#milestone-modding | Logs for 2012-02-03

[00:00:27] <zewelor> and that event2 i found in some old script for 2ndboot to simulate power button pressing to turn on screen before doing 2ndboot
[00:00:41] <psychuil> How about a device with no keyboard?
[00:00:55] <zewelor> maybe some other event
[00:01:07] <zewelor> add device 5: /dev/input/event2 name: "cpcap-key"
[00:01:44] <zewelor> maybe there is some more better generic way i dont know it
[00:06:51] <zewelor> maybe u can find some way to do it in android program using api ?
[00:44:05] <nadlabak> pontomedon: I haven't checked the code, but I'm pretty sure it's done via vold command
[00:46:10] <nadlabak> Skrilax_CZ: congrats :)
[00:50:14] <Skrilax_CZ> thanks
[00:52:42] Redh3ad is now known as Red|Off
[01:02:24] -!- n0life has quit [Quit: KTHXBYE]
[01:09:17] -!- RCFout has quit [Quit: This war is mine]
[01:09:31] milaq is now known as milaq|afk
[01:11:22] -!- K4sperle has quit [Ping timeout: 252 seconds]
[01:12:07] -!- K4sperle [K4sperle!~lutz@0x0badc0.de] has joined #milestone-modding
[01:50:27] <pontomedon> nadlabak: thanks
[01:50:48] -!- pontomedon has quit [Quit: I'll be back.]
[01:54:04] -!- `Wedge_ [`Wedge_!~wedge@93-181-238-161.adsl.yaroslavl.ru] has joined #milestone-modding
[01:57:13] -!- `Wedge has quit [Ping timeout: 245 seconds]
[02:11:57] <nadlabak> Skrilax_CZ: compcache source is here https://github.com/a853/a853_compcache , kernel diff here: https://github.com/a853/a853_compcache/blob/master/patches/patch_swap_notify_core_support_2.6.32.diff
[02:14:14] <Skrilax_CZ> nadlabak: ok
[02:35:28] -!- Ave [Ave!ave@kapsi.fi] has joined #milestone-modding
[02:35:34] <Ave> hello
[02:36:46] <Ave> I was trying luks manager on latest cm7 and there is no device mapper?
[02:37:05] <Ave> in kernel or as modules
[02:37:17] <Ave> is it missing on purpose
[02:41:44] Skrilax_CZ is now known as Skrilax_CZ|Away
[02:43:33] <nadlabak> Skrilax_CZ: all that is needed on the android side is here: https://github.com/nadlabak/android_device_motorola_umts_sholes/commit/275629ef95f3b5f1dd831199d039299bee8655e9 (gingerbread-zram branch)
[02:46:09] <nadlabak> all other needed scripts are already in place, compcache can be enabled in the performance menu
[02:51:57] -!- fernando_ [fernando_!~fernandot@189-18-215-26.dsl.telesp.net.br] has joined #milestone-modding
[02:54:00] -!- fernandotakai has quit [Ping timeout: 272 seconds]
[03:07:18] -!- Orgg has quit [Ping timeout: 272 seconds]
[03:11:24] <nadlabak> Ave: device-mapper is built in Milestone kernel
[03:25:33] <Ave> hmm ok
[03:25:41] <Ave> there is no /dev/mapper
[03:26:21] <Ave> /dev/block/vold rather?
[03:26:56] <Ave> either way, the luks script/cryptsetup fails due to hardcoded expectations
[03:29:34] <Ave> compcache coming to milestone?
