmedia.com) |
| 08:01.19 | *** join/#htc-linux ALoGeNo (~alogeno@243.Red-217-125-20.staticIP.rima-tde.net) |
| 08:01.20 | *** join/#htc-linux ALoGeNo (~alogeno@unaffiliated/alogeno) |
| 08:10.01 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-fff8dd00-39.dhcp.inet.fi) |
| 08:24.12 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-75-159.dynamic.mnet-online.de) |
| 08:28.04 | *** join/#htc-linux Cotulla (~myfakemai@nat100-255-205-109.tvoe.tv) |
| 08:49.58 | *** join/#htc-linux marc1706 (~Marc@phpbb/modifications/marc1706) |
| 09:00.35 | *** join/#htc-linux skodde (~skodde@unaffiliated/skodde) |
| 09:10.20 | *** join/#htc-linux rob_w (~bob@ppp-93-104-17-125.dynamic.mnet-online.de) |
| 09:10.22 | *** join/#htc-linux rob_w (~bob@unaffiliated/rob-w/x-1112029) |
| 09:22.24 | *** join/#htc-linux Kookstar (~Kookstar@s3109.dyn.hrz.tu-darmstadt.de) |
| 09:42.45 | *** join/#htc-linux leviathan_ (~quassel@2001:470:26:484:6ef0:49ff:fee6:8dca) |
| 09:55.39 | *** join/#htc-linux BabelO (~wdlxtv@AMontpellier-553-1-176-108.w92-133.abo.wanadoo.fr) |
| 10:02.54 | *** join/#htc-linux Kookstar (~Kookstar@s3109.dyn.hrz.tu-darmstadt.de) |
| 10:43.36 | *** join/#htc-linux _kiozen (~kiozen@ppp-93-104-75-159.dynamic.mnet-online.de) |
| 10:52.20 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-75-159.dynamic.mnet-online.de) |
| 10:58.20 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-75-159.dynamic.mnet-online.de) |
| 11:00.29 | *** join/#htc-linux ccube (ccube@nx.ccube.de) |
| 11:07.57 | *** join/#htc-linux Kookstar (~Kookstar@s3109.dyn.hrz.tu-darmstadt.de) |
| 11:08.39 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
| 11:21.21 | *** join/#htc-linux walter79_ (~walter79@31-17-62-189-dynip.superkabel.de) |
| 11:29.28 | *** join/#htc-linux MethoS- (~clemens@134.102.106.250) |
| 11:36.53 | *** join/#htc-linux ccube (ccube@nx.ccube.de) |
| 11:41.20 | *** join/#htc-linux Kookstar (~Kookstar@s3109.dyn.hrz.tu-darmstadt.de) |
| 12:03.46 | *** join/#htc-linux skodde (~skodde@unaffiliated/skodde) |
| 12:24.17 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
| 12:36.22 | *** join/#htc-linux ccube (ccube@nx.ccube.de) |
| 12:38.45 | *** join/#htc-linux Kookstar (~Kookstar@s3109.dyn.hrz.tu-darmstadt.de) |
| 13:30.08 | *** join/#htc-linux Rajko (~rajkosto@cable-94-189-239-212.dynamic.sbb.rs) |
| 13:30.56 | *** join/#htc-linux bitrot (~rajkosto@wan.rajkonet.info) |
| 14:01.16 | *** join/#htc-linux Kookstar (~Kookstar@s3109.dyn.hrz.tu-darmstadt.de) |
| 14:15.58 | *** join/#htc-linux AstainHellbring (~AstainHel@unaffiliated/astainhellbring) |
| 14:22.12 | *** join/#htc-linux conantroutman (~chatzilla@cpc5-pert4-2-0-cust205.sgyl.cable.virginmedia.com) |
| 14:26.01 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
| 14:31.41 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
| 15:24.53 | *** join/#htc-linux pbaxter (~pbaxter@host164-113-dynamic.6-79-r.retail.telecomitalia.it) |
| 15:25.02 | *** join/#htc-linux Alex[sp3dev]_ (d5551202@gateway/web/freenode/ip.213.85.18.2) |
| 15:49.06 | *** join/#htc-linux Kookstar (~Kookstar@p4FFF1916.dip.t-dialin.net) |
| 16:13.23 | *** join/#htc-linux jonpry (~jon@c-24-17-200-206.hsd1.wa.comcast.net) |
