00:10.47 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi) |
00:39.07 | *** join/#htc-linux |lippa|^ (~lippa@CPE-121-214-10-65.lnse3.win.bigpond.net.au) |
00:42.23 | *** join/#htc-linux the-leviathan (~quassel@c-82-192-226-27.customer.ggaweb.ch) |
00:46.08 | *** join/#htc-linux lipp[a] (~lippa@CPE-58-170-236-230.lnse4.win.bigpond.net.au) |
00:49.10 | *** join/#htc-linux the-leviathan (~quassel@c-82-192-226-27.customer.ggaweb.ch) |
00:50.09 | jonpry | detule, you cracked the code? |
00:58.38 | *** join/#htc-linux lipp[a] (~lippa@CPE-137-147-16-195.lnse7.lon.bigpond.net.au) |
01:02.38 | *** join/#htc-linux |lippa|^ (~lippa@CPE-121-219-76-204.lnse1.lon.bigpond.net.au) |
01:07.09 | *** join/#htc-linux lipp[a] (~lippa@CPE-137-147-3-241.lnse7.lon.bigpond.net.au) |
01:11.09 | *** join/#htc-linux |lippa|^ (~lippa@CPE-121-214-110-208.lnse4.lon.bigpond.net.au) |
01:15.08 | *** join/#htc-linux d3tul3 (~detule@unaffiliated/d3tul3) |
01:17.39 | *** join/#htc-linux lipp[a] (~lippa@CPE-137-147-179-251.lnse7.win.bigpond.net.au) |
01:19.34 | d3tul3 | jonpry, insert enough debugging statements and something will come up |
01:20.43 | d3tul3 | the first fifteen buffers or so have normal addresses 0xb59f4000 and looks like they are actually getting processed....at some point weird buffers start getting queued and the vfe starts spitting out axi errors |
01:22.02 | jonpry | ah |
01:23.08 | *** join/#htc-linux lipp[a] (~lippa@CPE-124-180-131-215.lnse4.lon.bigpond.net.au) |
01:23.38 | d3tul3 | how's the n4? |
01:24.29 | jonpry | i'm trying to boot a custom kernel on it |
01:24.56 | d3tul3 | can't be that hard |
01:25.31 | jonpry | we shall see |
01:25.34 | d3tul3 | just make sure your ramdisk is at a correct offset !:p |
01:29.40 | *** join/#htc-linux |lippa|^ (~lippa@CPE-121-220-77-143.lnse2.win.bigpond.net.au) |
01:35.40 | *** join/#htc-linux lipp[a] (~lippa@CPE-121-220-82-171.lnse2.win.bigpond.net.au) |
01:36.04 | jonpry | it isn't terribly happy |
01:44.49 | d3tul3 | shows signs of life though? |
01:44.58 | jonpry | fixed now |
01:45.05 | jonpry | fully functional |
01:45.19 | d3tul3 | broke a sweat there eh |
01:45.40 | jonpry | i made the mistake of reading a guide |
01:45.50 | jonpry | of foolishness |
01:46.24 | jonpry | this fastboot thing is pretty nice |
01:46.51 | d3tul3 | memboot |
01:47.07 | jonpry | yeah same as touchpad |
01:47.34 | jonpry | not entirely sure why we were trying to flash bootloaders and things |
01:56.01 | jonpry | i think it was that guy "craptain" |
01:56.23 | jonpry | he appealed to my jealous nature :p |
01:56.44 | d3tul3 | who's that |
01:56.52 | d3tul3 | didn't he have to do something with the touchpad |
01:56.55 | d3tul3 | back in the day |
01:57.00 | jonpry | yeah |
01:57.09 | jonpry | <PROTECTED> |
01:57.30 | jonpry | and then i was like wth do i want fastboot, and later i just had to have it |
01:58.04 | d3tul3 | i don't think the s3 has fastboot |
01:58.35 | jonpry | in theory that is fixable |