[03:34:02] <nadlabak> Ave: that's not clear at all at this point as it depends on miracleous swap module that does not exist yet
[03:34:23] <nadlabak> gotta go to sleeep, good night
[04:16:59] -!- Orgg [Orgg!orgg@terranigma.snes.org] has joined #milestone-modding
[04:43:36] -!- mavhc has quit [Ping timeout: 240 seconds]
[04:45:31] -!- rogerduran has quit [Ping timeout: 253 seconds]
[04:48:22] -!- mavhc [mavhc!~mavhc@cpc3-basf10-2-0-cust617.12-3.cable.virginmedia.com] has joined #milestone-modding
[05:05:35] -!- k1l has quit [Read error: Operation timed out]
[05:10:02] -!- k1l [k1l!~k1l@ubuntu/member/k1l] has joined #milestone-modding
[05:35:15] -!- cubi_ [cubi_!~mcd@p4FFE314F.dip.t-dialin.net] has joined #milestone-modding
[05:37:13] -!- cubi has quit [Ping timeout: 245 seconds]
[05:37:14] cubi_ is now known as cubi
[08:01:33] -!- TPoint75_ [TPoint75_!~tkrause@PCX142.ipht-jena.de] has joined #milestone-modding
[08:09:07] -!- typn has quit [Ping timeout: 240 seconds]
[08:09:15] -!- typn [typn!~quassel@adsl-99-188-112-54.dsl.ksc2mo.sbcglobal.net] has joined #milestone-modding
[08:47:47] -!- pontomedon [pontomedon!~poseidon@chello062178135187.2.14.univie.teleweb.at] has joined #milestone-modding
[08:47:48] -!- mode/#milestone-modding [+o pontomedon] by ChanServ
[09:41:25] -!- jonathantaylor [jonathantaylor!~jonathant@] has joined #milestone-modding
[09:45:40] -!- psychuil has quit [Ping timeout: 272 seconds]
[09:46:57] -!- broucaries [broucaries!~bastien@phy1.landsofr00t.eu] has joined #milestone-modding
[09:50:44] -!- psychuil [psychuil!~psychuil@bzq-79-177-141-160.red.bezeqint.net] has joined #milestone-modding
[09:54:18] -!- jonathantaylor has quit [Quit: jonathantaylor]
[10:02:09] -!- typn has quit [Ping timeout: 252 seconds]
[10:11:28] -!- typn [typn!~quassel@adsl-99-188-112-54.dsl.ksc2mo.sbcglobal.net] has joined #milestone-modding
[10:15:03] -!- RCFwork [RCFwork!~klaas@] has joined #milestone-modding
[10:25:52] Skrilax_CZ|Away is now known as Skrilax_CZ
[10:38:35] -!- endstille [endstille!~endstille@] has joined #milestone-modding
[10:59:30] -!- endstille_ [endstille_!~quassel@dsdf-5d8288d4.pool.mediaWays.net] has joined #milestone-modding
[11:07:43] -!- nadlabak_ [nadlabak_!~nadlabak@ip-62-245-82-172.net.upcbroadband.cz] has joined #milestone-modding
[11:08:27] -!- nadlabak has quit [Ping timeout: 245 seconds]
[11:08:41] -!- nadlabak_ has quit [Client Quit]
[11:09:50] -!- nadlabak [nadlabak!~nadlabak@ip-62-245-82-172.net.upcbroadband.cz] has joined #milestone-modding
[11:09:50] -!- mode/#milestone-modding [+o nadlabak] by ChanServ
[11:10:58] -!- jonathantaylor [jonathantaylor!~jonathant@] has joined #milestone-modding
[11:26:30] <endstille> hehe, finally, i should be able to compile faster :D http://dl.dropbox.com/u/12896017/Pics/compilier.png
[11:28:56] <k1l> uhh
[11:29:50] <k1l> is that a server? or a new desktop pc?
[11:31:11] <endstille> my notebook =) a had aVM on that one, but yesterday i wanted to switch to full power, so i created another dualboot :D
[11:31:26] <endstille> the ACER i postet one or 2 days ago
[11:31:32] <k1l> which notebook is that?
[11:32:20] <endstille> http://www.cyberport.de/notebook/notebooks/1C26-8FP/acer-aspire-7750g-2674g50mnkk-core-i7-2670qm-16gb-hd6850m.html
[11:32:27] <Ave> well shit, I cant figure out the device mapper in cm7
[11:32:42] <Ave> is /dev/device-mapper the control interface?
[11:32:50] <k1l> ahh i see. but that is more like a flat pc with monitor attached :)
[11:32:57] <endstille> XD
[11:33:27] <endstille> it's quite flat (without battery)
[11:33:41] <endstille> if you want much power, i can recommend it
[11:35:47] -!- mmodder [mmodder!~mmodder@] has joined #milestone-modding
[11:36:27] <mmodder> If you got swap to work, ICS would be so much more usable!
