--- Log for 13.11.114 Server: tepper.freenode.net Channel: #rockbox --- Nick: logbot- Version: Dancer V4.16 Started: 15 hours and 27 minutes ago 00.04.49 Quit xorly (Ping timeout: 255 seconds) 00.06.56 Quit lebellium (Quit: ChatZilla 0.9.91 [Firefox 34.0/20141106201515]) 00.11.05 Quit bertrik (Remote host closed the connection) 00.16.36 Quit petur (Remote host closed the connection) 00.25.16 Quit edhelas (Ping timeout: 245 seconds) 00.38.12 *** Saving seen data "./dancer.seen" 00.41.10 # [Franklin] It was just an idea I was throwing out there... 00.43.19 # Yeah it would be cool to integrate it all into the build system and have the option in rbutil to pick and choose 00.44.47 Quit ender` (Quit: Because nothing says "I have faith in God" like 3 inches of bulletproof glass on popemobile.) 00.56.55 Join JdGordon [0] (~jonno@ppp118-209-184-248.lns20.mel8.internode.on.net) 00.58.54 Quit JdGordon_ (Ping timeout: 272 seconds) 01.01.58 Join JdGordon_ [0] (~jonno@ppp118-209-232-6.lns20.mel8.internode.on.net) 01.02.04 Quit JdGordon (Ping timeout: 272 seconds) 01.26.48 Quit foolsh (Remote host closed the connection) 01.30.29 Join foolsh [0] (~foolsh@c-24-11-243-148.hsd1.in.comcast.net) 02.05.30 # <[Franklin]> foolsh: the thing that contributes to .rock size is bitmaps, not code 02.08.26 # <[Franklin]> which is why clock is an order of magnitude bigger than xworld 02.10.16 Join JdGordon [0] (~jonno@ppp118-209-140-221.lns20.mel8.internode.on.net) 02.10.41 Quit JdGordon_ (Ping timeout: 250 seconds) 02.20.22 # <[Franklin]> foolsh: so how were you going to classify "extra" vs non-"extra" plugins? 02.23.58 # Oh, its not bin size that concerns, it was that once a plugin goes in the rockbox tree, its almost as if it become part of the core, needing rockbox porters to wade through useless demo after demo, plugin after plugin, if the plugins grow "too" big, I feel someone some where will wish people to stop adding games and stuff to the tree 02.25.33 # if it moves to a more modular work flow, it means less work porting to new hardware, and more freedom to add any amount of plugins 02.26.21 # <[Franklin]> foolsh: aha 02.26.25 # <[Franklin]> now I get it 02.26.34 # <[Franklin]> yeah, that would be a problem 02.26.43 # <[Franklin]> but then again, its nice for every rockbox install to have plugins 02.26.53 # <[Franklin]> no need to worry about needing to install "extras" each time 02.27.34 # when is the last time you needed bounce? 02.27.39 # <[Franklin]> foolsh: lol 02.27.41 # <[Franklin]> never 02.27.57 # <[Franklin]> yeah, I agree that at least some of the demos are *useless* 02.28.04 # <[Franklin]> I thought you were referring to the games 02.28.19 # <[Franklin]> foolsh: then the real problem is that plugins are not portable 02.28.42 # <[Franklin]> it should be that all that nothing needs to be changed from platform to platform 02.28.51 # <[Franklin]> not even keymaps 02.28.57 # <[Franklin]> but unfortunately, that's not possible 02.29.25 # It has many cavets 02.32.15 # yes extra-plugins would have to be maintained by the community, but as it is now themes are exactly that 02.32.26 # just a thought 02.32.35 # I'm never serious ;) 02.32.42 # <[Franklin]> foolsh: neither am I 02.33.18 # <[Franklin]> foolsh: I like it the way it is now 02.34.33 Quit ZincAlloy2 (Quit: Leaving.) 02.34.42 # It would be nice to be able to just dump a source directory into plugins and get a compiled rock at ciimpile time with no other changes needed to rockbox 02.36.26 # set a few flags in a SOURCES file and tell the preprocessor to scan the whole directory 02.37.05 # <[Franklin]> foolsh: I still think that's too complicated for the average Joe 02.37.34 # <[Franklin]> foolsh: I think that the way it is now is just fine 02.37.47 # <[Franklin]> because any changess simply complicate things 02.38.13 *** Saving seen data "./dancer.seen" 02.38.23 # <[Franklin]> so you're saying that there'd be two links on the download page for each target? 