01:58.43 | jonpry | just chain boot it |
02:00.03 | jonpry | you can probably run moboot on that thing |
02:01.16 | d3tul3 | yeah lk would be a nice idea |
02:01.51 | jonpry | so i guess to work on dsp. i need to make a kernel that breaks audio and pil |
02:21.06 | jonpry | hrm |
02:21.30 | jonpry | i figure this pil_trusted stuff is no good |
02:21.35 | jonpry | you have that? |
02:30.09 | d3tul3 | i don't understand this interaction between pmem and ion on my kernel |
02:30.19 | d3tul3 | pil_trusted? |
02:30.33 | jonpry | pil_q6v4.c |
02:30.45 | jonpry | static struct pil_reset_ops pil_q6v4_ops_trusted |
02:31.09 | jonpry | any idea where the dsp hole is? |
02:31.24 | d3tul3 | yeah it's my 3.4 kernel |
02:31.38 | d3tul3 | there's a whole in my kernel at afe |
02:31.41 | d3tul3 | 0xafe00 |
02:31.50 | d3tul3 | nevermind |
02:32.02 | jonpry | like the memory map is atags right? |
02:32.18 | d3tul3 | yeah |
02:32.36 | jonpry | Memory: 134MB 74MB 7MB 1MB 253MB 1456MB = 1925MB total |
02:33.05 | jonpry | obviously some serious pmem |
02:33.44 | d3tul3 | you see something like 08-11 22:01:05.240 WARN : Early memory PFN ranges |
02:33.45 | d3tul3 | 08-11 22:01:05.240 WARN : 0: 0x00080200 -> 0x00088d00 |
02:33.45 | d3tul3 | 08-11 22:01:05.240 WARN : 0: 0x0008f200 -> 0x0008f700 |
02:33.45 | d3tul3 | 08-11 22:01:05.240 WARN : 0: 0x00090000 -> 0x000afe00 |
02:33.45 | d3tul3 | 08-11 22:01:05.240 WARN : 0: 0x000ba702 -> 0x000ea000 |
02:33.45 | d3tul3 | 08-11 22:01:05.240 WARN : 0: 0x000ea100 -> 0x000f7cff |
02:33.52 | d3tul3 | plenty of holes there |
02:34.15 | d3tul3 | i know afe000 is where vidc loads firmware on my kernel so that hole is there for a reason |
02:34.28 | jonpry | so how to correlate a hole with something? |
02:34.57 | d3tul3 | yeah no idea write to it with dev/mem and see what breaks |
02:35.15 | d3tul3 | i came across what afe00 is used purely by chance on my kernel |
02:35.45 | d3tul3 | wait dev mem won't let you write there nevermind |
02:35.51 | jonpry | pfn_shift=12? |
02:37.38 | d3tul3 | yes |
02:38.27 | jonpry | 80200000-887fffff : System RAM |
02:38.28 | jonpry | <PROTECTED> |
02:38.28 | jonpry | <PROTECTED> |
02:38.28 | jonpry | 89000000-8d9fffff : System RAM |
02:38.28 | jonpry | 8f700000-8fdfffff : System RAM |
02:38.28 | jonpry | 8ff00000-8fffffff : System RAM |
02:38.30 | jonpry | 90100000-9fdfffff : System RAM |
02:38.32 | jonpry | a3900000-fe9fefff : System RAM |
02:40.25 | *** join/#htc-linux DuperMan (Duper@85-250-37-220.bb.netvision.net.il) |
02:46.50 | jonpry | fwiw the qdsp entry point is at 0x8da0 0000 |
02:47.41 | d3tul3 | how did you find it |
02:47.45 | d3tul3 | hole # 1 |
02:48.25 | jonpry | i patched pil loader thing |
02:48.43 | jonpry | i'm not sure that the entire hole belongs to it |
02:49.17 | fakker | slaps jonpry |
02:49.49 | jonpry | need to dump this whole elf header |
02:54.20 | d3tul3 | i don't get this, it's not like it allocates physical buffer multiple times here |
02:54.32 | d3tul3 | it allocates it once, and then keeps using the same thing |
