| 00:00:54 | | Quit gevaerts ("sleeping time") |
| 00:02:04 | petur | bah, I was one word wrong, the 1gb sandisk doesn't comply |
| 00:03:18 | amiconn | ehhhh... |
| 00:03:44 | amiconn | Our 'force pio mode' setting checks a rather arbitrary bit... |
| 00:03:53 | amiconn | It's doing so for years though |
| 00:04:11 | amiconn | 4 years to be precise |
| 00:05:10 | petur | all these feature bits are 'not to be depended on' |
| 00:05:25 | petur | so what do we use.... lousy spec |
| 00:06:24 | | Quit jgarvey ("Leaving") |
| 00:06:29 | amiconn | According to the ata specs those words are valid if the indicator bits are set approrpiately |
| 00:06:46 | petur | for pio modes we should look at word 64 |
| 00:07:04 | amiconn | Word 64 lists the modes, but not whether modes can be set |
| 00:07:27 | petur | 'Word 64: Advanced PIO transfer modes supported' |
| 00:07:32 | amiconn | Modes can always be set, hence there is no indicator bit for that, and Linus used a bit that's (nearly) guaranteed to be 1 |
| 00:10:07 | petur | so we should check the indicator bits before usin g words 82-84 |
| 00:11:02 | amiconn | Perhaps. But that a-data card does indicate words 82..84 as valid |
| 00:11:05 | | Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon) |
| 00:11:17 | petur | yes (I was just typing that) |
| 00:11:29 | amiconn | Words 85..87 are indicated as invalid though |
| 00:12:23 | petur | maybe nobody bothers to use advanced powermanagement... |
| 00:12:27 | | Join Zarggg [0] (n=zarggg@216-15-73-111.c3-0.eas-ubr6.atw-eas.pa.cable.rcn.com) |
| 00:12:59 | * | petur is doing a battery benchmark on his H10 with and without to see the difference |
| 00:13:45 | amiconn | H10 is probably a bad target for that test |
| 00:13:57 | petur | contrary |
| 00:14:02 | amiconn | Those microdrive do at least partially implement the cf standard |
| 00:14:26 | amiconn | And on cf apm probably doesn't matter because cf power management is pretty much automatic |
| 00:14:27 | petur | I am testing the patch that disables powermanagement for CF |
| 00:14:48 | petur | so I want to know if this influences H10 |
| 00:15:04 | petur | because the microdrive reports as CF |
| 00:15:07 | amiconn | It could also influence the mini, or the mrobe100, you never know |
| 00:15:30 | amiconn | No, it doesn't |
| 00:15:57 | petur | still need a nano owner to test the modified soft_reset |
| 00:16:06 | amiconn | The CF signature is word 0 == 0x848A according to the standard |
| 00:16:25 | amiconn | And the H10's drive reports 0x0C5A |
| 00:16:47 | petur | I mean the new check which looks at the sector count |
| 00:16:52 | * | amiconn wants to see this cf check removed |
| 00:16:55 | | Join csc` [0] (n=csc@archlinux/user/csc) |
| 00:17:08 | petur | sure.... me too |
| 00:17:38 | petur | why do you think I bother trying and testing stuff? |
| 00:18:07 | amiconn | At least in the form of checking words which aren't actually related to the problematic features |
| 00:18:58 | amiconn | Btw, all the CF cards in the list either report to support the CFA feature set (word 83 bit 2), or have an invalid word 83 |
| 00:19:15 | amiconn | That also applies to the darn microdrives though |
| 00:20:29 | | Join wvmac [0] (n=wvmac@c-71-61-88-3.hsd1.wv.comcast.net) |
| 00:21:19 | | Quit OlivierBorowski (Remote closed the connection) |
| 00:23:22 | | Join DerDome [0] (n=DerDome@dslb-082-083-226-208.pools.arcor-ip.net) |
| 00:24:41 | amiconn | There must be a way to reliably distinguish CF and HDD. The CF standard also specifies PIO modes 5 and 6. The ATA standard does not (not even ATA-7) |