02.38.28 # <[Franklin]> one "basic" and one "full" 02.38.34 # <[Franklin]> or something like that? 02.39.20 Quit bluebrother (Disconnected by services) 02.39.25 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother) 02.39.26 # more like rockbox.zip and plugin-extras.zip I was thinking 02.39.40 # unzip extras over .rockbox 02.39.53 # but it's not an issue 02.39.57 # yet 02.40.11 # * [Franklin] still finds it too complicated 02.40.16 # <[Franklin]> bins are complicated 02.40.33 # <[Franklin]> they're tightly coupled with core 02.40.46 # <[Franklin]> so there'd be the problem of mismatches 02.41.02 # <[Franklin]> (not so much with release builds, but there hasn't been a release in *forever*) 02.41.20 # <[Franklin]> speaking of which, I really think a release is *long overdue* 02.41.35 Quit fs-bluebot (Ping timeout: 265 seconds) 02.43.56 # yes I understand that memory addresses change with each compile and as gevaerts and wodz talked about this morning they're version-ed extremely strictly because of that, I understand that it would "have" to be tightly coupled with each build somehow 02.44.37 # * [Franklin] goes and reads the logs 02.44.44 # but I'm also not saying it has to be, it just an idea 02.45.25 # <[Franklin]> yeah 02.46.41 # <[Franklin]> lol 06:56foolshhe will never stop 02.47.18 # lol, I haven't seen you slow an instant in two weeks 02.49.35 # <[Franklin]> next thing to do: wolf3d 02.49.56 Join fs-bluebot [0] (~fs-bluebo@g224237078.adsl.alicedsl.de) 02.51.21 # <[Franklin]> foolsh: well, I don't care about sound for xworld :) 02.51.28 # <[Franklin]> I guess I'll leave that for someone else 02.51.44 # no one does I think ;) 02.51.50 # <[Franklin]> ok then 02.51.56 # <[Franklin]> that simplifies things *a lot* 02.52.24 # <[Franklin]> then all that's really left to do is to get the thing to run on gray+mono devices 02.52.34 # <[Franklin]> and optimize it like crazy 02.53.49 # <[Franklin]> and document it 02.56.18 # * foolsh starts downloading 500+ mb of LaTeX 02.56.34 # * [Franklin] never cared much about manual entries 02.56.37 # <[Franklin]> at most 2 sentences 02.56.41 # <[Franklin]> no pictures 02.57.21 # fine with me 02.57.35 # * [Franklin] hopes no one reads the manual 02.57.41 # <[Franklin]> or uses the help screen 03.00.00 Quit AlexP (Remote host closed the connection) 03.01.13 # <[Franklin]> also, I need to put yield()s in 03.01.42 # <[Franklin]> ok lets see here... 03.01.56 # <[Franklin]> "XWorld is an interpreter for Another World, a fantastic game by Eric Chahi" 03.02.03 # [Franklin] I wouldn't bother with targets that can't do "good" bmp scalling, it makes it imposible to get past the code imput screen with out it 03.02.11 # <[Franklin]> for the help screen, how's that? 03.02.26 # that sound good 03.02.26 # <[Franklin]> foolsh: so that'd be everything but color then 03.02.30 # right 03.02.47 # or zoom 03.02.52 # but htats more work 03.02.55 # <[Franklin]> zoom? 03.03.11 # sure zoom and pan 03.03.16 # <[Franklin]> how? 03.03.32 # the problem is you can't read the code say "O23" or what ever 03.03.52 # on tiny screens, the "good" scalling solves that 03.03.54 # * [Franklin] hasn't tried yet 03.04.04 # Ah I've seen it on my e280 03.04.08 # <[Franklin]> let's try this distcc 03.09.19 # <[Franklin]> ooh yes I see 03.09.26 # <[Franklin]> it is *unreadable* on e200 without good scaling 03.09.55 # yes its completely useless with out good scalling 03.10.08 # so oh well so long 03.10.13 # <[Franklin]> unless the code screen can be bypassed 03.21.04 # [Franklin], fine with me if you skip the gray scales then all together either way it's more your baby than mine ;) 03.41.43 Quit [Franklin] (Remote host closed the connection) 03.46.13 Join Strife89 [0] (~Strife89@adsl-98-80-212-67.mcn.bellsouth.net) 03.51.24 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs) 04.15.05 Quit stripwax (Read error: Connection reset by peer) 04.22.22 Quit cmhobbs (Ping timeout: 245 seconds) 04.38.17 *** Saving seen data "./dancer.seen" 04.40.10 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma) 04.40.11 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn) 04.40.11 Quit pixelma (Disconnected by services) 04.40.11 Quit amiconn (Disconnected by services) 05.04.54 Quit Scr0mple (Read error: Connection reset by peer) 05.05.12 Join Scr0mple [0] (~Simon@27.127.199.230) 05.12.26 Quit ender| (Ping timeout: 272 seconds) 05.12.26 Quit Riviera (Ping timeout: 272 seconds) 05.12.27 Quit yuriks (Ping timeout: 272 seconds) 05.12.28 Quit mpmc (Ping timeout: 272 seconds) 05.12.33 Join yuriks_ [0] (~quassel@opentyrian/developer/yuriks) 05.12.53 Quit the-kyle (Ping timeout: 272 seconds) 05.13.25 Join mpmc [0] (~mpmc@unaffiliated/mpmc) 05.13.52 Join the-kyle [0] (~kyle@kyle.tk) 05.21.54 Join ender| [0] (krneki@2a01:260:4094:1:42:42:42:42) 05.51.08 Quit [7] (Disconnected by services) 05.51.20 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven) 06.31.57 Quit Strife89 (Ping timeout: 245 seconds) 06.38.21 *** Saving seen data "./dancer.seen" 06.41.16 Quit prof_wolfff (Ping timeout: 244 seconds) 07.01.24 # <[Saint]> TheSeven: Do you have any idea whay only _some_ people would be having issues with http://git.rockbox.org/?p=rockbox.git;a=commit;h=582035c ? 