02:54.55 | d3tul3 | so how the hell does it relocate to an abnormal physical location |
02:57.10 | jonpry | memory corruption maybe? |
03:00.23 | d3tul3 | maybe when buffers are released to userspace something bad happens.... |
03:21.09 | jonpry | <PROTECTED> |
03:21.10 | jonpry | <PROTECTED> |
03:21.10 | jonpry | <PROTECTED> |
03:21.10 | jonpry | <PROTECTED> |
03:21.10 | jonpry | <PROTECTED> |
03:21.11 | jonpry | <PROTECTED> |
03:23.44 | jonpry | looks like all of 8da to 8f7 - 1 |
03:23.54 | jonpry | yours is the same btw |
03:25.16 | d3tul3 | you see some 8960 tags there? |
03:25.29 | jonpry | you sent me the firmware |
03:25.30 | d3tul3 | or it's not device specific |
03:25.46 | jonpry | readelf -a q6.mdt |
03:26.01 | d3tul3 | thanks |
03:29.02 | jonpry | gtg |
04:05.17 | d3tul3 | Chad " Installing a custom version of Linux that is used by Qualcomm developers. It has all their tools installed by default. " |
04:05.28 | d3tul3 | i heard Chad |
04:05.34 | d3tul3 | s/heard/heart/ |
04:23.05 | *** join/#htc-linux Ondalf (~ondalf@cable-roi-50ddf8-39.dhcp.inet.fi) |
04:46.29 | *** join/#htc-linux d3tul3 (~detule@unaffiliated/d3tul3) |
04:46.52 | d3tul3 | jonpry, fixed it |
04:52.11 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
05:06.22 | *** join/#htc-linux DuperMan (~Duper@85-250-37-220.bb.netvision.net.il) |
05:12.47 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
05:42.42 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
05:52.29 | jonpry | d3tul3, no shit |
05:52.36 | jonpry | what was it? |
06:42.08 | *** join/#htc-linux eR^zeRa` (~zzeratul@ip-84-42-202-42.net.upcbroadband.cz) |
06:53.34 | *** join/#htc-linux kiozen (~kiozen@p578a42db.dip0.t-ipconnect.de) |
07:03.34 | *** join/#htc-linux lamikr (lamikr@nat/nokia/x-layzkbbfumkmqtkd) |
07:21.43 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
07:26.43 | *** join/#htc-linux |lippa|^ (~lippa@CPE-121-219-187-124.lnse2.lon.bigpond.net.au) |
08:17.12 | *** join/#htc-linux lipp[a] (~lippa@CPE-138-217-88-187.lnse4.lon.bigpond.net.au) |
09:05.37 | *** join/#htc-linux a-st (~a-st@unaffiliated/a-st) |
10:55.12 | *** join/#htc-linux |lippa|^ (~lippa@CPE-124-176-178-128.lns3.win.bigpond.net.au) |
10:59.42 | *** join/#htc-linux lipp[a] (~lippa@CPE-121-219-38-122.lnse1.lon.bigpond.net.au) |
11:12.42 | *** join/#htc-linux |lippa|^ (~lippa@CPE-121-219-244-246.lnse1.win.bigpond.net.au) |
11:46.52 | *** join/#htc-linux teemodk (~teemodk@80-62-117-30-mobile.dk.customer.tdc.net) |
11:55.25 | *** join/#htc-linux d3tul3 (~detule@unaffiliated/d3tul3) |
11:55.42 | *** join/#htc-linux teemodk (~teemodk@80-62-117-30-mobile.dk.customer.tdc.net) |
12:06.01 | *** join/#htc-linux gauner1986 (~Adium@87.253.171.203) |
12:16.59 | *** join/#htc-linux nearst (~c0re@pdpc/supporter/monthlybyte/nearst) |
12:35.46 | *** join/#htc-linux helicopter88 (~helicopte@host83-54-dynamic.40-79-r.retail.telecomitalia.it) |
13:07.31 | *** join/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