| 16:28.43 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
| 16:48.05 | *** join/#htc-linux ALoGeNo (~alogeno@243.Red-217-125-20.staticIP.rima-tde.net) |
| 16:48.06 | *** join/#htc-linux ALoGeNo (~alogeno@unaffiliated/alogeno) |
| 17:06.24 | *** join/#htc-linux l0wbra1n (~l0wbra1n@unaffiliated/l0wbra1n) |
| 17:12.31 | *** join/#htc-linux Kookstar (~Kookstar@p4FFF1916.dip.t-dialin.net) |
| 17:13.24 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
| 17:28.24 | *** join/#htc-linux balans2 (~user@82-170-217-205.ip.telfort.nl) |
| 17:31.56 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-75-159.dynamic.mnet-online.de) |
| 17:40.17 | jonpry | only one more post until xda lets me in |
| 17:40.31 | jonpry | it's been years in the making |
| 17:45.48 | jonpry | ^&*( me |
| 17:46.33 | jonpry | it's a ruse! |
| 17:49.05 | detule | post in one of our tp2 threads |
| 17:49.20 | jonpry | i'm up to 10 |
| 17:49.21 | detule | everyone mentions jonpry but nobody has ever seen any evidence of him |
| 17:49.26 | jonpry | but still banned |
| 17:50.08 | detule | how did you achieve that again |
| 17:50.59 | jonpry | i posted in general forums |
| 17:51.19 | jonpry | but i think they don't count towards the real 10 until they are individually approved by moderators |
| 17:51.39 | detule | where's arrrghhh when you need him |
| 17:52.48 | jonpry | or thedeadcpu |
| 17:54.01 | jonpry | how is the kexec? |
| 17:54.45 | jonpry | is it the image pointer? |
| 17:56.19 | detule | i have no freaking clue....that virtual adress that's referenced in the failure |
| 17:56.26 | detule | is in the kernel .text section |
| 17:56.44 | detule | i am not sure it's supposed to be trying to place something there until it hits relocate_kernel or whatever |
| 17:58.06 | jonpry | your sure about .text? |
| 17:58.35 | jonpry | i'm still confused about where in machine_kexec() it's crashing |
| 17:59.13 | detule | <5>[ 0.000000] .init : 0xc0008000 - 0xc004d000 ( 276 kB) |
| 17:59.14 | detule | <5>[ 0.000000] .text : 0xc0100000 - 0xc0a7583c (9687 kB) |
| 17:59.14 | detule | <5>[ 0.000000] .data : 0xc0b00000 - 0xc0bd3c20 ( 848 kB) |
| 17:59.14 | detule | <5>[ 0.000000] .bss : 0xc0bd3c44 - 0xc10013fc (4278 kB) |
| 17:59.32 | jonpry | that image pointer is crates in kernel/kexec.c fwiw |
| 17:59.39 | detule | [ 86.899832] [c010b1a4] *pgd=8021140e(bad) |
| 18:00.01 | jonpry | well that could explain it |
| 18:00.21 | jonpry | kernel .text is probably read only |
| 18:00.32 | detule | what about when it tries to relocate the kernel |
| 18:00.41 | jonpry | mmu is off |
| 18:00.49 | detule | so what the hell is going on |
| 18:01.11 | jonpry | find out how that pointer is generated |
| 18:01.17 | jonpry | maybe its unitialized |
| 18:01.26 | jonpry | er |
| 18:01.33 | jonpry | bad math, who knows |
| 18:01.56 | detule | and i still don't understand this page_size in the kernel versus page_size in mkbootimg difference |
| 18:03.02 | detule | how can one be 2^11 and the other 2^12 |
| 18:03.17 | jonpry | i think its not terribly relevant if your specifying the load address |
| 18:03.35 | jonpry | and i think the problem they had is that the kernel wants to be loaded at like 0x8200 0800 |