| 00:25:37 | | Quit ender` (" Top reason why compilers are like women: Miss a period and they go crazy") |
| 00:27:29 | petur | hear hear |
| 00:27:42 | petur | you wanted the cf check removed... |
| 00:28:09 | amiconn | I still want that if at all possible |
| 00:28:48 | * | petur has an attention problem, looking at a documentary of The Smiths at the same time ;) |
| 00:29:47 | | Quit roolku () |
| 00:30:39 | * | scorche|sh wonders if petur is a shoplifter |
| 00:31:26 | petur | no, but devconpub must have caused some panic in the streets of london :p |
| 00:31:38 | | Quit Arathis ("Bye, bye") |
| 00:31:48 | | Quit n1s () |
| 00:33:07 | amiconn | Btw, what's the effect of the apm feature set operation failing? |
| 00:33:12 | | Quit kushal_12_27_200 ("This computer has gone to sleep") |
| 00:33:27 | petur | ABRT |
| 00:33:38 | petur | 0x04 |
| 00:34:41 | | Quit ol_schoola () |
| 00:35:05 | amiconn | eh? |
| 00:35:49 | petur | yes, and SENSE returns 0x1F (unsupported) |
| 00:36:34 | | Quit tessarakt ("Client exiting") |
| 00:38:25 | amiconn | hmm |
| 00:38:51 | petur | urgh wrong. it is 'Data Transfer Error / Aborted Command' |
| 00:40:15 | * | petur moves closer to a power socket. laptop and target both low on battery |
| 00:41:14 | | Join spiorf [0] (n=spiorf@host76-226-dynamic.25-79-r.retail.telecomitalia.it) |
| 00:41:45 | | Quit herrwaldo (Read error: 104 (Connection reset by peer)) |
| 00:48:22 | petur | why doesn't features[] use a bitmask iso a bitnumber |
| 00:52:02 | amiconn | Compactness |
| 00:52:33 | amiconn | Right now one array element is 4 bytes, i.e. a power of 2 |
| 00:52:59 | petur | I mean use 0x08 and not 3 |
| 00:53:05 | amiconn | If you relace the bit number (char) with a mask (necessarily unsigned short), it becomes 5 bytes |
| 00:53:27 | | Quit spiorf (Remote closed the connection) |
| 00:53:33 | amiconn | And since that's not a power of 5, addressing array elements needs more code |
| 00:53:38 | petur | would have been easier to check the validate bits at the same time |
| 00:53:40 | amiconn | Ehh, a power of 2 |
| 00:53:58 | petur | ah well |
| 00:54:02 | | Quit kies (Read error: 110 (Connection timed out)) |
| 00:54:13 | amiconn | That whole check should probably be redesigned based on those new findings |
| 00:54:42 | | Quit Soap_ (Read error: 110 (Connection timed out)) |
| 00:54:58 | | Join kies [0] (n=kies@adsl-76-199-14-32.dsl.chmpil.sbcglobal.net) |
| 00:56:08 | | Quit wvmac ("Lost terminal") |
| 00:56:18 | amiconn | petur: What about using the validity bits for these feature *settings* (i.e. word 87 bits 15 and 14) for deciding whether to refrain from setting apm? |
| 00:56:51 | * | petur checks those values for hdd |
| 00:57:15 | amiconn | The hdds all have valid word 87 |
| 00:57:23 | amiconn | _including_ the microdrives |
| 00:57:29 | petur | hmmm |
| 00:57:48 | * | amiconn wants more values |
| 00:58:29 | amiconn | I only have one single CF card - an 1GB Sandisk Ultra II |
| 00:59:17 | petur | what about all those dumps in the wiki? |
| 00:59:24 | amiconn | Hmm, the Sandisk Ultra II 2GB in the list indicates word 87 as valid |
| 01:00 |
| 01:00:09 | petur | same as my sandisk 1GB |
| 01:00:13 | amiconn | Those are about a dozen cards and harddisks. Also, for most of them the "pwrmgt fails" column just says "?" |
| 01:01:18 | | Join japc [0] (n=japc@bl7-242-163.dsl.telepac.pt) |
| 01:02:31 | petur | I wonder if the description of bit 3 of word 85 is a typo: talks about 'supported', not 'enabled' |
| 01:03:24 | amiconn | I am wondering the same, although ata-4, ata-5 and ata-7 all say "supported" |