13:09.18 | Cotulla | hey |
13:09.33 | nearst | hi |
13:31.02 | *** join/#htc-linux d3tul3 (~detule@unaffiliated/d3tul3) |
13:49.29 | *** join/#htc-linux jonpry (~jon@2602:306:c417:8aa0:e107:cfb:6290:8135) |
14:02.07 | Marc | hi |
14:02.33 | Marc | Cotulla: changing initrd offset to 16M fixes boot with initrd.gz |
14:03.00 | Marc | tested with cLK with 16M offset and kernel compiled with 0x12800000 offset (I think that was it) |
14:04.19 | Cotulla | yes |
14:06.07 | Marc | and audio init fail seems to be related to SMD |
14:07.18 | Cotulla | did u found reason? |
14:08.14 | Marc | dal_open_channel fails |
14:08.23 | Marc | I yet have to find out where exactly |
14:09.52 | Marc | yey, printk ftw |
14:11.23 | Cotulla | u slow |
14:11.24 | Cotulla | :( |
14:12.35 | Marc | what? :D |
14:12.39 | Marc | I'm making it pretty :P |
14:12.52 | Marc | and I didn't do anything yesterday |
14:13.35 | Cotulla | okay |
14:14.17 | Marc | btw, I'm actually seeing the second boot animation now ^^ |
14:14.42 | Cotulla | cool |
14:16.19 | Marc | smd_open: -19 |
14:16.37 | Marc | that's also what qmi_open is saying |
14:16.47 | detule | smd_open fails at start always |
14:17.00 | Marc | then fix it :P |
14:17.11 | Cotulla | detule u sure it's our case? |
14:17.22 | Marc | ENODEV |
14:17.24 | detule | it does for me...it fails before the modem is up |
14:17.38 | Marc | it just keeps reporting -19 ;) |
14:17.45 | Marc | so I'm guessing that is not the case here |
14:19.40 | detule | 08-19 05:37:26.164 INFO : pil pil2: modem_fw: Brought out of reset |
14:19.40 | detule | 08-19 05:37:28.715 INFO : pil pil1: modem: Brought out of reset |
14:19.40 | detule | 08-19 05:37:28.720 ERROR: smd_tty_open: DATA1 open failed -19 |
14:19.40 | detule | 08-19 05:37:28.724 ERROR: smd_tty_open: DATA1 open failed -19 |
14:19.40 | detule | 08-19 05:37:28.729 ERROR: smd_tty_open: DATA1 open failed -19 |
14:19.40 | detule | 08-19 05:37:28.733 ERROR: smd_tty_open: DATA1 open failed -19 |
14:20.38 | Cotulla | delute it's not applicable |
14:20.41 | Cotulla | it's MSM8960 |
14:20.50 | Cotulla | we are talking about our dear QSD8250B |
14:20.55 | Cotulla | ^^ |
14:21.28 | detule | i am just saying i've seen smd open failing on the rhod as well at boot....might not be the same for you guys |
14:22.06 | detule | Cotulla, ' detule ' doesn't mean anything in russian? |
14:23.11 | Cotulla | nothing |
14:24.25 | Marc | I have the feeling smd_get_channel fails big time |
14:24.27 | Marc | let's see |
14:25.27 | Marc | In case of emergency: blame detule |
14:25.40 | Cotulla | ? |
14:25.48 | Marc | just need someone to blame :P |
14:25.59 | Marc | why is fastboot boot so freaking slow? -.- |
14:25.59 | Cotulla | bad approach |
14:26.01 | Cotulla | just fix it |
14:26.07 | Cotulla | from LK? |
14:26.12 | Marc | yeah -.- |
14:26.22 | Marc | total time is always something like 55 seconds |
14:26.35 | Cotulla | cuz it's LK |
14:26.36 | Cotulla | :) |