[11:36:36] -!- mmodder has quit [Client Quit]
[11:39:10] milaq|afk is now known as milaq
[11:54:15] <Skrilax_CZ> gonna be pretty brutal patch in any case (page_index and page_mapping)
[11:57:17] <Skrilax_CZ> xvilka: do you think that looking up bytes for BL #function and all occurences of it's address; and patching them to branch / point to function would be enough (assuming there will be no pitfalls like BL #function in raw data)
[12:08:40] -!- n0life [n0life!~madhatter@dslb-088-067-043-027.pools.arcor-ip.net] has joined #milestone-modding
[12:14:39] -!- stlifey [stlifey!~stlifey@] has joined #milestone-modding
[12:24:56] -!- Martus [Martus!~chatzilla@] has joined #milestone-modding
[12:42:37] -!- Martus has quit [Quit: ChatZilla 0.9.88 [Firefox 9.0.1/20111220165912]]
[13:17:54] -!- rogerduran [rogerduran!~roger@] has joined #milestone-modding
[13:40:29] milaq is now known as milaq|afk
[13:50:27] -!- stlifey has quit [Quit: WeeChat 0.3.7-dev]
[14:16:12] -!- Bavilo [Bavilo!~Bavilo@ip-81-210-252-115.unitymediagroup.de] has joined #milestone-modding
[14:52:06] <endstille> damn, sync finished, i wonder how long it takes now, to build cm7 =)
[14:56:59] <TheSpiritof69> It's done when it's done. :P
[15:06:04] <endstille> ^^
[15:07:06] <k1l> but should be pretty fast on the quadcore with 16gb ram
[15:21:46] <endstille> done 29 minutes
[15:26:54] milaq|afk is now known as milaq
[15:29:43] <xvilka> wow. Traffic of our server increased from 50Gb in a day in Novermber 2011 to 250Gb in a day in January 2011
[15:29:51] <xvilka> *2012 of course
[15:29:52] <endstille> hmm, i try to build most recent ICS and it tells me, i run an unsupported version of java..... GB worked like a charm
[15:30:40] <endstille> oh, its also working, maybe i just did not see the messeage on gb
[15:30:59] <endstille> xvilka: that is much more data....
[15:30:59] <k1l> xvilka: what happend?
[15:31:39] <Ave> is there ANY hope for seeing usable accelerated ics for mile?
[15:31:54] <Ave> considering the binary nature of certain drivers
[15:31:59] <xvilka> k1l: i think because of large sbf database
[15:32:24] * xvilka going to buy new or additional network cards/channels
[15:32:28] <k1l> xvilka: well yes.
[15:33:27] <TPoint75_> http://www.androidpit.de/de/android/blog/402144/apple-verliert-verkaufsverbot-iphone-ipad
[15:33:43] <xvilka> also, are there any experts on tuning web-servers?
[15:35:00] <pontomedon> lol
[15:35:20] <endstille> TPoint75_: yes i've found this as well :D
[15:35:29] <endstille> xvilka: i'm no expert, sorry ;)
[15:40:19] -!- TPoint75_ [TPoint75_!~tkrause@PCX142.ipht-jena.de] has parted #milestone-modding
[15:53:45] -!- Bavilo has quit [Quit: Leaving.]
[16:03:12] milaq is now known as milaq|afk
[16:12:13] <xvilka> also, looks like our traffic going to grow more and more
[16:12:25] <xvilka> Skrilax_CZ: are you here?
[16:12:58] milaq|afk is now known as milaq
[16:15:16] -!- n0life has quit [Quit: KTHXBYE]
[16:24:47] <Skrilax_CZ> xvilka: yes
[16:27:50] <endstille> hmmmm 59 minutes for ics.. sorry i'll stop spamming now :D
[16:28:34] <zewelor> xvilka: maybe use something like varnish to lower the load ?
[16:52:27] <xvilka> zewelor: varnish good for multiserver sites
[16:52:35] <xvilka> i guess it doesnt help much
[16:54:57] -!- Bavilo [Bavilo!~Bavilo@ip-81-210-252-115.unitymediagroup.de] has joined #milestone-modding
[16:57:11] <xvilka> zewelor: just received letter, i'm going to setup Gbit internet channel
[17:00:15] -!- n0life [n0life!~madhatter@HSI-KBW-109-193-047-224.hsi7.kabel-badenwuerttemberg.de] has joined #milestone-modding
[17:03:03] -!- RCFwork has quit [Quit: Leaving.]