07.01.41 # <[Saint]> I've attempted to look into it, but it confuses the fuck out of me. 07.02.03 # <[Saint]> The main confusing thing, is I can't for the life of me see where that arror is actually coming from. 07.02.09 # <[Saint]> But its a repeatable address. 07.02.21 # <[Saint]> (I'm not experiencing it, but many are) 07.17.45 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn) 07.21.26 Quit shamus (Read error: Connection reset by peer) 07.26.27 Join shamus [0] (~shmaus@ip-206-192-193-180.marylandheights.ip.cablemo.net) 08.10.24 Join pamaury [0] (~quassel@rockbox/developer/pamaury) 08.17.50 Join ender` [0] (krneki@foo.eternallybored.org) 08.28.03 Join mortalis [0] (~kvirc@212.44.150.238) 08.37.06 Join xorly [0] (~xorly@m180.dkm.cz) 08.37.51 Join lebellium [0] (~chatzilla@i16-les01-ntr-212-194-176-149.sfr.lns.abo.bbox.fr) 08.38.22 *** Saving seen data "./dancer.seen" 08.47.12 Quit pamaury (Ping timeout: 256 seconds) 08.57.49 Join petur [0] (5bb7304d@gateway/web/freenode/ip.91.183.48.77) 08.58.13 Nick petur is now known as Guest85674 (5bb7304d@gateway/web/freenode/ip.91.183.48.77) 08.58.43 Join prof_wolfff [0] (~prof_wolf@82.158.1.206.dyn.user.ono.com) 08.59.34 Nick Guest85674 is now known as p3tur (5bb7304d@gateway/web/freenode/ip.91.183.48.77) 08.59.45 Quit p3tur (Changing host) 08.59.45 Join p3tur [0] (5bb7304d@rockbox/developer/petur) 09.12.19 Quit xorly (Ping timeout: 245 seconds) 09.42.08 Join pamaury [0] (~quassel@sphinx.lix.polytechnique.fr) 09.42.09 Quit pamaury (Changing host) 09.42.09 Join pamaury [0] (~quassel@rockbox/developer/pamaury) 09.46.58 Join einhirn [0] (~Miranda@bsod.rz.tu-clausthal.de) 10.01.20 Nick p3tur is now known as petur (5bb7304d@rockbox/developer/petur) 10.03.03 Join edhelas [0] (~edhelas@193.172.124.224) 10.05.00 Join maitreya [0] (4f2a005d@gateway/web/freenode/ip.79.42.0.93) 10.05.43 Quit maitreya (Client Quit) 10.19.32 Quit edhelas (Quit: Quitte) 10.19.58 Join edhelas [0] (~edhelas@193.172.124.224) 10.32.25 Quit mc2739 (Ping timeout: 250 seconds) 10.38.24 *** Saving seen data "./dancer.seen" 11.04.12 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma) 11.07.46 Quit foolsh (Quit: Leaving) 11.08.53 Quit akaWolf (Ping timeout: 250 seconds) 11.21.37 Join maruk1 [0] (~papier@titanium.v6.sdv.fr) 11.29.46 Quit mpmc (Quit: ZNC - http://znc.in) 11.30.37 Join mpmc [0] (~mpmc@unaffiliated/mpmc) 11.46.54 Quit lebellium (Quit: ChatZilla 0.9.91 [Firefox 34.0/20141110195804]) 12.14.25 Join krabador_ [0] (~krabador@host178-180-dynamic.11-87-r.retail.telecomitalia.it) 12.15.28 Quit krabador_ (Read error: Connection reset by peer) 12.23.05 Join wodz [0] (~wodz@iwl138.internetdsl.tpnet.pl) 12.23.09 # pamaury: ping 12.28.13 Join Riviera [0] (Riviera@2a03:b0c0:1:d0::10:b001) 12.38.25 *** Saving seen data "./dancer.seen" 12.57.16 Join krabador [0] (~krabador@unaffiliated/krabador) 13.12.20 # wodz: pong 13.13.35 # pamaury: Do you have experience with llvm .td files? I would like to add synci instruction to Mips but I fail miserably. 13.14.48 # At some point I got a bit familiar with it but then they changed many things so I'm not so sure anymore 13.15.02 # however the llvm irc channel and mailing list are quite helpful 13.16.00 # iirc you basically need to do two things: add the instruction itself (virtual node, how to encode/decode it), and add some rules to lower some construction to it 13.16.18 # (except if you plan to use it assembly directly) 13.16.48 # ? 