14:26.58 | Marc | fix magldr and I'm back to that :) |
14:27.34 | Marc | and yes, smd_get_channel fails |
14:29.55 | detule | there's all those magic offsets when looking for those smd channels |
14:30.42 | Marc | yeah -.- |
14:32.24 | Marc | I'll just add something like they fixed there: http://sourceware.org/ml/glibc-cvs/2013-q1/msg00115.html |
14:32.48 | Cotulla | :D |
14:34.17 | Marc | who would've thought that TWO is in fact -2.0? ^^ |
14:35.59 | Marc | ok, we are looking for channel: DSP_DAL_AQ_AUD, type: 0 |
14:36.21 | Marc | left is always the looked for value, right is what we actually have |
14:36.24 | Marc | name: DSP_DAL_AQ_AUD vs. DAL_AQ_AUD, type: 0 vs. 1 |
14:36.30 | Marc | fail there ^^ |
15:12.14 | Marc | Cotulla: looking at the CAF qdsp6 driver it looks like we also shouldn't use smd_open() anymore |
15:12.19 | Marc | at least in dal.c |
15:16.24 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
15:23.35 | Marc | Cotulla: http://pastebin.com/9uP6DzyS |
15:27.51 | fakker | http://pic.twitter.com/HZAuEBUm |
15:29.50 | fakker | is hard at work |
15:31.06 | helicopter88 | fakker, that dolan is very nice |
15:31.39 | fakker | thanks :) |
15:33.42 | fakker | helicopter88, http://pic.twitter.com/K53Lb1nB my other one |
15:34.03 | helicopter88 | fakker, the one on the top is nice |
15:34.09 | *** join/#htc-linux nearst (~c0re@pdpc/supporter/monthlybyte/nearst) |
15:34.12 | helicopter88 | the bigger one is a bit less dolan |
15:34.20 | fakker | yup, it was the first quick drawing |
15:34.26 | fakker | i am filling my notebook full of them |
15:34.31 | helicopter88 | lol |
15:35.05 | *** join/#htc-linux a-st (~a-st@unaffiliated/a-st) |
15:41.21 | Marc | fakker: I was just like "porn or bs?" |
15:41.31 | Marc | seems like the latter one |
15:42.12 | fakker | it's not bs |
15:46.07 | helicopter88 | Marc, it's dolan |
16:01.30 | Marc | Cotulla: I just booted into android :) |
16:01.49 | fakker | i can too |
16:01.51 | fakker | turns his phone off |
16:01.51 | Marc | then again, it just made a soft reboot :D |
16:01.53 | fakker | turns his phone on |
16:01.58 | Marc | with the 3.0.16 kernel :D |
16:02.17 | fakker | 3.0.31? |
16:02.26 | Marc | no |
16:02.29 | Marc | 3.0.16 |
16:02.37 | fakker | i did 3.0.31 |
16:02.42 | fakker | catch up, n00bs |
16:03.04 | arrrghhh | RHOD had 3.4 |
16:03.09 | fakker | fu |
16:03.09 | arrrghhh | for like a year now. |
16:03.17 | arrrghhh | trolololol |
16:03.20 | fakker | pls see my twitpic link |
16:03.22 | fakker | read text |
16:03.24 | fakker | and do so |
16:03.31 | helicopter88 | runs 3.0.62 |
16:03.34 | arrrghhh | hah |
16:03.45 | fakker | :p |
16:04.55 | fakker | fuck all of you |
16:04.59 | fakker | i am outta here |
16:06.47 | *** join/#htc-linux paulk-desktop (~paulk@lib33-1-82-233-88-171.fbx.proxad.net) |
16:10.21 | nearst | lol |
16:12.36 | Cotulla | fixed? |
16:17.55 | Marc | Cotulla: yep, audio seems to actually start now |
16:18.01 | Marc | and I'm at the lockscreen :D |
16:18.07 | Cotulla | how u fixed it? |