| 01:03:26 | petur | those hdd dumps were taken after the set_features was executed :/ |
| 01:03:38 | amiconn | Yes of course |
| 01:03:57 | petur | also, word 85 is 0 for those sandisk cards... |
| 01:04:21 | amiconn | The bootloader already does that (how else would it load rockbox to do the dump?) |
| 01:04:29 | * | petur looks for a bit that must be set |
| 01:05:02 | * | amiconn curses thread.c, especially sleep_core() |
| 01:06:01 | amiconn | That function is one big FOREVER loop - and putting my alignment-changing 'nop's in there causes gcc to shuffle instructions all over |
| 01:09:27 | * | petur gives transcend x133 2GB a strange look. word 85 = 1 (supports smart???) |
| 01:09:28 | amiconn | So all I could find out is that the alignment problem is within that function, because correcting the alignment before this call makes it work, and after that call it fails (sleep_core() is inlined within switch_thread()) |
| 01:09:58 | amiconn | petur: It doesn't. Word 85 isn't valid according to word 87 |
| 01:10:36 | petur | hmmm my cf spec says nothing about word 87 :( |
| 01:10:50 | petur | oh it does |
| 01:11:11 | petur | that stuff should be in nice tables, not text |
| 01:12:19 | | Quit amiconn (Nick collision from services.) |
| 01:12:25 | | Join amiconn [50] (n=jens@rockbox/developer/amiconn) |
| 01:12:37 | *** | Saving seen data "./dancer.seen" |
| 01:12:49 | | Quit Nico_P (Remote closed the connection) |
| 01:13:14 | * | amiconn wonders where the threading guru is when he needs him |
| 01:13:16 | amiconn | jhMikeS: ping |
| 01:13:20 | | Join rgould [0] (n=rgould@osgeo/member/rgould) |
| 01:14:25 | | Quit tvelocity__ (Remote closed the connection) |
| 01:14:53 | | Join Jessica464 [0] (n=Jessica@adsl-75-19-179-90.dsl.bltnin.sbcglobal.net) |
| 01:15:13 | Jessica464 | Hey folks. Anyone still have an old Archos Jukebox? |
| 01:16:03 | | Join cool_walking_ [0] (n=root@203-59-129-195.perm.iinet.net.au) |
| 01:16:31 | pixelma | some do, just ask your actual question and see (have a bit patience too) |
| 01:17:17 | Jessica464 | I dropped 2500mAh niMH in my jukebox, but the battery life is like 25% what it used to be? any clues? The battery change happened at the same time i installed rockbox. |
| 01:17:48 | Jessica464 | i figured with 2500mAh the thing would run foreever |
| 01:18:23 | Jessica464 | Also, anyone replaced the 20GB HD with a newer 160GB? any issues doing so? |
| 01:19:17 | Jessica464 | thx for the tip, pix. Unfortunately, I catch my flight for Shenzhen in a few hours. At the mo, it seems like it is gonna be a silent flight. |
| 01:20:32 | amiconn | Did you try charging the batteries in an external charger and check runtime? |
| 01:20:51 | pixelma | also those high capacity cells often suffer from high self discharge (there is a new generation out there that doesn't) |
| 01:20:57 | amiconn | Also, what jukebox is that, Player/Studio or Recorder? |
| 01:20:58 | Jessica464 | doing that now, but I don't know if they'll fully cook b4 i need to leave. |
| 01:21:04 | Jessica464 | recorder |
| 01:21:36 | scorche|sh | Jessica464: i have replaced it with a 160, but keep in mind that unless you enable LBA48 support, you will only get 128/137 GiB/GB |
| 01:21:39 | amiconn | Could be that our charging algorithm doesn't fully charge the batteries. What rockbox build are you running? |
| 01:22:10 | Jessica464 | hmmm... not sure on the build. lemme check |
| 01:23:53 | | Quit DerDome ("Leaving.") |
| 01:25:30 | Jessica464 | the date is 7/24/07 |
| 01:25:35 | Jessica464 | on my RB build |
| 01:25:42 | | Join WatchDragon [0] (n=a@ip68-4-103-162.oc.oc.cox.net) |
