00:04.52 | *** part/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
03:14.53 | *** join/#htc-linux Bry8Star (~Bry8Star@gateway/tor-sasl/bry8star) |
03:45.43 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
04:10.58 | *** join/#htc-linux bartman (~bart@2607:f2c0:f00e:700::dd) |
04:58.36 | *** join/#htc-linux d3tul3 (~detule@unaffiliated/d3tul3) |
05:16.13 | *** join/#htc-linux eR^zeRa` (~zzeratul@ip-84-42-202-42.net.upcbroadband.cz) |
06:00.16 | *** join/#htc-linux DuperMan (Duper@85-250-109-40.bb.netvision.net.il) |
06:16.40 | *** join/#htc-linux Rajko (~rajkosto@wan.rajkonet.info) |
06:28.55 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
06:50.50 | *** join/#htc-linux kiozen (~kiozen@p578a42db.dip0.t-ipconnect.de) |
07:23.10 | *** join/#htc-linux DuperMan (Duper@93-172-160-252.bb.netvision.net.il) |
08:20.55 | *** join/#htc-linux lamikr (lamikr@nat/nokia/x-mwlghdedojnejdud) |
09:08.29 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
09:38.26 | *** join/#htc-linux Bry8Star (~Bry8Star@gateway/tor-sasl/bry8star) |
10:36.35 | *** join/#htc-linux noobhands_ (d03032a2@gateway/web/freenode/ip.208.48.50.162) |
10:43.16 | *** join/#htc-linux gauner1986 (~Adium@87.253.171.220) |
10:43.49 | fakker | sup gauner1986 |
10:45.35 | gauner1986 | hey fakker |
10:50.01 | fakker | hows tricks? |
10:51.04 | gauner1986 | currently boring myself at work |
10:51.56 | fakker | SNAP |
10:52.23 | stybla | gauner1986, you could be my twin. |
10:56.26 | gauner1986 | hehe.. but today is an exception actually. ;) |
10:56.50 | fakker | mine isn't |
10:57.07 | gauner1986 | hope to buy a htc m7 soon |
10:57.18 | fakker | hmm |
10:57.30 | fakker | you still got the iphone? |
10:57.42 | gauner1986 | yeah - company phone |
10:57.50 | fakker | i still got mine - but i left that company |
10:57.51 | fakker | haha |
11:08.41 | stybla | gauner1986: lucky you. I'm quite constantly bored at work. |
11:08.59 | gauner1986 | styble: what's your job? |
11:09.17 | stybla | gauner1986: quality assurance/sustaining. |
11:09.31 | gauner1986 | software? |
11:09.37 | stybla | yes |
11:09.58 | gauner1986 | ah ok.. i'm sitting right next to my qa-team. ;) |
11:10.08 | gauner1986 | but i'm on the developer-side luckily |
11:11.40 | stybla | I don't think that's the problem in general way. |
11:13.35 | stybla | anyway ... |
11:14.18 | *** join/#htc-linux Rajko (~rajkosto@cable-94-189-245-214.dynamic.sbb.rs) |
11:15.02 | *** join/#htc-linux bitrot (~rajkosto@wan.rajkonet.info) |
11:53.01 | *** join/#htc-linux d3tul3 (~detule@pool-96-234-132-77.bltmmd.east.verizon.net) |
12:41.34 | *** join/#htc-linux conantroutman (~conantrou@cpc5-pert5-2-0-cust768.16-4.cable.virginmedia.com) |
13:03.03 | noobhands | http://www.gizmodo.co.uk/2013/02/this-is-what-happens-when-your-baker-doesnt-understand-english/ |
13:16.33 | *** join/#htc-linux d3tul3 (~detule@unaffiliated/d3tul3) |
14:05.00 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
14:10.07 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
14:13.42 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
14:31.55 | *** join/#htc-linux arif-ali (~arif-ali@81.144.163.60) |