16:18.11 | Marc | but it reboots after a few seconds of doing anything |
16:18.44 | nearst | what u all working atm |
16:19.13 | Marc | https://github.com/marc1706/android_kernel_leo_3.0.16/commit/bacfe30e82e0b0095cc633f86f26f90468457cf1 |
16:19.42 | Marc | lol, I actually just saw that I nubbed something :) |
16:20.23 | nearst | ouh, mach-msm :) |
16:21.27 | Marc | now: https://github.com/marc1706/android_kernel_leo_3.0.16/commit/25ac01da5809435b4a8a2f9d9a82244023586b39 |
16:22.54 | Cotulla | looks not bad |
16:23.02 | Marc | Cotulla: looking at the 3.0.16 qdsp6 from codeaurora actually helped :) |
16:23.37 | Marc | mounting actually seems to fail as far as I can see |
16:24.06 | Marc | and the stupid rom is trying to load bcm4329 |
16:24.13 | Marc | which we do not have in the 3.0.16 kernel right now |
16:27.15 | Cotulla | just 3.0? |
16:32.37 | Marc | what do you mean with just 3.0? |
16:33.37 | Marc | well I could add the bcm4329 driver to the 3.0 kernel but the bcmdhd driver is actually the newer driver |
16:33.47 | Marc | a combined driver for bcm4329 and bcm4330 |
16:34.38 | Cotulla | which version is there? |
16:34.43 | Cotulla | 5.X? |
16:37.55 | Marc | 5.90.195.75 |
16:38.27 | Marc | took a while to open that file as my computer is freakishly slow right now :D |
16:38.34 | Marc | converting a video ^^ |
16:39.05 | Marc | and I have like 420 MB of RAM left xD |
16:39.21 | Marc | of 6 GB :D |
16:40.26 | detule | we had bcm4329 running on 3.x on the rhod....perhaps grabbing them from our tree is not a bad idea |
16:40.46 | detule | i never figured out where jonpry got them from....some weird lineage |
16:41.38 | Cotulla | yes new |
16:41.50 | Cotulla | code quality is still same |
16:41.54 | Cotulla | however |
16:42.15 | Marc | hehe |
16:46.10 | Marc | it really sucks that you can see the pixels of the leo screen while shooting a video in full hd |
16:46.18 | Marc | stupid leo screen :P |
16:46.43 | Marc | smashes the screen with a hammer to increase pixel density |
16:50.56 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
17:03.06 | Cotulla | jonpry ping |
17:07.35 | Marc | jumps in front of junpry |
17:07.40 | Marc | jumps in front of jonpry |
17:07.41 | Marc | :P |
17:07.43 | Marc | pong |
17:07.44 | Marc | :P |
17:11.29 | Marc | Cotulla: also fixed the missing code for creating the partitions in msm_nand: https://github.com/marc1706/android_kernel_leo_3.0.16/commit/0dec1ff54f777f4f2cf795709b107a828f31b1b1 |
17:12.45 | Cotulla | why no comments? |
17:12.57 | Marc | comments for that function? |
17:13.02 | Marc | or in the commit? ^^ |
17:13.06 | Cotulla | for commit, for changes |
17:13.19 | Marc | should be self-explanatory |
17:13.23 | Marc | can add them if you want though |
17:13.36 | Cotulla | well |
17:13.42 | Cotulla | I don't really know why u removed "\n" :D |
17:13.47 | Cotulla | I have a guess |
17:13.48 | Cotulla | but still |
17:13.51 | Marc | because it's stupid |
17:13.58 | Cotulla | but it worked before |