13.18.15 # Instruction definition is in *InstrFormat.td and InstrInfo.td and this other bits? 13.18.25 # yes I think so 13.19.02 # in lib/Target/Mips/Mips32r6InstrFormats.td you probably need to add the opcode and in lib/Target/Mips/Mips32r6InstrInfo.td the intruction itself 13.20.13 # you must specify the OutOperandList and InOperandList, the AsmString, and probably a few others things if it acts a barrier or load/store some memory 13.22.51 # this instruction basically do writeback of dcache followed by icache invalidate 13.23.38 # ok, let me read through the description files, I'll try to give you a basis to start on 13.25.13 # in InstrFormat.td, you need "def OPCODE5_SYNCI : OPCODE5<0b11111>;" (I hope I got the opcode right) 13.26.38 # in InstrIntr.td, first declare the encoding: "class SYNCI_R6_ENC : REGIMM_FM;" (I guess the R6 should be replaced by the correct revision it was introduced in, doesn't really matter anyway) 13.27.07 # define the instruction: def SYNCI_R6 : SYNCI_R6_ENC, SYNCI_R6_DESC, ISA_MIPS32R6; 13.27.26 # define the description (SYNCI_R6_DESC), that's probably the hardest part 13.33.10 # wodz: https://gist.github.com/pamaury/7386b1765657f7ae7bac 13.33.14 # work in progress ;) 13.33.21 # * wodz looking 13.33.47 # I think it needs to specify that it acts as a memory barrier 13.34.03 # or pretend it touches memory 13.34.53 # I have updated it with "bit isBarrier = 1;" 13.35.23 # you also need to specify the delay slot if it has one 13.35.33 # no it does not 13.36.42 # hmm where is defined mapping of the fields in actual 32bit word? 13.37.06 # REGIMM_FM? 13.37.35 # yes 13.38.21 # but this only creates the instruction, you cannot use it for the moment I think 13.38.31 # and you probably want to have some llvm guy check that 13.38.59 # I think you need to hook something in the MC assembly parser if you want to use it in inline assembly which I guess you want 13.39.27 # well I actually want to use it in disassembly (llvm-mc) 13.40.16 # oh ok, I think this won't suffice, you need something more 13.41.12 # hell, MipsInstrInfo.td and Mips32r6InstrInfo.td are vastly different 13.42.28 # I think you need to specify a DecoderMethod, unless llvm is smart enough to decode it only using REGIMM_FM which I doubt 13.43.45 # hum I don't know 13.45.03 # does synci belongs to mips32r6 by the way ? 13.45.38 # no :-) this is mips32r2 instruction 13.46.21 # so rather MipsInstr.td then 13.47.28 # hum indeed the file is written in a completely different style 13.51.28 # it's weird that SYNC is declared with a i32 immediate by the way 13.54.47 # You mean that according to isa manual it doesn't take arguments? 13.56.04 # according to mips manual, sync a single immediate argument which is 5-bit 13.56.24 # ah yes, thats true 14.02.04 # I'm wondering what is the bare minimal to be able to use it in the disassembler only... 14.02.31 Join krnlyng [0] (~liar@83.175.90.24) 14.06.56 # wodz: maybe something like this: https://gist.github.com/pamaury/7386b1765657f7ae7bac 14.08.43 Quit krnlyng (Remote host closed the connection) 14.09.21 # * wodz tries 14.11.18 # I think ISA_MIPS32 is not right but we don't really care, however I'm not so sure the disassembler will pick the instruction just like this 14.12.37 Join krnlyng [0] (~liar@83.175.90.24) 14.12.50 # it should be ISA_MIPS32R2 but as you said this doesn't matter now 14.13.28 Join TheLemonMan [0] (~lemonboy@unaffiliated/thelemonman) 14.14.40 # pamaury: http://pastie.org/9716671 14.15.08 # urg, very helpful ^^ 14.19.51 # CCAction seems related to calling convention 14.21.55 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf) 14.22.34 # wodz: just for consistency, have you checked it compiles without the added code ? ;) 14.22.57 # pamaury: just now, and it doesn't :P 14.23.10 # yeah, because that didn't make any sense 14.38.29 *** Saving seen data "./dancer.seen" 14.47.53 # pamaury: http://pastie.org/9716745 14.56.29 # wodz: I'm compiling llvm on my box to try it, give me the afternoon so I can find some free time to try it and I'll tell you what I managed to achieve 14.56.52 # sure, thanks for help 15.01.08 Join mortalis|2 [0] (~kvirc@194.133.18.67) 15.01.28 Join mortalis|3 [0] (~kvirc@212.44.150.238) 15.02.19 Quit mortalis|3 (Client Quit) 15.04.36 Quit mortalis (Ping timeout: 258 seconds) 15.05.17 Quit mortalis|2 (Ping timeout: 245 seconds) 15.08.16 Quit wodz (Quit: Leaving) 15.11.34 Join ZincAlloy [0] (~Adium@pD9EEB747.dip0.t-ipconnect.de) 15.24.31 Join amayer [0] (~amayer@mail.weberadvertising.com) 15.30.46 Quit TheLemonMan (Quit: leaving) 16.17.51 Quit polemon__ (Ping timeout: 264 seconds) 16.25.15 Join Cunctator [0] (4f2a005d@gateway/web/freenode/ip.79.42.0.93) 16.25.56 # anyone here able to help? 16.32.53 Quit krabador (Ping timeout: 264 seconds) 16.34.39 # Cunctator: maybe, but we won't know until we know what your problem is :) 16.36.26 # awesome. I'm having issues with a rockboxed ipod 160gb 7g. Followed the freemyipod.org instructions carefully. It all works fine, except, I cannot add files on it. Recognized by my Win8.1 laptop, but as I copy&paste it gets stuck and needs rebooting. 