[17:17:17] -!- endstille_ has quit [Remote host closed the connection]
[17:30:14] -!- broucaries has quit [Quit: Leaving.]
[17:31:53] -!- mifritscher [mifritscher!~michi@static.] has joined #milestone-modding
[17:37:27] -!- emper0r [emper0r!~administr@] has joined #milestone-modding
[17:37:38] <emper0r> hi
[17:42:18] -!- broucaries [broucaries!~bastien@phy1.landsofr00t.eu] has joined #milestone-modding
[17:46:15] -!- endstille has quit [Quit: I'll be back.]
[17:51:59] -!- broucaries has quit [Quit: Leaving.]
[17:53:15] <xvilka> yay! We have 1Gbit internet connection!
[17:55:07] -!- k1l has quit [Ping timeout: 240 seconds]
[17:56:39] <Skrilax_CZ> good
[17:57:11] -!- k1l [k1l!~k1l@ubuntu/member/k1l] has joined #milestone-modding
[18:05:30] -!- raubvogel [raubvogel!~raubvogel@shop.monetra.com] has parted #milestone-modding
[18:06:36] -!- raubvogel [raubvogel!~raubvogel@shop.monetra.com] has joined #milestone-modding
[18:31:01] -!- DAGr8 has quit [Remote host closed the connection]
[18:31:19] milaq is now known as milaq|afk
[18:33:27] <mioze7ae> Skrilax_CZ: In 2.6.29 page_index doesn't depend on CONFIG_SWAP. Is that true in 2.6.32?
[18:33:46] <Skrilax_CZ> that's indirect dependency
[18:34:31] Red|Off is now known as Redh3ad
[18:44:31] <mioze7ae> Oh, I see where that goes now. Yikes.
[18:45:26] <mioze7ae> Maybe ksplice would work?
[18:46:10] <mioze7ae> I started reading the docs, but I haven't built anything with it yet. I think swap is one of those ones that would need helper code.
[18:47:06] <mioze7ae> It was a little hard to find the downloads now that Oracle's got their greedy finger in it.
[18:47:31] <mioze7ae> So, I'm still working on my setup for that
[19:02:24] -!- mifritscher has quit [Read error: Operation timed out]
[19:08:15] milaq|afk is now known as milaq
[19:13:43] -!- Luxxi has quit [Remote host closed the connection]
[19:19:20] -!- Luxxi [Luxxi!~Luxx@] has joined #milestone-modding
[19:19:20] -!- Luxxi has quit [Changing host]
[19:19:21] -!- Luxxi [Luxxi!~Luxx@unaffiliated/gumbler] has joined #milestone-modding
[19:29:22] -!- jonathantaylor has quit [Ping timeout: 255 seconds]
[19:37:11] <Skrilax_CZ> mioze7ae: ksplice - well just partially, I'm intending to use it to find differences in ELF object to easen the work
[19:37:44] <Skrilax_CZ> but I'm also intending to patch the kernel by rewriting the branches / addresses in it
[19:41:13] <mioze7ae> I think just about all of it is isolated to mm/
[19:41:28] <mioze7ae> There are a few things like ./arch/arm/mm/init.c
[19:41:42] <Skrilax_CZ> so I'm making a list of all "symbols" (not symbols exactly, just need addresses) and will scan for them (on PC) and generate a small header file containing patchdata, where the module at runtime (interrupts disabled at that time) will patch it with new addresses (BL instruction)
[19:41:45] <mioze7ae> and ./fs/proc/meminfo.c
[19:41:49] <Skrilax_CZ> also, no
[19:41:52] <Skrilax_CZ> API - yes
[19:41:54] <Skrilax_CZ> ABI - no
[19:41:57] <Skrilax_CZ> => inlined
[19:42:38] <mioze7ae> Right, I've been indexing the code to find where all the inlined things are used
[19:42:52] <mioze7ae> Well, the explict inlining anyway.