| 18:05.36 | jonpry | image->control_code_page |
| 18:05.43 | detule | yeah mkbootimg always loads the kernel at base + 800 regardless of what page size is |
| 18:05.48 | jonpry | i don't yet understand how it find that page |
| 18:06.47 | jonpry | maybe fastboot itself won't load it to the specified address if page size is set too coarse |
| 18:07.53 | detule | so i don't know if it's control_page because i have a physical adress for it |
| 18:08.05 | Cotulla | +8000 u mean? |
| 18:08.19 | Cotulla | u need at least 16 KB for FirstPTE . . . |
| 18:08.25 | detule | [ 86.899008] About to calculate reboot_code_buffer_phys |
| 18:08.25 | detule | [ 86.899130] reboot_code_buffer_phys: a1b1d000 |
| 18:09.55 | jonpry | need to print out the virt too |
| 18:11.01 | jonpry | you can always use my machine_kexec |
| 18:11.06 | detule | is it magical |
| 18:11.11 | jonpry | seems to be |
| 18:11.40 | jonpry | http://pastebin.com/7XK3QzAz |
| 18:11.59 | jonpry | will take a little bit to make it work with your relocate_kernel.S |
| 18:12.27 | jonpry | also you probably need a different outer_disable() |
| 18:13.31 | detule | that's ok i can kang the soft_restart from 3.4 |
| 18:13.41 | detule | but wat's the key here disable_preempt? |
| 18:14.10 | jonpry | identify_map(reboot_code_buffer_phys) |
| 18:14.23 | jonpry | and flush_icache_all |
| 18:15.03 | jonpry | i like to disable all that irq crap early |
| 18:15.16 | jonpry | but it didn't actually help tbh |
| 18:15.57 | jonpry | oh and someway to hotplug your second core is probably a good idea |
| 18:22.10 | *** join/#htc-linux ccube (ccube@nx.ccube.de) |
| 18:23.58 | jonpry | all you need is to leave that pos on :p |
| 18:27.10 | detule | the virtual for the control page looks ok too |
| 18:27.10 | detule | [ 84.963711] reboot_code_buffer_phys: a11dc000 |
| 18:27.11 | detule | [ 84.963894] reboot_code_buffer: daadc000 |
| 18:27.20 | detule | i think it's image->head |
| 18:28.07 | jonpry | so page_list? |
| 18:28.14 | detule | yes |
| 18:28.47 | jonpry | i don't see where page_list is dereferenced though |
| 18:29.14 | detule | indirection_page gets passed to relocate kernel but yeah at that point the cpu is off |
| 18:29.19 | detule | mmu is off |
| 18:31.56 | jonpry | actually i know what it is |
| 18:32.02 | detule | damnit [ 84.963345] page_list: a1205000 |
| 18:32.18 | jonpry | &kexec_start_address |
| 18:32.25 | jonpry | all those things are in .text |
| 18:32.32 | jonpry | because of the crazy ass way it works |
| 18:33.55 | jonpry | they are pc relative data stored at the end of relocate_new_kernel |
| 18:34.26 | jonpry | i think there is some linker patch that causes that function to not go into .text |
| 18:34.49 | jonpry | but you could do something else anyways |
| 18:35.36 | detule | but aren't they supposed to point to addresses in .text since they get used in relocate_new_kernel ...as in |
| 18:36.02 | jonpry | *(reboot_code_buffer + (kexec_start_address-relocate_new_kernel)) = image->start |
| 18:36.33 | jonpry | + copius casting |
| 18:36.40 | jonpry | do that after the memcpy |