15:51.23 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-93-124.dynamic.mnet-online.de) |
17:04.11 | *** join/#htc-linux rudeblunt (4fa80aec@gateway/web/freenode/ip.79.168.10.236) |
17:04.27 | rudeblunt | hey cockpoo? |
17:04.43 | rudeblunt | ~seen Cotulla |
17:04.54 | apt | cotulla <~myfakemai@109.205.253.11> was last seen on IRC in channel #htc-linux, 1d 1h 7m 59s ago, saying: 'he is not here'. |
17:04.55 | rudeblunt | ~seen Cottula |
17:04.56 | apt | i haven't seen 'cottula', rudeblunt |
17:05.10 | rudeblunt | ~seen Marc |
17:05.10 | apt | marc is currently on #htc-linux (7d 1h 50m 16s). Has said a total of 119 messages. Is idling for 1d 1h 24m 50s, last said: 'I did a quick search and couldn't find it'. |
17:05.21 | rudeblunt | so |
17:05.35 | fakker | hey cockpoo |
17:05.46 | rudeblunt | I've managed to boot the current 3.06 leo kernel |
17:06.06 | Marc | ok |
17:06.13 | Marc | with loading the stupid initrd? |
17:06.17 | rudeblunt | not by creating a boot.img |
17:06.26 | rudeblunt | yep |
17:06.40 | rudeblunt | i think I've confirmed cockpoo's theory |
17:07.05 | rudeblunt | if you put initrd inside kernel |
17:07.14 | rudeblunt | it actually boots |
17:07.42 | rudeblunt | complains about not finding some stuff |
17:07.54 | Marc | which is normal |
17:07.56 | jonpry | uncompressed kernels require less space between kernel and initrd |
17:08.34 | rudeblunt | and stays alive ... (but i think all processes spawned by init crash) |
17:08.46 | Marc | hm, we are currently only using gzip |
17:08.54 | rudeblunt | not even initlogo.rle displays |
17:09.23 | Marc | that might actually be an mdp fail |
17:09.41 | rudeblunt | care to see my dmesg? |
17:09.50 | Marc | sure |
17:09.58 | Marc | it'd be nice if you'd upload your compiled kernel, too |
17:10.24 | Marc | I was actually trying to compile cLK with a 16M offset yesterday |
17:10.35 | Marc | got kind of stuck while creating the nbh ... |
17:10.41 | Marc | and didn't want to brick my HD2 ^^ |
17:12.08 | rudeblunt | http://pastebin.com/0MA3BBtf |
17:13.31 | rudeblunt | actually all I did was uncompressing my initrd and copy it over /usr/initramfs_data.cpio |
17:14.12 | Marc | ok |
17:14.25 | Marc | did you enable blk_dev_ram? |
17:14.49 | Marc | also: the first part of init is complaining about you having the 2.6.32 modules ^^ |
17:14.56 | rudeblunt | yes |
17:14.59 | rudeblunt | I know |
17:15.09 | rudeblunt | but that should't be an issue |
17:15.21 | rudeblunt | (or so I think) |
17:16.05 | Marc | ok, qmi_open is also failing |
17:16.10 | Marc | who would've thought? ^^ |
17:18.20 | *** join/#htc-linux the-leviathan (~quassel@2001:470:26:484:6ef0:49ff:fee6:8dca) |
17:18.49 | rudeblunt | anyway... still kinda busy |
17:19.19 | rudeblunt | but I tink now you have sort of a way to continue Marc ... |
17:19.24 | rudeblunt | *think |
17:19.43 | rudeblunt | even whitout MAGLDR2 |
17:19.44 | rudeblunt | :) |
17:25.03 | rudeblunt | last but not least ... i was trying to boot a NativeSD rom |
17:25.40 | *** join/#htc-linux a-st (~a-st@unaffiliated/a-st) |
17:25.54 | rudeblunt | and kernel config is default |