[19:43:14] -!- stlifey [stlifey!~stlifey@] has joined #milestone-modding
[19:44:13] <mioze7ae> The SwapCache flag isn't used outside mm/ much at all and mostly inside the objects that aren't present without CONFIG_SWAP
[19:45:52] <Skrilax_CZ> yeah, but if you start looking on the inline functions, their use is fairly common
[19:47:22] <Skrilax_CZ> and also of course not all is used by the config
[19:47:47] <Skrilax_CZ> so I thing it's perhaps easier to compare the ELF objects with swap disabled and swap enabled
[19:50:03] <Skrilax_CZ> s/thing/think/
[19:53:35] <Ave> crw------- 1 root root 10, 43 Feb 2 00:01 /dev/device-mapper
[19:53:42] <Ave> what is this minor 43?
[19:54:02] <Ave> the function of this char device
[19:55:19] <mioze7ae> Right, but then why not just let ksplice update the majority/all of the functions?
[19:55:45] <mioze7ae> What I've been hunting for is a struct that changed, and I haven't seen that yet
[19:56:21] <mioze7ae> If structs don't change it should just be a matter of updating the code and then walking the state. You have to walk the state either way.
[19:56:35] <mioze7ae> But ksplice would let you skip the updating
[19:57:37] <mioze7ae> And we can modify the kernel source to get ksplice partway if needed.
[19:57:54] <mioze7ae> and then do the rest by hand
[20:03:36] <mioze7ae> i.e. instead of full swap vs no swap, do no swap vs no swap with the inlines "fixed"
[20:14:42] -!- stlifey has quit [Quit: WeeChat 0.3.7-dev]
[20:17:51] -!- mifritscher [mifritscher!~michi@188-194-220-209-dynip.superkabel.de] has joined #milestone-modding
[20:18:15] <Skrilax_CZ> mioze7ae: it's not just patching those function
[20:18:18] <Skrilax_CZ> *functions
[20:18:24] -!- mifritscher has quit [Client Quit]
[20:18:35] -!- mifritscher [mifritscher!~michi@188-194-220-209-dynip.superkabel.de] has joined #milestone-modding
[20:26:54] -!- endstille [endstille!~endstille@dslb-092-072-186-131.pools.arcor-ip.net] has joined #milestone-modding
[20:28:13] -!- stlifey [stlifey!~stlifey@] has joined #milestone-modding
[20:32:06] <mioze7ae> Right. But patching those functions is a big part of it that ksplice could do easily.
[20:34:25] -!- rah has quit [Ping timeout: 252 seconds]
[20:35:45] -!- mifritscher1 [mifritscher1!~michi@188-194-220-209-dynip.superkabel.de] has joined #milestone-modding
[20:35:45] -!- mifritscher has quit [Disconnected by services]
[20:36:08] -!- mifritscher [mifritscher!~michi@188-194-220-209-dynip.superkabel.de] has joined #milestone-modding
[20:45:33] milaq is now known as milaq|afk
[20:48:33] -!- endstille has quit [Quit: I'll be back.]
[21:00:08] -!- DAGr8 [DAGr8!~v@modemcable049.8-202-24.mc.videotron.ca] has joined #milestone-modding
[21:21:20] -!- Ave [Ave!ave@kapsi.fi] has parted #milestone-modding
[21:26:02] -!- rah [rah!rah@myrtle.6gnip.net] has joined #milestone-modding
[21:30:47] -!- stlifey has quit [Quit: WeeChat 0.3.7-dev]
[21:39:51] -!- emper0r has quit [Quit: leaving]
[21:50:15] -!- eiyee [eiyee!~eiyee@ip-95-223-12-17.unitymediagroup.de] has joined #milestone-modding
[21:59:46] -!- mifritscher1 has quit [Quit: Leaving.]
[22:13:51] -!- fernando_ has quit [Quit: Leaving...]
[22:51:31] <Skrilax_CZ> mioze7ae: I am not entirely convinced of that. ksplice was designed for applying security patches without rebooting the system
[22:54:48] <mioze7ae> Does exactly what the hook function does.
[22:55:09] <Skrilax_CZ> and we're not doing that
[22:55:30] <mioze7ae> But reading more about ksplice there seems to be other challenges, like you can't hotpatch functions that are active.