| 01:26:23 | amiconn | Umm, that's old.... although, it's at least not the 2.5 release from the rockbox stone age |
| 01:26:34 | Jessica464 | are niMH ok, or should I be using nicad? |
| 01:26:51 | amiconn | Regarding the high self-discharge rate of high capacity NiMH - I'd recommend everyone to use the new generation cells instead. They are "only" avaliable as 2100mAh cells (Sanyo: 2000mAh), but lengths better than ordinary high capacity cells |
| 01:26:52 | jhMikeS | amiconn: yez |
| 01:27:01 | scorche|sh | Jessica464: NiMH only |
| 01:27:25 | Jessica464 | ok scor |
| 01:27:26 | * | amiconn experienced ordinary high capacity cells emptyig themselves within 1..2 days |
| 01:28:01 | Jessica464 | i'm getting 2~3 hours from these 2500mAHs!!!! |
| 01:28:09 | | Part toffe82 |
| 01:28:34 | amiconn | Directly after charging? |
| 01:28:39 | scorche|sh | that definitely seems off if they are brand new...can you get longer runtimes by putting in your old betteries? |
| 01:29:05 | Jessica464 | amicon, yes |
| 01:29:22 | Jessica464 | scor: they were new when i put them in the archos |
| 01:29:32 | Jessica464 | i had to recycle the old ones |
| 01:30:33 | scorche|sh | Jessica464: well, it *could* be other things as well, but testing different batteries is the easiest thing to check |
| 01:30:54 | Jessica464 | yeah, shame i have only a couple of hours... |
| 01:31:04 | Jessica464 | and no other batteries! |
| 01:32:22 | * | jhMikeS read what amiconn said and experiences a 'wtf' moment |
| 01:33:01 | amiconn | jhMikeS: Regarding what part? |
| 01:33:17 | scorche|sh | the charging algorithym? |
| 01:33:28 | petur | Jessica464: get some non-rechargeables at the airport? |
| 01:33:49 | scorche|sh | Jessica464: and if you get non-rechargables, make sure not to accidently charge them ;) |
| 01:33:55 | jhMikeS | amiconn: putting nops in sleep_core having the effect you reported. I think at first commit the function was in IRAM. |
| 01:33:57 | Jessica464 | pet: like alcaline? |
| 01:33:57 | pixelma | that could be dangerous if one then inserts a charger... |
| 01:33:59 | petur | scorche|sh: the threading problem |
| 01:34:07 | scorche|sh | ah |
| 01:34:37 | saratoga | jhMikeS: someone suggested you knew something about why the backlight might be very slow to active lately? |
| 01:34:38 | amiconn | I tracked down the location where the PP5002 obviously originates, but unfortunately it's not very precise, because I cannot put nops *into* sleep_core() without making gcc shuffling instructions all around |
| 01:35:11 | amiconn | But why does *code* alignment have an effect at all? And why there? |
| 01:35:41 | amiconn | petur, Jessica464: Alkalines shouldn't be used in the archos |
| 01:36:07 | jhMikeS | saratoga: it's the priority and a side effect of actually fixing the scheduling. I'd suggest just making it a normal UI prority thread for the time being. |
| 01:36:12 | Jessica464 | ami, then what non-rechareables should I use? |
| 01:36:37 | petur | Jessica464: seems I gave wrong advice then... |
| 01:36:48 | amiconn | Jessica464: None. The archos jukeboxes are designed to run from NiMH. Other batteries will either work poorly, or not at all |
| 01:36:58 | saratoga | jhMikeS: how are thread priorities adjusted? I'd like to play with this a bit |
| 01:37:15 | Jessica464 | ok. so it's rechargeables or nuthin |
| 01:37:49 | Jessica464 | anyone know what voltage and current I should be charging the archos with? |
| 01:37:55 | scorche|sh | amiconn: i would think lithium would sate the power demands just fine, no? |