17:30.49 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
17:34.51 | *** join/#htc-linux gauner1986 (~Adium@hmbg-4d06f88b.pool.mediaWays.net) |
17:42.26 | *** join/#htc-linux gassed (faxed@c-76-30-161-228.hsd1.tx.comcast.net) |
17:53.15 | *** join/#htc-linux the-leviathan (~quassel@c-82-192-226-27.customer.ggaweb.ch) |
18:17.57 | *** join/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
18:18.10 | Cotulla | hello |
18:20.25 | Marc | hi |
18:20.29 | Marc | and nice pics on twitter |
18:20.35 | Cotulla | yea |
18:20.35 | noobhands | fu |
18:23.27 | Cotulla | did u try with boot.img? |
18:24.04 | Marc | <rudeblunt> if you put initrd inside kernel |
18:24.12 | Cotulla | what? |
18:24.19 | Marc | <rudeblunt> I've managed to boot the current 3.06 leo kernel |
18:24.31 | Cotulla | good |
18:24.34 | Cotulla | like I tell |
18:25.11 | Cotulla | I never was cockpoo :P |
18:25.15 | Cotulla | it was some another guy |
18:36.16 | rudeblunt | sorry cockpoo |
18:36.18 | rudeblunt | :) |
18:37.20 | rudeblunt | well, now there is another issue |
18:38.45 | rudeblunt | i tried to flash a boot image with a bigger initrd offset |
18:39.19 | rudeblunt | but MAGLDR could never find the kernel |
18:39.20 | Cotulla | ? |
18:39.34 | rudeblunt | remember? |
18:39.50 | rudeblunt | you said : create a boot.img and flash it? |
18:40.16 | rudeblunt | i used abootimg |
18:40.33 | rudeblunt | and created it |
18:41.16 | rudeblunt | but MAGLDR always thought NAND device was yaffs2 and failed with "Kernel not found" or something... |
18:41.52 | rudeblunt | so I gave up and ended placing initramfs inside kernel |
18:44.33 | Cotulla | hm |
18:44.37 | Cotulla | but it supports two formats |
18:44.41 | Cotulla | I tested that |
18:57.52 | rudeblunt | anyway... everything will be fixed in MAGLDR2 ... riiiiiiiiiiiiight? |
18:58.02 | Cotulla | I increased offset to 16 |
18:58.05 | Cotulla | maybe should set to 32 |
18:58.10 | Cotulla | it works for me |
18:58.39 | rudeblunt | it must be that |
18:58.40 | Cotulla | but new region system makes not clear which config it should take by loading kernel as file from SD card |
18:58.50 | Marc | just fyi: offset is set to 16M in cLK's main repository (not in the HD2 version though O.o) |
18:59.10 | Cotulla | but it's not important? |
18:59.19 | Cotulla | boot.img specify initrd separately |
18:59.37 | Marc | it's important if you just flash it to boot |
18:59.51 | Marc | or use the fastboot boot |
18:59.52 | Cotulla | yes but ever fastboot is working with such format |
19:02.01 | rudeblunt | maybe I was doing someting wrong ... I created a boot.img and flashed it to boot partition with flash_image ... |
19:04.16 | Cotulla | u need attributes for partition |
19:04.28 | Cotulla | there were 4 types |
19:04.39 | Cotulla | YAFFS2 Kernel/Recovery and RAW |
19:05.54 | rudeblunt | got it ... so my bad |
19:06.24 | Cotulla | MAGLDR2 doesn't include YAFFS2 support |
19:06.27 | Cotulla | only RAW images |
19:07.20 | rudeblunt | k |
19:11.40 | jonpry | Cotulla, how is qdsp port? |
19:15.44 | Cotulla | looking to stuffs jonpry |