17:15.25 | Marc | there, fixed: https://github.com/marc1706/android_kernel_leo_3.0.16/commit/ef3a4c7e1975feaeebe12b4c58f3e9baad82510d |
17:15.45 | Cotulla | yeah better now |
17:16.20 | fakker | stupid |
17:16.33 | Cotulla | fakker. |
17:16.36 | fakker | nice work Marc |
17:16.52 | Marc | thx |
17:17.04 | Marc | I was wondering why adb wasn't working; this explains it: [USB] init(): Failed to register androidplatform driver |
17:17.34 | *** join/#htc-linux nearst (~nearst@pdpc/supporter/monthlybyte/nearst) |
17:17.35 | Marc | oh, and rpc is still really broken ^^ |
17:26.10 | Cotulla | <PROTECTED> |
17:26.25 | Cotulla | pff in kernel |
17:26.26 | helicopter88 | /Tiem to eat them |
17:30.18 | Cotulla | "which should not roll over until Hell freezes over" :O |
18:07.57 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
18:22.49 | *** join/#htc-linux a-st (~a-st@unaffiliated/a-st) |
18:54.12 | *** join/#htc-linux kiozen (~kiozen@ppp-93-104-67-180.dynamic.mnet-online.de) |
19:13.52 | *** join/#htc-linux rajkosto (~rajkosto@wan.rajkonet.info) |
19:51.03 | *** join/#htc-linux ltxda (~anon@c-98-194-107-10.hsd1.tx.comcast.net) |
19:51.06 | *** join/#htc-linux ltxda (~anon@unaffiliated/ltxda) |
19:52.06 | jonpry | Cotulla, pong |
19:53.06 | Cotulla | hey |
19:53.12 | Cotulla | looked today to q6 kernel more |
19:53.33 | Cotulla | MMU still not clear |
19:54.05 | jonpry | what is the problem? |
19:54.17 | jonpry | i'm working on using PIL on v4 |
19:54.28 | Cotulla | most SMP are using CPU id inside switch_mm() I assume it's done inside HVM |
19:56.17 | Cotulla | also |
19:56.30 | Cotulla | page table entries should be mapped to TLB |
19:56.39 | Cotulla | itself |
19:56.48 | Cotulla | to be able access them at tlbmiss |
19:57.06 | Cotulla | for the first level it's not hard |
19:57.09 | Cotulla | just 6 x 4096 |
19:57.23 | Cotulla | but the second level is not clear |
19:57.54 | Cotulla | there is pte_alloc_one() which just allocate one kernel page |
19:58.01 | Cotulla | but not clear how it's mapped |
19:59.47 | jonpry | i'll have to look at this |
19:59.59 | jonpry | not clear on how this is all implemented on linux |
20:00.35 | Cotulla | do u got my doubts? |
20:00.47 | Cotulla | the first level 1024 x 4 |
20:00.54 | Cotulla | each describe 4mb |
20:01.03 | Cotulla | second level 1024 x 4 - each describe 4K |
20:02.27 | Cotulla | so we need to be able access second level |
20:02.30 | Cotulla | inside tlbmiss |
20:09.33 | *** join/#htc-linux a-st_ (~a-st@unaffiliated/a-st) |
20:19.51 | *** join/#htc-linux BabelO (~fcr@2a01:e35:8adc:3150:21e:8cff:fe2d:7d3f) |
20:19.51 | *** join/#htc-linux BabelO (~fcr@unaffiliated/babelo) |
20:52.11 | *** join/#htc-linux a-st (~a-st@unaffiliated/a-st) |
21:13.44 | *** join/#htc-linux helicopter88_2 (~helicopte@host161-93-dynamic.22-79-r.retail.telecomitalia.it) |
21:25.14 | *** part/#htc-linux Cotulla (~myfakemai@109.205.253.11) |
23:07.37 | *** join/#htc-linux detule (~d3tul3@unaffiliated/d3tul3) |