| 01:38:03 | jhMikeS | amiconn: that's weird. Now that you found that I'm brainstorming for an answer on that. |
| 01:38:05 | amiconn | Most probably not |
| 01:38:22 | scorche|sh | why not? |
| 01:38:41 | jhMikeS | saratoga: the create_thread call in backlight.c inside IF_PRIO() |
| 01:38:50 | amiconn | scorche: The circuitry is designed for the typical openloop voltage of NiMH cells, i.e. 4x 1.2V. Non-rechargeables start at 1.5V |
| 01:39:07 | jhMikeS | saratoga: use PRIORITY_USER_INTERFACE |
| 01:39:13 | amiconn | And the jukeboxes need cells with a very low impedance, because of the high hdd spinup current |
| 01:39:39 | scorche|sh | yes...the lithium should be fine with regards to impedance |
| 01:39:55 | saratoga | jhMikeS: instead of "PRIORITY_SYSTEM"? |
| 01:40:16 | jhMikeS | saratoga: yes |
| 01:40:22 | Jessica464 | lithium wont charge properly in a device designed to charge nimh, will it? |
| 01:40:37 | scorche|sh | Jessica464: no...i am talking just about usage |
| 01:40:57 | saratoga | putting lithium in anything but a lithium charger will probably result in an on fire battery |
| 01:41:06 | Jessica464 | ok |
| 01:41:30 | Jessica464 | so, for niMH in an archos rockbox, what voltage and current should I charge it at? |
| 01:44:03 | | Join csc`` [0] (n=csc@archlinux/user/csc) |
| 01:44:13 | saratoga | jhMikeS: that fix seems to have worked fairly well |
| 01:44:14 | amiconn | When charging them in the device, you should either use the original charger (9V unregulated), or a *regulated* 12V charger. Less won't fully charge the batteries, more will produce unnecessary heat and eventually break the charging circuit |
| 01:44:15 | * | jhMikeS has to get this priority inheritance done and it's the hardest piece and most time-consuming of programming undertaken yet but what's done works extremely well (and as a bonus alot of kernel stuff is more simple) |
| 01:44:22 | saratoga | takes about half a second to come on now |
| 01:44:28 | saratoga | instead of the 2-5 seconds before |
| 01:44:45 | amiconn | If you charge them externally, use a proper NiMH charger |
| 01:45:47 | Jessica464 | ami, but at what current? |
| 01:45:54 | jhMikeS | saratoga: with the changes I've made (somewhat of a 2nd-pass rewrite), even at PRIORITY_SYSTEM, the backlight delay is hardly perceivable :) |
| 01:46:43 | | Quit csc` (Nick collision from services.) |
| 01:46:46 | | Nick csc`` is now known as csc` (n=csc@archlinux/user/csc) |
| 01:46:48 | * | amiconn managed to track down the crash location a bit further |
| 01:46:50 | amiconn | jhMikeS: The problem must be somewhere between start of sleep_core(), and the core_sleep() call (line 1466) |
| 01:47:08 | amiconn | core_sleep() is asm, so I could sneak in 2 extra nops without gcc going wild |
| 01:48:41 | saratoga | jhMikeS: these are going to be committed relatively soon? |
| 01:48:58 | saratoga | i'm wondering if its worth working around this issue in SVN in the meantime |
| 01:49:33 | amiconn | There are 55 instrucitons between the 2 known points |
| 01:49:56 | jhMikeS | saratoga: I'd just commit it. Really I think all UI-based threads should be equal since they're expected to response well to input. |
| 01:50:08 | saratoga | ok will do |
| 01:50:34 | jhMikeS | amiconn: you put nops both at the start and the end? |
| 01:50:41 | amiconn | yes |
| 01:51:16 | jhMikeS | that code didn't even exist for PP5002 at the initial commit though. that only came with finishing dual-core support there and the same problem creeps in even if it's the C reference version. |