19:15.53 | Cotulla | looking for possible replacements |
19:15.55 | Cotulla | for commands |
19:16.02 | Cotulla | *HVM commands |
19:17.21 | *** part/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
19:17.40 | *** join/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
19:17.53 | Cotulla | u jonpry? |
19:19.49 | Cotulla | also fundamental problems like divide 64 TLBs to 6 fixed |
19:19.53 | Cotulla | or make it dynamic |
19:20.01 | Cotulla | tracked by tlbmiss |
19:20.25 | Cotulla | like 10 TLB per thread |
19:20.32 | Cotulla | and 4 system ones |
19:22.18 | Cotulla | I think I will test SW MMU in qMAGLDR before kernel |
19:30.39 | jonpry | i don't think its't that complicated |
19:31.06 | jonpry | like there is a kernel page table |
19:31.19 | jonpry | and then per process page table |
19:32.13 | jonpry | and when a core schedules a task. we set just keep track of which user pagetable is assigned to every core |
19:32.26 | jonpry | so on tlb miss |
19:32.51 | jonpry | if it's kernel region. just use kernel page tables and create a global entry in the FILO |
19:33.10 | jonpry | if not we look up what process is schedules, and create ASID specific entry |
19:33.56 | Cotulla | right |
19:34.03 | Cotulla | but u need debug that |
19:34.33 | jonpry | even that isn't so bad |
19:34.46 | jonpry | cause like the kernel will boot all the way to init with no user page table support |
19:35.12 | jonpry | so we can have ramconsole and everything working first |
19:35.25 | Cotulla | right too |
19:39.33 | jonpry | how does ring0 protection work? |
19:39.58 | jonpry | i don't see anything in the docs that suggest how to protect memory |
19:40.33 | Cotulla | what kind protection? |
19:43.58 | Cotulla | u mean |
19:44.01 | Cotulla | kernel / user mode I guess |
19:44.02 | Cotulla | ? |
19:45.19 | jonpry | yeah |
19:46.02 | Cotulla | hm I don't see inside TLB |
19:46.05 | Cotulla | such entries |
19:46.31 | Cotulla | maybe |
19:46.35 | Cotulla | also via different ASID? |
19:47.22 | Cotulla | or some special mode |
19:47.25 | Cotulla | hm |
19:47.38 | Cotulla | I guess it should exist |
19:48.38 | jonpry | syscall is done with a trap0 |
19:49.32 | Cotulla | Linux one |
19:49.34 | Cotulla | , |
19:49.36 | Cotulla | ? |
19:49.58 | jonpry | user to kernel |
19:50.02 | Cotulla | yes |
20:06.00 | jonpry | there is PTE_U |
20:07.14 | Cotulla | ? |
20:07.55 | jonpry | page table flag |
20:08.02 | jonpry | page accessible to user mode |
20:08.30 | jonpry | the trick is dropping into usermode |
20:08.45 | jonpry | GSM.UM looks like it has something to do with it |
20:09.07 | jonpry | er. GSR |
20:09.18 | Cotulla | oh in HVM |
20:09.51 | Cotulla | what is GSR? |
20:10.04 | jonpry | guest status register |
20:10.12 | Cotulla | hm |
20:10.17 | Cotulla | sound like it fake |
20:10.43 | jonpry | must have some hardware support |
20:10.47 | jonpry | or TLB would hit |
20:11.01 | Cotulla | well it can use ASIDs as well |
20:11.13 | Cotulla | or clear TLB at switch |
20:11.36 | jonpry | but then kernel can't use global entries |
20:11.56 | Cotulla | right |
20:13.40 | jonpry | and have to flush tlb on context switch exit |
