00:37.38 | *** join/#htc-linux jonpry (~jon@c-67-170-101-76.hsd1.wa.comcast.net) |
01:01.20 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
01:05.44 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
01:09.34 | detule | hey jonpry |
02:07.51 | *** join/#htc-linux leviathan (~quassel@2001:470:26:484:6ef0:49ff:fee6:8dca) |
02:09.43 | jonpry | hi detule |
02:10.16 | detule | jonpry, this is what you guys are using https://gitorious.org/droid3-kexec/droid3-kexec-kernel/blobs/3.0.8-dev/arch/arm/mm/proc-v7.S#line63 |
02:10.51 | jonpry | mmm |
02:11.07 | jonpry | well depends |
02:11.41 | jonpry | that is not used in the kexec module loaded into the .35 host kernel |
02:12.08 | jonpry | in fact we don't use any cpu_reset in kexec. it kept busting for some reason |
02:12.30 | detule | not turning the mmu off at all |
02:12.44 | jonpry | we turn mmu off |
02:12.47 | jonpry | for sure |
02:12.51 | detule | where |
02:13.00 | jonpry | its built into relocate-kernel.S |
02:13.13 | jonpry | in the actual relocate-kernel entry point |
02:13.25 | detule | i thought to turn the mmu off one needs to flib a bit on the control register |
02:13.31 | jonpry | yeah it does |
02:13.32 | detule | don't see that in relocate kernel |
02:14.29 | jonpry | https://gitorious.org/droid3-kexec/droid3-kexec-modules/blobs/master/arch/arm/kernel/relocate_kernel.S#line97 |
02:14.38 | jonpry | 0x7 is all bs off |
02:14.49 | detule | oh yeah |
02:14.50 | detule | that's what i need |
02:15.01 | detule | i think this turning the mmu off in setup_mm_for_reboot is bad mojor |
02:15.10 | detule | s/mojor/mojo/ |
02:15.15 | jonpry | it doesn't does it? |
02:15.16 | detule | way too earl |
02:15.18 | detule | y |
02:15.28 | jonpry | it just prepares vm for mmu off |
02:15.44 | detule | i keep thinking it wipes out the instruction stream coz this pos keeps stalling |
02:16.11 | jonpry | ouch |
02:16.31 | jonpry | i thought it was working except cpu1? |
02:16.41 | detule | well god damnit i can't replicate my earlier success |
02:17.12 | detule | either my tree is polluted somehow or it was just dumb luck that it worked when i pulled that last_kmsg |
02:17.16 | jonpry | adb only works for me in one branch |
02:17.20 | jonpry | and i can't find the diff |
02:18.24 | jonpry | i really don't think setup_mm-for_reboot it turning mmu off |
02:18.37 | jonpry | you should probably reanalyze that |
02:20.13 | detule | oh yeah i said something definitely not true |
02:20.41 | detule | it just takes out an identity map |
02:21.24 | detule | still that mmu_off in cpu reset is earlier than putting it in relocate_kernel |
02:22.18 | detule | not by much |
02:25.15 | jonpry | the trouble is that the place relocate_kernel gets copied to. might be above the identity map |
02:25.27 | jonpry | that isn't the case on my lowly 512mb phone |
02:25.53 | detule | ok but don't we pass the physical location of where it gets copied to |
02:26.16 | jonpry | it still must be in the identity map |
02:26.27 | detule | oh |
02:26.39 | detule | that would be my problem |
02:26.56 | jonpry | only a function running inside of an identity map can achieve mmu off |
02:28.30 | detule | i have cpu_reset in the mapping but not relocate_kernel |
02:29.21 | detule | if i leave cpu_reset turn the mmu off is this still an issue? |
02:31.05 | *** part/#htc-linux ali1234 (~ajbuxton@s15821883.onlinehome-server.info) |
02:31.35 | *** join/#htc-linux ali1234 (~ajbuxton@s15821883.onlinehome-server.info) |
02:36.44 | jonpry | detule, no i don't think so |
03:36.55 | *** join/#htc-linux Rajko (~rajkosto@wan.rajkonet.info) |
03:40.01 | *** join/#htc-linux leviathan_ (~quassel@2001:470:26:484:6ef0:49ff:fee6:8dca) |
06:40.44 | *** join/#htc-linux kiozen (~kiozen@p578a42db.dip0.t-ipconnect.de) |
06:44.15 | *** join/#htc-linux kiozen (~kiozen@p578a42db.dip0.t-ipconnect.de) |
06:46.37 | *** join/#htc-linux walter79 (~walter79@31-17-62-189-dynip.superkabel.de) |
06:55.09 | *** join/#htc-linux ychavan (ychavan@nat/redhat/x-qstxpeigahkqjkpp) |
07:02.34 | *** join/#htc-linux Alex[sp3dev] (~alexander@195.208.175.17) |
07:21.13 | *** join/#htc-linux eR^zeRa` (~zzeratul@88.103.98.168) |
07:25.20 | *** join/#htc-linux kiozen (~kiozen@p578a42db.dip0.t-ipconnect.de) |
07:27.59 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-fff8dd00-39.dhcp.inet.fi) |
07:28.03 | *** join/#htc-linux tmzt (~tmzt@76.244.156.141) |
07:32.34 | *** join/#htc-linux DuperMan (~Duper@93-173-169-101.bb.netvision.net.il) |
07:46.47 | *** join/#htc-linux khorben__ (~dont@mahce.defora.org) |