[22:55:45] <mioze7ae> or suspended
[22:56:11] <mioze7ae> So ksplice will try repeatedly until the functions are unused
[22:56:29] <mioze7ae> It seems likely that the mm is going to be active a lot
[22:57:17] <mioze7ae> Oh, you're not doing something like the mmcfix?
[22:57:45] <Skrilax_CZ> no, I'll analyze the kernel statically for the calls and patch them
[22:58:19] <Skrilax_CZ> not put a pseudo instruction on the function (+ symbols, lots of functions are static)
[22:58:40] <Skrilax_CZ> bleh ... pseudo instruction = jump to new function
[22:58:42] <Luxxi> wait
[22:59:09] <Luxxi> kabaldan has an mm which doesnt have a locked bootloader?
[22:59:10] <Luxxi> :o
[22:59:17] <Skrilax_CZ> he has dev mm, yes
[22:59:25] <Luxxi> :o nice
[23:00:06] <mioze7ae> But most of the changes are code deletions, there's no space to stuff new code
[23:00:17] <mioze7ae> esp the inline things
[23:01:32] <mioze7ae> I mean I assume when things like SwapCache are redefined to noops, the optimizer will throw them out
[23:04:19] <mioze7ae> Oh, I see what you're saying. NM
[23:09:30] -!- dx has quit [Ping timeout: 260 seconds]
[23:15:35] -!- dx [dx!~dicks@host249.190-139-222.telecom.net.ar] has joined #milestone-modding
[23:23:45] -!- endstille [endstille!~endstille@dslb-092-072-186-131.pools.arcor-ip.net] has joined #milestone-modding
[23:27:19] <Skrilax_CZ> mioze7ae: also nadlabak has one (optional) patch to modify a structure in blkdev (forgot the exact name)
[23:30:25] <mioze7ae> The ones I know are dsifix and mmcfix
[23:34:07] <mioze7ae> I've been meaning to look at those, but I didn't realize mmcfix was altering structs.
[23:42:10] <k1l> just made a cm9 build this morning. will try it now :)
[23:43:57] <mioze7ae> Skrilax_CZ: I've sort of ignore mmc-fix because the code that's being patched seems so different from XT720|2.6.29
[23:44:52] <mioze7ae> our set_data_timeout() takes two args instead of three for example.
[23:45:25] <mioze7ae> and the struct has a different name/is all sorts of different
[23:45:26] -!- Bavilo has quit [Read error: Connection reset by peer]
[23:45:33] <Skrilax_CZ> mioze7ae: mmc-fix doesn't alter any structures
[23:46:08] <mioze7ae> No, I meant 2.6.29 to 2.6.32 did.
[23:46:26] <Skrilax_CZ> no, my point was smth different
[23:46:27] <Skrilax_CZ> https://github.com/a853/a853_compcache/blob/master/patches/patch_swap_notify_core_support_2.6.32.diff
[23:46:48] <Skrilax_CZ> struct block_device_operations
[23:47:22] <Skrilax_CZ> is given an extra one
[23:47:52] <mioze7ae> Right, that patch comes from the compcache project.
[23:48:31] <mioze7ae> It's recommended but optional
[23:48:59] <Skrilax_CZ> well I'll hack it via struct module
[23:49:01] <Skrilax_CZ> :P
[23:49:44] <mioze7ae> I'll be interested to see it. I learn a lot from your code :D
[23:50:20] -!- raubvogel [raubvogel!~raubvogel@shop.monetra.com] has parted #milestone-modding
[23:50:31] <Skrilax_CZ> I'll declare struct like: struct blkdev_patch { struct blkdev; void (*swap_slot_free_notify) (struct block_device *, unsigned long); }
[23:50:44] <Skrilax_CZ> and use it for compcache
[23:51:44] <Skrilax_CZ> then I'll check if the "struct module *owner" is ME (let's say by comparing init function address)
[23:52:33] <Skrilax_CZ> and if it is, use the blkdev in (disk->fops) in the swap_entry_free function as blkdev_patch
[23:57:20] <Skrilax_CZ> sed 's/blkdev/block_device_operations/'