| 01:52:04 | amiconn | I first tested the effect of a varying number of nops in main(). With my current build, 0 or 3 nops work fine, 1 or 2 nops make it crash. The effects are modulo-4 |
| 01:52:38 | amiconn | Then I put 4 nops into main() - everything ok. Then I started moving 2 of the 4 nops around the code. |
| 01:52:43 | Jessica464 | ami: do you know the current i should be charging this thing at? |
| 01:53:21 | amiconn | If I moved them too far to the end, it was crashing, moved them back to the beginning, worked again etc (bisecting) |
| 01:53:37 | amiconn | Jessica464: The current will be determined by the device. |
| 01:53:55 | amiconn | YOur charger should be capable of providing at least 400mA |
| 01:54:12 | amiconn | All this must be somewhere in the wiki |
| 01:54:28 | Jessica464 | err... i no longer have the original charger, and there is no spec on the archos itself. |
| 01:55:16 | amiconn | jhMikeS: I finally found that putting the nops *directly* before the sleep_core() call in switch_thread(), it works, and if I put them directly after that call, it crashes |
| 01:55:24 | amiconn | (sleep_core() is inlined) |
| 01:55:30 | Jessica464 | is there a wiki for the archos? |
| 01:56:11 | amiconn | And the tiny bit further I got is that if I put the nops into core_sleep() it still crashes, so the problem must be before that |
| 01:57:44 | | Quit BlakeJohnson86 ("Leaving.") |
| 01:59:07 | petur | amiconn: uploaded a new CF patch that should be committable, feel free to comment (no hurry) http://www.rockbox.org/tracker/task/8644 |
| 01:59:15 | amiconn | jhMikeS: http://pastebin.ca/931215 is the whole, inlined sleep_core(). The problem must be somewhere between the beginning and 0x4404c |
| 01:59:42 | | Join jurrie [0] (n=jurrie@adsl-068-209-041-021.sip.asm.bellsouth.net) |
| 02:00 |
| 02:00:07 | petur | amiconn: my mk8007gah still reports powermanagement in word 85 even if not set, so looks good :) |
| 02:00:32 | amiconn | The bootloader would set it... |
| 02:00:49 | petur | argh... right :/ |
| 02:01:13 | petur | will play with that tomorrow - bedtime now |
| 02:01:21 | | Quit petur ("Zzzzz") |
| 02:02:55 | jhMikeS | amiconn: why do you cut it off at the nops in core_sleep? |
| 02:03:04 | amiconn | ? |
| 02:03:49 | jhMikeS | 0x4404c is the first nop of core_sleep assembly |
| 02:03:53 | amiconn | yes |
| 02:04:27 | amiconn | And if I put the 2 extra nops there (my only chance to put extra nops into sleep_core()), it still crashes, so the problem must be before that |
| 02:05:17 | | Quit Jessica464 ("http://irc.netsplit.de/") |
| 02:06:08 | amiconn | ...becasue my system of bisecting was designed so that the build crashes if the nops come after the alignment sensitive code, and does not crash if they come before that code |
| 02:07:22 | | Quit jurrie__ (Read error: 110 (Connection timed out)) |
| 02:07:47 | * | amiconn realises that he can put the testing nops a little bit further up |
| 02:07:48 | jhMikeS | this is twilight zone stuff. 1) Why only PP5002 2) why there of all places |
| 02:08:52 | jhMikeS | hmmm...one thing happens there that happens nowhere else |
| 02:10:27 | jhMikeS | it's the place where context is whatever context exists at the time no threads are actually scheduled but it's running in the context of the last thread that became non-runnable |
| 02:10:32 | | Join toffe82 [0] (n=chatzill@adsl-70-235-227-56.dsl.frs2ca.sbcglobal.net) |
| 02:11:07 | amiconn | eh.... |
| 02:11:13 | amiconn | Seems we're lucky |
| 02:11:28 | amiconn | The problem seems to be *in* core_sleep() itself |
| 02:12:25 | amiconn | Now I'm moving the nops intsruction-by-instruction |