16.38.32 *** Saving seen data "./dancer.seen" 16.40.40 # I should say: I managed to move my music library as soon as I installed it. Then played with it for a day. Then wanted to add new files, and all of a sudden anything bigger than a few MB will make the file transfer freeze. 16.41.14 # (and no, it's not full. I got 15GB of free space, says my computer) 16.42.49 # I suspect [Saint] is the person to help you. He seems to keep track of ipod classic issues fairly well 16.43.16 # He might wake up any minute now 16.46.40 # alright, I'll be AFK for a while now, but back later. Thanks for your help! 16.47.11 # You can always check the logs to see if someone answered 16.47.19 # Even if you're disconnected, I mean 16.49.00 # thanks :) 17.17.52 Quit edhelas (Ping timeout: 265 seconds) 17.21.48 Quit petur (Ping timeout: 246 seconds) 17.43.30 Join Guest66888 [0] (Slayer@c-69-143-187-144.hsd1.va.comcast.net) 17.45.29 Nick yuriks_ is now known as yuriks (~quassel@opentyrian/developer/yuriks) 17.45.54 Quit Guest66077 (Ping timeout: 244 seconds) 17.46.14 Join AlexP [0] (~alex@rockbox/staff/AlexP) 18.16.11 Join ikeboy [0] (~ikeboy@ool-435622d3.dyn.optonline.net) 18.27.07 Quit maruk1 (Quit: Leaving.) 18.38.34 *** Saving seen data "./dancer.seen" 18.43.04 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org) 18.46.50 # wodz (logs): I have something working but it's unsatisfactory in several points: 1) the disassembler outputs "synci imm, base" instead of "synci imm(base)" 2) also the immediate is always printed as a sign number, don't know why 18.53.21 # wodz (logs): but I think I know to fix this, I need to understand how it works for load/store, it uses custom parsing procedure and representation, let's reuse that 19.04.42 Join y4n [0] (~y4n@unaffiliated/y4ndexx) 19.10.27 Quit Makinit (Quit: bye) 19.11.39 # wodz (logs): current state: https://gist.github.com/pamaury/7386b1765657f7ae7bac 19.11.47 # if I uncomment //let DecoderMethod = "DecodeMem"; 19.11.55 Join Makinit [0] (makinit@makinit.nl) 19.12.25 # parsing is wrong because DecodeMem decodes an extra register (like for sw and ld), and without it crashes, so I think I need a custom decoder, I'll write that later, stay tuned ! 19.13.19 Quit pamaury (Remote host closed the connection) 19.16.53 Join edhelas [0] (~edhelas@77-173-104-232.ip.telfort.nl) 19.21.10 Join bertrik [0] (~quassel@rockbox/developer/bertrik) 19.53.17 Quit edhelas (Ping timeout: 264 seconds) 20.02.06 Join pamaury [0] (~quassel@rockbox/developer/pamaury) 20.05.53 Join krabador [0] (~krabador@unaffiliated/krabador) 20.09.48 Quit Cunctator (Ping timeout: 246 seconds) 20.09.51 Join ZincAlloy1 [0] (~Adium@pD9EE9574.dip0.t-ipconnect.de) 20.11.16 Quit ZincAlloy (Ping timeout: 255 seconds) 20.32.52 Join wodz [0] (~wodz@89-75-106-114.dynamic.chello.pl) 20.35.18 # pamaury: hmm, that is interesting. cache instruction also declares Reg operand which is actually not used (cache $hint, $addr where addr is imm16(base reg)). I tried to mimic how is cache modeled but llvm-mc crashes with my version. 20.36.23 # I had a look at the code and I think it will be quite easy to write a custom decoder, I'm busy right now but I will write it before the end of the day 20.37.05 # I am eager to understand how this works. 20.37.18 # pamaury: .td is enough to have disasm working? 