20:14.16 | Cotulla | I am loking |
20:27.01 | *** join/#htc-linux BabelO (~fcr@unaffiliated/babelo) |
20:29.37 | jonpry | doing it with ASID is going to be nasty |
20:29.54 | jonpry | the whole syscall will need to be redone |
20:30.22 | Cotulla | yea |
20:30.26 | jonpry | cause like trap0 will then cause a TLB miss |
20:31.08 | jonpry | somehow the miss will have to cause the syscall to happen |
20:31.46 | jonpry | but i guess that is probably the easiest thing |
20:32.30 | Cotulla | looks like it has supervisor and user mode |
20:33.42 | jonpry | but we don' have any idea how it is programmed |
20:43.41 | Cotulla | I will find it I think |
21:02.26 | *** join/#htc-linux gauner1986 (~Adium@hmbg-4d06f88b.pool.mediaWays.net) |
21:14.02 | *** join/#htc-linux zeusk_ (dfbe5524@gateway/web/freenode/ip.223.190.85.36) |
21:14.04 | zeusk_ | hi |
21:14.32 | Cotulla | hey |
21:15.11 | zeusk_ | it gets to init now ? |
21:15.17 | Cotulla | who |
21:15.25 | zeusk_ | 3.0, leo |
21:15.28 | zeusk_ | linux |
21:15.31 | Cotulla | long time already |
21:15.48 | zeusk_ | hm |
21:23.15 | zeusk_ | jonpry, you got nexus 4 ? |
21:26.39 | Cotulla | he is busy |
21:26.51 | Cotulla | he is trying to start engine in his boat :) |
21:27.31 | zeusk_ | lol |
21:36.51 | Cotulla | looks like they have a special tlb bit |
21:37.03 | Cotulla | which is kinda super visor access flag |
21:40.09 | zeusk_ | it has same modes as arm ? (svc, sys, usr etc..) |
21:40.18 | Cotulla | no |
21:40.49 | zeusk_ | then ? |
21:45.20 | Cotulla | svc and normal seems |
21:46.24 | zeusk_ | hmm |
21:51.35 | Cotulla | his brain blows up |
22:01.56 | jonpry | i got engine going good |
22:02.57 | jonpry | http://www.youtube.com/watch?v=Wl2ZayrBlL4 |
22:04.14 | jonpry | it's better now. turns out that was not running on all cylinders |
22:05.22 | detule | what the hell |
22:06.26 | jonpry | ? |
22:07.07 | detule | how much better? |
22:07.22 | detule | damnit should've been an engineer |
22:07.26 | jonpry | well it can move the boat |
22:07.33 | Cotulla | cool |
22:07.36 | jonpry | and doesn't smoke |
22:07.46 | jonpry | or leave diesel fuel all over the water |
22:07.54 | Cotulla | how many cylinders? |
22:07.58 | jonpry | 4 |
22:08.28 | jonpry | it's like 30kw and 400 kg |
22:08.38 | Cotulla | how many h.p? |
22:08.42 | jonpry | 40 |
22:08.53 | Cotulla | good |
22:09.04 | Cotulla | ;) |
22:09.29 | jonpry | i might take it on the maiden voyage this week |
22:09.34 | Cotulla | intersting if trap0 and trap1 goes to same supervisor mode or not |
22:09.46 | jonpry | i think so |
22:09.59 | jonpry | and hvm drops back to quest for trap0 |
22:10.08 | jonpry | er guest |
22:10.16 | Cotulla | hm yes inside real handler |
22:11.59 | Cotulla | okay |
22:12.11 | Cotulla | so I think it half clear |
22:12.25 | jonpry | i'm going to try and write the TLB in pseudo code later |
22:12.47 | jonpry | i think it needs 7 ASID |
22:12.53 | Cotulla | do u have extinguisher in the boat? |
22:13.04 | jonpry | and 2 permanent entries per core |
22:13.15 | jonpry | it's called the ocean |
22:13.20 | Cotulla | :O |