| 02:16:40 | jhMikeS | usage of str<cond>b instead of str<cond> to the processor control reg? <== big guess |
| 02:17:35 | * | amiconn suspects that one instructions and skips a few to hopefully get a result earlier |
| 02:18:10 | amiconn | Does the OF uses 32 bit accesses? |
| 02:18:13 | amiconn | *use |
| 02:18:38 | | Join BlakeJohnson86 [0] (n=bjohnson@c-67-190-234-51.hsd1.mn.comcast.net) |
| 02:18:55 | * | jhMikeS is checking a disassembly if he can locate it |
| 02:22:02 | amiconn | OF uses either STRH or STR |
| 02:22:14 | amiconn | Some places also have 3 NOPs before *and* 3 NOPs after t |
| 02:22:19 | amiconn | *it |
| 02:22:39 | amiconn | Actually I checked 2nd Gen ROM code |
| 02:22:42 | jhMikeS | The pipeline seems to be a factor for that |
| 02:23:38 | jhMikeS | or prefetch. for example, a sleep write immediately followed by a wake write will actually not sleep because the next instruction gets executed after the sleep write before it sleeps |
| 02:24:20 | amiconn | It's definitely the streqb |
| 02:24:28 | amiconn | I'm trying streqh next |
| 02:24:57 | amiconn | If that doesn't help, I'll put 3 nops in front as well (and balance with further nops for my test) |
| 02:25:41 | jhMikeS | core_wake should also be altered then as well as the CORE_CTL definition |
| 02:25:48 | amiconn | meh, streqh doesn't like the addressing mode |
| 02:26:21 | | Quit axionix_ ("Lost terminal") |
| 02:26:35 | | Join axionix [0] (n=axion@cpe-74-70-239-192.nycap.res.rr.com) |
| 02:27:30 | * | amiconn tries tsreq instead |
| 02:27:33 | amiconn | *streq |
| 02:28:58 | amiconn | It's not the byte access - streq also crashes |
| 02:31:23 | jhMikeS | but that instruction crashes or doesn't crash depending on whether the nops precede or follow? |
| 02:32:35 | | Quit DaCapn (Read error: 110 (Connection timed out)) |
| 02:32:42 | | Quit axionix (Read error: 104 (Connection reset by peer)) |
| 02:33:10 | amiconn | It crashes or doesn't crash depending on how it's aligned |
| 02:33:17 | | Join axionix [0] (n=axion@cpe-74-70-239-192.nycap.res.rr.com) |
| 02:33:43 | jhMikeS | just the alignment of that one instruction? |
| 02:33:50 | amiconn | yes |
| 02:34:29 | jhMikeS | what precise alignment is needed for that? |
| 02:35:25 | jhMikeS | or did you imply 16 bytes? |
| 02:35:55 | | Join fasmaie [0] (n=yohann@c-98-216-170-85.hsd1.ma.comcast.net) |
| 02:36:27 | amiconn | It must be aligned at xxxxxx4 or xxxxxxx8 in order to work, if I didn't make a mistake |
| 02:38:58 | jhMikeS | possible to pad before/after with words? I know you can pad with specific byte values in gas. |
| 02:43:30 | | Quit jhulst (Connection timed out) |
| 02:43:33 | | Join ztidwell [0] (n=48c8cf6c@gateway/web/cgi-irc/labb.contactor.se/x-5dcf72c10b336b3e) |
| 02:45:18 | | Quit ztidwell (Client Quit) |
| 02:45:24 | | Join ztidwell [0] (n=48c8cf6c@gateway/web/cgi-irc/labb.contactor.se/x-848304306e3fa820) |
| 02:46:10 | | Quit ztidwell (Client Quit) |
| 02:46:28 | amiconn | The ROM seems to have this instruction at all possible alignments, although most instances handle the cop only |
| 02:47:34 | jhMikeS | the cop never has to see any of the thread code for the alignment crashing to surface |
| 02:48:02 | amiconn | Yes, but the cop also crashes if it's not aligned properly |
| 02:48:38 | amiconn | It alost seems there's something in the deisgn of the PP5002 that makes it transfer values from cpu_control to the PC |
| 02:48:51 | amiconn | I'fe seen crash addresses of 0xCF004084 several times now... |
| 02:49:53 | * | amiconn wonders whether the addressing mode using the shifter is involved |