20.38.06 # yes, except I need the decoder in this case but it will generated the disassembler all by itself 20.38.36 *** Saving seen data "./dancer.seen" 20.41.16 # pamaury: I don't quite understand why it is not possible to simply modify what CACHE does (drop the $hint) 20.44.12 Join Cunctator [0] (4f2a005d@gateway/web/freenode/ip.79.42.0.93) 20.45.48 # wodz: because since it uses this complicated address structure imm(rs) which is represented by a MipsMemAsmOperand, it has to use a custom decoder, DecodeCacheOp in the case of cache, and this decode handles all arguments ($hint and $addr) 20.47.30 # damn, llvm trunk compiles at my lab's computer but not on my laptop, apparently it uses new c++ features 20.47.33 # *sight* 20.48.34 Join edhelas [0] (~edhelas@77-173-104-232.ip.telfort.nl) 20.48.36 # pamaury: I came up with this http://pastie.org/9717458 which crashes actually 20.50.27 # no that won't work, mem already contains the immediate and a register 20.51.30 # wait a sec, I will upload some code, I cannot compile it at the moment but you can 20.53.44 # wodz: https://gist.github.com/pamaury/7386b1765657f7ae7bac 20.56.29 Join rela [0] (~x@pdpc/supporter/active/rela) 20.58.57 Quit Cunctator (Quit: Page closed) 21.12.20 Join TheLemonMan [0] (~lemonboy@unaffiliated/thelemonman) 21.19.55 Join Cunctator [0] (4f2a005d@gateway/web/freenode/ip.79.42.0.93) 21.28.11 Quit rela (Read error: Connection reset by peer) 21.52.58 Quit y4n (Quit: PANTS OFF!) 21.58.22 Quit yuriks (Remote host closed the connection) 22.02.11 Join petur [0] (~petur@rockbox/developer/petur) 22.02.31 Join Cunctator_ [0] (4f2a005d@gateway/web/freenode/ip.79.42.0.93) 22.03.54 Quit Cunctator (Ping timeout: 246 seconds) 22.07.25 Quit mpmc (Quit: ZNC - http://znc.in) 22.07.39 Join lebellium [0] (~chatzilla@128-79-0-151.hfc.dyn.abo.bbox.fr) 22.07.46 Join mpmc [0] (~mpmc@unaffiliated/mpmc) 22.08.18 Quit mpmc (Remote host closed the connection) 22.10.41 Join yuriks [0] (~quassel@2001:1b40:5ffe:6e2f:f0f6:7dba:ced:6bd4) 22.11.53 Join mpmc [0] (~mpmc@unaffiliated/mpmc) 22.15.58 Quit mpmc (Client Quit) 22.18.51 Quit yuriks (Changing host) 22.18.51 Join yuriks [0] (~quassel@opentyrian/developer/yuriks) 22.22.38 Join RiD [0] (~RiD@31.22.153.38) 22.23.18 Quit RiD (Client Quit) 22.30.21 Quit ikeboy (Quit: Leaving) 22.38.39 *** Saving seen data "./dancer.seen" 22.39.24 Join franklin [0] (~franklin@cpe-071-071-039-006.triad.res.rr.com) 22.39.24 Nick franklin is now known as [Franklin] (~franklin@cpe-071-071-039-006.triad.res.rr.com) 22.49.27 # <[Franklin]> I need some help profiling xworld 22.51.17 Join rela [0] (~x@pdpc/supporter/active/rela) 22.51.22 # wodz: it's working :) 22.51.26 # https://gist.github.com/pamaury/7386b1765657f7ae7bac 22.51.43 # # echo "0xef 0xbe 0x7f 0x04" | ./Debug+Asserts/bin/llvm-mc -disassemble -triple=mipsel-unknown-linux -mcpu=mips32r2 22.51.43 # .text 22.51.43 # synci -16657($3) 22.52.59 # can you try 0x05 0x1f 0x00 0x00 ? 22.53.23 # invalid instruction 22.53.40 # this is not a synci instruction 22.53.51 # oh wait, that's big endian 22.54.01 # synci 0($8) 22.54.01 # yeah 22.54.44 # ok, so the final thing is that $8 is rather unusual way of expressing mips register (although correct AFAIK) 22.55.18 # yeah llvm seems to output registers this ways on mips 22.55.25 # maybe there is an assembler variant to change this 22.56.53 # <[Franklin]> what should I put for the copyright of xworld? 22.57.22 # <[Franklin]> the way it is now, it has (C) 2004 Gregory Montoir 22.57.30 # wodz: if you want to tweak or better understand the code, I suggest you run this with the patch: 22.57.30 # echo "0x05 0x1f 0x00 0x00" | ./Debug+Asserts/bin/llvm-mc -disassemble -triple=mips-unknown-linux -mcpu=mips32r2 -show-encoding -show-inst-operands -show-inst 22.57.38 # synci 0($8) # encoding: [0x04,0x1f,0x00,0x00] 22.57.38 # # # 22.57.38 DBUG Enqueued KICK pamaury 22.57.38 # # > 22.57.41 # very useful ^ 22.58.35 # why encoding differs from actual instruction sequence? 22.59.08 Quit TheLemonMan (Remote host closed the connection) 22.59.42 # pamaury: ^ 22.59.51 # that's a good question 23.00.13 # my guess is that this is the encoding of the decoded instruction, now why is that different, no idea 23.00.35 # Build Server message: 3New build round started. Revision edb0c6c, 255 builds, 25 clients. 23.01.31 # I'm not familar with mips encoding of registers but 0x05 0x1f 0x00 0x00 would be imm=0, base=16 right ? 