22:13.20 | Cotulla | :D |
22:13.36 | Cotulla | well |
22:13.50 | Cotulla | why 7 ASIDs? |
22:14.11 | jonpry | context switch just switches to ASID 0 |
22:14.36 | Cotulla | hm |
22:14.41 | Cotulla | and kernel in global |
22:14.46 | Cotulla | supervisor |
22:14.47 | Cotulla | one |
22:14.52 | jonpry | yeah |
22:15.17 | jonpry | each thread needs a PTE_X PTE_R for the tlb handler |
22:15.31 | jonpry | and a PTE_W that the handler can store the context in |
22:15.32 | Cotulla | yes |
22:15.43 | Cotulla | a fixed entry |
22:15.50 | jonpry | 2 |
22:15.59 | Cotulla | maybe cover kernel by one 16M entry |
22:16.21 | jonpry | no highpage? |
22:16.36 | jonpry | dma, or ioremap? |
22:16.41 | Cotulla | hm |
22:16.46 | Cotulla | or we can just copy tlbmiss code |
22:16.49 | Cotulla | to special location |
22:17.00 | jonpry | i think so |
22:17.08 | Cotulla | it can be 64K |
22:17.10 | Cotulla | XRW |
22:17.11 | Cotulla | together |
22:17.17 | jonpry | no! |
22:17.17 | *** join/#htc-linux Mavy-bnc (~mavfree@46.249.37.173) |
22:17.25 | jonpry | can't write kernel from userland |
22:17.38 | Cotulla | but it should be switched to supervisor mode |
22:17.48 | jonpry | yeah but we can't use that |
22:17.52 | Cotulla | at tlbmiss entry |
22:17.56 | jonpry | ASID won't switch on tlbmiss |
22:17.57 | Cotulla | I guess? |
22:18.14 | jonpry | at least i don't think so |
22:18.33 | Cotulla | well but TLB entry with Global should be always valid |
22:18.36 | Cotulla | for any ASIDs? |
22:18.42 | jonpry | hrm. can user programs just change ASID? |
22:18.56 | Cotulla | don't think so |
22:18.56 | jonpry | i think global is not an option |
22:19.01 | jonpry | that's why 7 ASID |
22:19.02 | Cotulla | I think SSR is not accessable from user mode |
22:19.22 | jonpry | how exactly do we set something as user mode? |
22:19.25 | Cotulla | also it looks like TLB with supervisor bit can access anything without rights |
22:19.33 | jonpry | if that can be done, then everything is different |
22:19.36 | Cotulla | I should try that in MAGLDR |
22:19.45 | Cotulla | hm it's good question |
22:19.53 | Cotulla | trapX user -> svc |
22:20.02 | Cotulla | return svc -> user |
22:20.11 | Cotulla | but the initial state . . . |
22:20.44 | jonpry | i couldn't find anything in the kernel source |
22:21.29 | jonpry | just uses vmrte |
22:21.46 | Cotulla | __switch_to |
22:21.47 | Cotulla | maybe? |
22:25.05 | jonpry | don't think so |
22:26.08 | jonpry | i was looking at the GAS sources |
22:26.23 | jonpry | figuring maybe they compile the HVM with it |
22:26.37 | Cotulla | it supports registers |
22:26.42 | Cotulla | yes |
22:26.50 | jonpry | no mention of GSR |
22:26.57 | Cotulla | where u found GSR? |
22:27.05 | Cotulla | I see that name the first time |
22:27.10 | jonpry | 80-nb419-3 |
22:27.15 | Cotulla | oh |
22:27.19 | Cotulla | it's accessed via function |
22:27.25 | Cotulla | I think it's fake |
22:27.28 | Cotulla | just variable |
22:27.46 | Cotulla | there is remark that it can be accessed directly in the feature fversions |
22:27.54 | jonpry | there must be some way to set it |
22:28.04 | jonpry | usermode that is |