| 18:38.24 | detule | cool i need something similar for boot_atags? |
| 18:38.42 | jonpry | for all 4 |
| 18:38.50 | jonpry | *((u32*)reboot_code_buffer + ((u32*)&kexec_start_address-(u32*)relocate_new_kernel)) = image->start |
| 18:39.36 | detule | thanks i'll try that after the game....ttyl |
| 18:40.41 | *** join/#htc-linux FlawlesStyle (~LOL@173.255.136.122) |
| 18:40.42 | *** join/#htc-linux FlawlesStyle (~LOL@unaffiliated/flawlesstyle) |
| 18:54.35 | *** join/#htc-linux FlawlesStyle (~LOL@unaffiliated/flawlesstyle) |
| 19:00.07 | *** join/#htc-linux wk_ (wk@141.0.136.22) |
| 19:01.16 | *** join/#htc-linux zeusk_ (77eb3613@gateway/web/freenode/ip.119.235.54.19) |
| 19:01.25 | zeusk_ | hi |
| 19:02.27 | *** join/#htc-linux sdub (~sdub@Aircrack-NG/Friend) |
| 19:23.40 | *** join/#htc-linux skodde (~skodde@unaffiliated/skodde) |
| 19:51.14 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-fff8dd00-39.dhcp.inet.fi) |
| 19:57.20 | *** join/#htc-linux LargePrime (~LargePrim@108.101.162.67) |
| 20:30.08 | *** join/#htc-linux mes (~mes@S0106000ce55186df.cc.shawcable.net) |
| 20:50.45 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
| 21:09.32 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
| 21:19.48 | detule | jonpry, what the hell trying to make those assignments is a serious bug |
| 21:20.00 | detule | (it took me a while to parse what you were sayign) |
| 21:20.32 | jonpry | i'm not sure if read only text is normal |
| 21:44.03 | *** join/#htc-linux jonpry (~jon@unaffiliated/jonpry) |
| 21:45.37 | jonpry | detule got it fixed? |
| 21:47.09 | *** join/#htc-linux BabelO (~wdlxtv@AMontpellier-553-1-176-108.w92-133.abo.wanadoo.fr) |
| 21:47.11 | detule | wife's working on the windows machine...can't flash just yet |
| 21:49.25 | jonpry | ic |
| 21:56.56 | *** join/#htc-linux ALoGeNo (~alogeno@243.Red-217-125-20.staticIP.rima-tde.net) |
| 21:56.57 | *** join/#htc-linux ALoGeNo (~alogeno@unaffiliated/alogeno) |
| 21:57.31 | detule | odin comes conveniently equipped with a flash counter |
| 21:57.52 | detule | i am at 37...i think i've probably voided my warranty |
| 21:58.18 | detule | need kexec, at this rate i'll burn through /boot pretty fast |
| 21:59.52 | jonpry | how many do you get? |
| 22:09.39 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-fff8dd00-39.dhcp.inet.fi) |
| 22:17.19 | *** join/#htc-linux ALoGeNo (~alogeno@unaffiliated/alogeno) |
| 22:39.05 | detule | apparently more than 38 |
| 22:39.09 | detule | http://pastebin.com/2QBSCr72 |
| 23:11.39 | *** part/#htc-linux Cotulla (~myfakemai@nat100-255-205-109.tvoe.tv) |
| 23:26.00 | detule | my math must be wrong since it shouldn't be erroring out at d9c94000 if reboot_code_buffer: dafe2000 (everything comes after it right)? |
| 23:31.54 | *** join/#htc-linux ALoGeNo (~alogeno@unaffiliated/alogeno) |
| 23:48.21 | jonpry | detule, where is the code? |
| 23:54.50 | detule | temporarily undid it while i am trying something else this is where am at currently http://pastebin.com/dZuKhYps |
| 23:59.23 | jonpry | there is a slight problem there |
| 23:59.35 | jonpry | like your offsets will be computed in bytes |
| 23:59.42 | jonpry | but then applied in dwords |