23.03.34 # this sequence should be decoded as synci 0($t0) t0 is $8 so decoding is correct 23.04.19 # imm is 0, base should be 8 I guess 23.05.46 # ah right my mistake 23.06.09 # so this encoding output is wrong, that's weird 23.07.35 # anyway the decoding is correct 23.07.54 # apparently mips doesn't have an asm variant so there is no way to print something else than $8 23.08.34 # Right, thanks a bunch. Now I need to study this more to understand completely 23.09.20 Quit rela (Ping timeout: 244 seconds) 23.09.38 # I suggest you submit submit this on the llvm mailing list to get some feedback (otherwise I will), just to check this is the right way of doing it 23.09.47 # and ask about the register printing 23.10.34 # well, you are the author 23.11.06 # ok I'll do it, let's pretend "synci" will one day be useful to someone ^^ 23.11.49 # pamaury: it will be imported to capstone :-) 23.12.01 # ah yeah right 23.12.42 # gtg, good night 23.12.50 # pamaury: For merging in llvm I guess assemebler must be provided as well 23.13.01 # Build Server message: 3Build round completed after 746 seconds. 23.13.02 Join xorly [0] (~xorly@m180.dkm.cz) 23.13.06 Quit Makinit (Quit: bye) 23.13.14 # yes indeed 23.14.05 Join Makinit [0] (makinit@makinit.nl) 23.14.13 # <[Saint]> Cunctator_: you still around? 23.14.57 Ctcp Ignored 1 channel CTCP requests in 0 seconds at the last flood 23.14.57 # * gevaerts kicks lillebror1-zagor out 23.14.59 # lillebror1-zagor has problems 23.15.27 # I'll try to catch zagor tomorrow 23.15.35 # For now though, that client is blocked 23.15.35 # <[Saint]> Cunctator_: boot to the emCORE main menu; select the 'Tools' menu item; select 'Run Rockbox fallback image'; mount the device; transfer your files; shutdown and reboot the device to return to the user installed Rockbox image. 23.16.21 # <[Franklin]> [Saint]: what should I put for xworld's copyright? 23.17.21 # <[Saint]> Cunctator_: USB is known to be fairly problematic on this target for a while now, this is (one of) the reason(s) a known stable Rockbox fallback image is embedded in the emCORE image. 23.17.25 Quit lebellium (Quit: ChatZilla 0.9.91 [Firefox 34.0/20141110195804]) 23.18.00 # <[Saint]> Cunctator_: for some odd reason, the USB issues only seem to predictably affect Windows operating systems. 23.18.30 Quit pamaury (Ping timeout: 265 seconds) 23.18.33 # <[Saint]> Cunctator_: on linux based operating systems, it appears to 'Just Work(TM)'. 23.20.57 # hey Saint] 23.21.01 # [Saint] 23.21.06 # yes im still around 23.21.36 # <[Saint]> Try the fallback image as described above and gimme a yell afterwards. 23.22.02 # I tried that. But either 1) my fallback image doesn't load (stuck on checking database or something) or 2) even if I manage to launch it, the very same issue with blocked transfers happen 23.22.12 # I will try once again right now, just to be sure 23.22.17 # but I already did that twice 23.22.23 Quit wodz (Quit: Leaving) 23.22.30 # I'd start suspecting possible disk issues 23.22.47 # <[Franklin]> Cunctator_: do you have all the data backed up? 23.22.56 # on a brand new iPod? I literally tore the plastic from the box yesterday 23.23.05 # yeah, all the music is stored on my laptop 23.23.39 # That does make it less likely, yes 23.23.40 # <[Saint]> Nuke the /.rockbox fodler entirely and extract a new image to it. 23.23.46 # <[Saint]> *folder 23.23.55 # <[Franklin]> [Saint]: could it be that ATA driver commit? 23.24.02 # <[Saint]> No. 23.24.11 # use this build if in doubt: http://caudec.net/rockbox/rockbox-ipod6g-440ff95M-140517.zip 23.24.34 # old build from before all the new stuff 23.24.36 # "old" 23.24.46 # <[Saint]> Can we not push people on to unofficial builds please? 23.24.46 # 's all relative, eh 23.24.59 # <[Saint]> Lets try to actually debug this first. 23.25.07 # so, plug it in, delete the .rockbox folder, and then copy and paste a new one? 23.25.15 # <[Saint]> Correct. 23.25.28 # lets try 23.25.35 # <[Saint]> That way we'll know exactly what we're working with. 23.26.47 # the version I've got here sitting on my desktop is: bcca111-141110 23.26.49 # is that alright? 23.26.58 # <[Saint]> Yeah, that's fine. 23.27.58 # (is it normal that Win 8.1 always tells me "there's a problem with this drive. Scan the drive and fix it now"?) 23.28.08 # upon plugging in 23.28.48 # <[Saint]> In some cases it will bitch about this, yes. Its fine to let it do whatever it thinks it wants to do. 23.29.49 # <[Saint]> Windows is a strange creature and I'm not entirely sure why it has a cry about the way the drive is set up. 23.30.16 # <[Saint]> Maybe it doesn't like large FTA32 superfloppy volumes. I honestly don't know. 23.30.20 # <[Saint]> *FAT32 23.31.49 # OK copied successfully. Now all that is shows on the iPod is: the new ".rockbox" folder, an ".apps" folder, a "Music" folder (which I created) and an invisible "iPod Controls" folder. 