22:28.05 | Cotulla | it can be SW solution |
22:28.26 | jonpry | sw solution to make tlb miss? |
22:28.48 | Cotulla | maybe it's possible to change SSR |
22:28.53 | Cotulla | svc -> user |
22:28.59 | jonpry | what is SSR? |
22:29.25 | Cotulla | Supervisor Status Register |
22:29.31 | Cotulla | I think |
22:29.39 | jonpry | where did you here about this? |
22:30.02 | Cotulla | I read it somewhere I think |
22:30.16 | jonpry | gas generates code for it? |
22:30.35 | Cotulla | yes |
22:30.45 | Cotulla | I am used it in qMAGLDR to get thread number for example |
22:30.46 | Cotulla | <PROTECTED> |
22:30.46 | Cotulla | get_hwthread_num: |
22:30.46 | Cotulla | <PROTECTED> |
22:30.46 | Cotulla | <PROTECTED> |
22:30.46 | Cotulla | <PROTECTED> |
22:32.10 | jonpry | <PROTECTED> |
22:32.10 | jonpry | <PROTECTED> |
22:32.10 | jonpry | <PROTECTED> |
22:32.10 | jonpry | <PROTECTED> |
22:32.33 | Cotulla | that ones I also used |
22:32.41 | jonpry | ssr is register 6 on v4 |
22:32.41 | Cotulla | u setup TLBIDX |
22:32.53 | Cotulla | and then specify 64bit value |
22:35.55 | Cotulla | TLB format is differnet between V2 and V4 AFAIK |
22:36.47 | jonpry | REG_ELR? |
22:37.14 | jonpry | http://pastebin.com/ECbw4eb1 |
22:37.26 | Cotulla | ELR is Exception return address |
22:37.52 | Cotulla | like in tlbmiss handler I can get code address |
22:37.55 | Cotulla | with problem |
22:38.09 | jonpry | in v4 gsr is real |
22:38.11 | Cotulla | oh shit stid and htid |
22:38.22 | Cotulla | super visor tid and hyper visor tid? |
22:38.35 | jonpry | gsr is g1 |
22:46.17 | Cotulla | hm bad |
22:50.28 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
22:56.19 | *** join/#htc-linux jonpry (~jon@2602:306:c417:8aa0:15a4:8791:5d38:cca3) |
23:04.16 | jonpry | maybe it's a property of the hypervisor |
23:04.32 | jonpry | and gsr writes work fine on v2? |
23:06.17 | Cotulla | hm don't think so |
23:06.21 | Cotulla | v2 is rather odl |
23:06.22 | Cotulla | old |
23:06.53 | jonpry | somehow the hypervisor is setting this usermode thing |
23:07.53 | jonpry | int reg_htid; /* v2/v3 id was store as part of reg_ssr */ |
23:08.04 | jonpry | int reg_sgp0; /* User mode same as reg_resgp on v2/3 */ |
23:08.14 | jonpry | int reg_stid; /* Re-named TID register */ |
23:11.45 | Cotulla | yes |
23:11.46 | Cotulla | so |
23:11.51 | Cotulla | my code takes tid |
23:11.54 | Cotulla | from ssr |
23:11.56 | Cotulla | on V2 |
23:17.28 | Cotulla | so u need to do a lot of things via different way |
23:17.30 | Cotulla | on V4 |
23:17.32 | Cotulla | looks like |
23:21.22 | jonpry | ugh |
23:21.32 | jonpry | at least usermode is possible on v4 |
23:44.23 | *** join/#htc-linux arrrghhh (~arrrghhh@unaffiliated/arrrghhh) |
23:50.03 | Cotulla | on V2 too |
23:50.09 | Cotulla | there is super bit seems |
23:56.23 | *** join/#htc-linux arrrghhh_ (~arrrghhh@unaffiliated/arrrghhh) |
23:57.03 | jonpry | you cleared it? |
23:57.46 | Cotulla | not yet |
23:57.51 | Cotulla | will try it |
23:58.00 | Cotulla | but in qMAGLDR it's clear |
23:58.47 | Cotulla | maybe create git project |
23:59.10 | Cotulla | or too early |