09:33.53 | *** join/#htc-linux zeusk (~zeusk@h2041382.stratoserver.net) |
09:33.59 | zeusk | hi |
09:34.34 | zeusk | jonpry, why're you invalidating instruction tlb and data tlb after invalidating UTLB ? |
09:39.27 | jonpry | that's how the kernel does it |
09:40.07 | jonpry | i can't speak to the wisdom |
09:40.26 | jonpry | also i think armv7 doesn't actually have UTLB |
09:41.31 | zeusk | ARM docs mention ULTB, infact i can see UTLB operation in HTC SPL |
09:41.47 | zeusk | (for Cortex-A8, armv7-scorpion) |
09:42.17 | zeusk | what problem are you having ? |
09:43.28 | zeusk | you should be disabling caches along/before MMU |
09:43.45 | zeusk | s/MMU/disabling MMU/ |
09:45.59 | zeusk | invalidate L2, disable fpu if on, DSB, Invalidate icache and dcache, disable dcache, icache and mmu, invalidate tlb |
09:46.04 | *** join/#htc-linux Alex[sp3dev] (~alexander@nat.rnd.stcnet.ru) |
09:46.29 | Alex[sp3dev] | finally stupid jtag works on a stupid pandaboard in my stupid hands |
09:46.31 | jonpry | no problem is happening |
09:46.37 | jonpry | it all works pretty good |
09:46.51 | jonpry | good enough to be kanged |
09:47.26 | zeusk | alex, woot, how does it feel to halt the core at own wish ? :D |
09:47.50 | Alex[sp3dev] | the problems were (1) my hands which connected arm 14-pin header to the 20-pin one in the wrong way and (2) a j-link clone that failed to work. olimex usb tiny works fine |
09:47.55 | Alex[sp3dev] | zeusk: feels so 1337 |
09:48.03 | Alex[sp3dev] | (gdb) x 0x4a002204 |
09:48.03 | Alex[sp3dev] | 0x4a002204:0x6b95c02f |
09:48.24 | jonpry | invalidating TLB is probably not critical since it isn't used until mmu is reenabled |
09:48.49 | Alex[sp3dev] | now I need to figure out how to debug linux under jtag |
09:49.39 | zeusk | the only board i have with such capability is a STM cortex-m0 with 8kb ram and no mmu, can i run android on it ? |
09:50.36 | zeusk | jonpry, that's how most software do it, invalidate TLB after turning MMU off but you could be right.. |
09:53.18 | jonpry | that's how i do it too |
09:53.33 | jonpry | just get everything squeaky clean |
10:39.52 | zeusk | scorpion's cache line is 64kb or 128kb ? |
10:52.13 | *** join/#htc-linux BabelO (~wdlxtv@AMontpellier-553-1-42-41.w90-48.abo.wanadoo.fr) |
11:23.37 | *** join/#htc-linux marc1706 (~Marc@phpbb/modifications/marc1706) |
11:34.42 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
11:59.57 | *** join/#htc-linux khorben__ (~dont@mahce.defora.org) |
12:42.20 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
13:00.42 | *** join/#htc-linux kiozen (~kiozen@p578a42db.dip0.t-ipconnect.de) |
14:11.46 | *** join/#htc-linux ychavan (ychavan@nat/redhat/x-zacvvcxqrrndiyam) |
14:22.28 | *** join/#htc-linux AstainHellbring (~AstainHel@unaffiliated/astainhellbring) |
14:37.32 | *** join/#htc-linux BabelO__ (~fcr@AMontpellier-553-1-186-139.w92-133.abo.wanadoo.fr) |
14:39.01 | *** join/#htc-linux BabelO__ (~fcr@AMontpellier-553-1-186-139.w92-133.abo.wanadoo.fr) |
14:39.36 | *** join/#htc-linux BabelO (~fcr@unaffiliated/babelo) |
15:03.51 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-74-207.dynamic.mnet-online.de) |
15:15.14 | *** join/#htc-linux Willd (willd@citu-202.citu.kth.se) |
15:38.17 | *** join/#htc-linux jonpry (~jon@c-67-170-101-76.hsd1.wa.comcast.net) |
15:38.22 | *** join/#htc-linux BabelO (~fcr@unaffiliated/babelo) |
16:09.10 | *** join/#htc-linux detule (~detule@unaffiliated/d3tul3) |
16:14.39 | *** join/#htc-linux rob_w (~bob@unaffiliated/rob-w/x-1112029) |
16:17.01 | *** join/#htc-linux Alex[sp3dev] (d5551202@gateway/web/freenode/ip.213.85.18.2) |
16:46.58 | *** join/#htc-linux Cotulla (~myfakemai@nat100-255-205-109.tvoe.tv) |
17:55.43 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-74-207.dynamic.mnet-online.de) |
18:13.00 | *** join/#htc-linux skodde (~skodde@unaffiliated/skodde) |
20:08.57 | *** join/#htc-linux Alex[sp3dev] (d5551202@gateway/web/freenode/ip.213.85.18.2) |
20:23.18 | *** join/#htc-linux MethoS- (~clemens@134.102.106.250) |
21:00.06 | *** join/#htc-linux rajkosto (~rajkosto@cable-94-189-239-212.dynamic.sbb.rs) |
21:01.08 | *** join/#htc-linux bitrot (~rajkosto@wan.rajkonet.info) |
21:02.16 | *** join/#htc-linux khorben__ (~dont@mahce.defora.org) |
21:58.32 | *** join/#htc-linux arif-ali (~arif-ali@94-192-24-56.zone6.bethere.co.uk) |
22:02.20 | *** join/#htc-linux Alex[sp3dev] (d5551202@gateway/web/freenode/ip.213.85.18.2) |
23:57.35 | *** join/#htc-linux walter79_ (~walter79@31-17-62-189-dynip.superkabel.de) |
23:58.29 | *** join/#htc-linux walter79 (~walter79@31-17-62-189-dynip.superkabel.de) |