23.31.55 # now i safely extract and reboot 23.32.35 # <[Saint]> I...hmmmm. That ipod controls' folder shouldn't be there. 23.32.57 # <[Saint]> The volume should have been formatted during the installation. 23.33.19 # OK bad news. As I launched RockBox a red error message came out, saying there was some sort of error, and that it is trying to load fallback image 23.33.34 # ...which it just did 23.34.11 # <[Saint]> Do you recall the error that was presented to you? 23.34.22 Quit yuriks (Remote host closed the connection) 23.34.25 # it was pretty quick...sorry 23.34.32 # <[Saint]> If you could see if it is repeatable, its useful information. 23.34.40 # OK 23.36.07 # odd 23.36.12 # it didnt do it anymore 23.36.33 # i resetted, relaunched, SCANNING DISK for a while, and then it launched fine 23.37.10 Join yuriks [0] (~quassel@opentyrian/developer/yuriks) 23.37.24 # <[Saint]> Hmmm. Ok. Well, if it decided to work, I'm cool with that. Now, this time just try transferring your media solely through the Rockbox fallback image. 23.37.46 # ok, so I go back again and try to launch the fallback image 23.37.58 # <[Saint]> Correct. 23.38.19 # <[Saint]> The basic premise is that you'll need to launch the fallback image for all USB interactions. 23.38.38 # Balls. 23.39.00 # ok it launched 23.39.05 # now I plug in the usb 23.39.29 # (how do i make sure it indeed IS the fallback image?) 23.41.42 # <[Saint]> By way of launching the fallback image via the 'emCORE -> Tools -> Run Rockbox fallback image' menu flow. 23.42.12 # yeah thats what i did 23.42.19 # anyway, crap. The same thing happened 23.42.32 # starts moving the files, 2 maybe 3% 23.42.41 # stops 23.43.06 # even worse, now the iPod all of a sudden got unplugged by himself 23.43.28 # * [Franklin] is leaning towards hardware failure 23.43.50 # <[Saint]> [Franklin]: its literally brand new. 23.44.05 # now its unresponsive, black screen. I can see the USB icon, but no backlight 23.44.20 Quit ender` (Quit: Calling ‘atheism’ a religion is like calling ‘bald’ a hair color. -- Don Hirschberg) 23.44.22 # yes guys, I promise. I opned the box yesterday 23.44.33 Quit amayer (Quit: Leaving) 23.44.35 # <[Saint]> Cunctator_: I'm thinking that the prior errors may have left the disk in a somewhat funky state. WHat we'll try to do now is format the disk via the emCORE main menu. 23.44.44 # alright 23.44.59 # "reformat data partition"? 23.45.08 # <[Saint]> Again, go to the emCORE main menu; select 'Tools'; and then select reformat data partition'. 23.45.14 # <[Saint]> right, correct. 23.45.17 # OK 23.45.43 # yellow message: "reformatting data partition" 23.45.48 # blue progress bar 23.45.49 # <[Saint]> After which, when you boot Rockbox from the emCORE main menu, the fallback image will load automatically. 23.46.31 # <[Saint]> ANd then wel'll try to add the new Rockbox image and some media at the same time. 23.46.39 # alright 23.46.40 # <[Saint]> bah - can't type. 23.46.45 Quit markun (Remote host closed the connection) 23.49.30 # almost there 23.49.53 # <[Franklin]> so what should I put for the copyright notice? 23.49.57 # <[Franklin]> (for xworld) 23.51.40 # "data partition has been formatted" 23.51.47 # I try and relaunch Rockbox? 23.51.57 # <[Franklin]> yeah 23.51.59 # <[Franklin]> go ahead 23.52.02 # <[Saint]> Correct. 23.52.27 # loading rockbox ipod failed, loading fallback image (as planned) 23.52.40 # OK, I plug it in 23.52.41 # <[Franklin]> ok, now try a transfer 23.52.45 # <[Saint]> Correct.. 23.53.05 # <[Franklin]> [Saint]: sorry, I'll let you do this :) 23.53.25 # <[Saint]> the additional . was a typographical error. :) 23.53.58 # <[Franklin]> lol 23.55.30 # ummm. OK, the drive isnt called iPod Classic anymore, but Removable Disk. Also, it's not opening. 23.55.51 # just froze my windows explorer 23.55.59 # <[Franklin]> try waiting a bit 23.56.08 # sigh 23.57.10 # <[Saint]> This is all quite strange. Are the any iTunes processes still running? 23.58.01 # it totally fucked my explorer. I had to plug it out 23.58.10 # <[Saint]> What method of installation was it that you used? 23.58.10 # umm, no 23.58.25 # the one on the FreeMyIpod website, step by step 23.58.43 # <[Saint]> right, but, the w/ itunes, without itunes, etc.