| 02:50:08 | amiconn | But I need to stop experimenting for now and get some sleep |
| 02:50:34 | jhMikeS | the writes to the control register after the first DC commit didn't use that for pp5002 |
| 02:56:14 | | Quit csc` ("Powering Off") |
| 03:00 |
| 03:07:21 | | Nick pikhq_ is now known as pikhq (n=pikhq@209.248.124.243) |
| 03:07:51 | | Nick fxb is now known as fxb__ (n=felixbru@h1252615.stratoserver.net) |
| 03:12:38 | *** | Saving seen data "./dancer.seen" |
| 03:17:32 | | Quit argumentD (Read error: 110 (Connection timed out)) |
| 03:24:25 | | Part pixelma |
| 03:26:45 | | Join DaCapn [0] (n=dacapn@c-76-105-220-239.hsd1.or.comcast.net) |
| 03:36:13 | | Nick fxb__ is now known as fxb (n=felixbru@h1252615.stratoserver.net) |
| 03:36:33 | | Quit fasmaie (Read error: 113 (No route to host)) |
| 03:40:56 | | Nick fxb is now known as fxb__ (n=felixbru@h1252615.stratoserver.net) |
| 03:42:08 | | Nick fxb__ is now known as fxb (n=felixbru@h1252615.stratoserver.net) |
| 03:44:50 | | Nick fxb is now known as fxb__ (n=felixbru@h1252615.stratoserver.net) |
| 03:52:46 | | Join jcollie [0] (n=jcollie@dsl-ppp239.isunet.net) |
| 04:00 |
| 04:50:15 | | Quit miepchen^schlaf (Read error: 110 (Connection timed out)) |
| 04:50:29 | | Join miepchen^schlaf [0] (n=el_miep@p54BF42F3.dip.t-dialin.net) |
| 05:00 |
| 05:06:37 | | Join major_works [0] (n=chatzill@ool-45745dcc.dyn.optonline.net) |
| 05:09:58 | | Join webguest09 [0] (n=40e62bc8@gateway/web/cgi-irc/labb.contactor.se/x-1d3abc194bc0b934) |
| 05:11:03 | | Quit webguest09 (Client Quit) |
| 05:11:17 | | Quit Horscht (Read error: 104 (Connection reset by peer)) |
| 05:12:41 | *** | Saving seen data "./dancer.seen" |
| 05:12:50 | major_works | Is there a knowledgeable person on who could look at this and help me understand why it's causing a make to fail? |
| 05:12:54 | major_works | http://pastebin.com/ddaa336c |
| 05:18:22 | XavierGr | make sure to: have the latest source and update your repository, make clean, reconfigure |
| 05:18:49 | XavierGr | or even make veryclean |
| 05:18:59 | major_works | It was the latest source. There are patches applied, however. |
| 05:19:16 | JdGordon | svn up |
| 05:19:23 | JdGordon | looks like your not latest |
| 05:19:29 | JdGordon | or a patch is borken |
| 05:19:32 | | Quit Thundercloud (Remote closed the connection) |
| 05:20:05 | | Join Horscht [0] (n=Horscht@xbmc/user/horscht) |
| 05:20:06 | major_works | Oh yeah, I did have to hand-patch some stuff. Multifont for one; viewports for another. Both are out of sync at this point. |
| 05:20:39 | major_works | But it was a fairly straightforward hand patch, which is good, because I am not a coder by any stretch. |
| 05:21:24 | XavierGr | then type svn diff | patch -R to remove all differences and then svn up as JdGordon said |
| 05:29:47 | | Quit mchua ("Holy shiny distraction, Batman!") |
| 05:30:02 | major_works | I'll start over with your suggestions. thanks. |
| 05:31:07 | JdGordon | svn revert -R <path> is a better way to blow away all changes |
| 05:37:59 | | Quit Horscht ("User was distributing pornography on server; system seized by FBI") |
| 05:42:26 | | Quit DaCapn (Remote closed the connection) |
| 05:43:00 | | Join aliask [0] (n=chatzill@rockbox/developer/aliask) |
| 05:43:38 | | Join x86ia [0] (n=chatzill@189.156.220.10) |
| 05:58:13 | | Quit kies (Connection timed out) |
| 05:59:17 | | Join kies [0] (n=kies@adsl-76-199-3-132.dsl.chmpil.sbcglobal.net) |
| 06:00 |
| 06:02:34 | | Join jurrie_ [0] (n=jurrie@adsl-068-209-041-021.sip.asm.bellsouth.net) |
| 06:09:53 | | Join argumentD [0] (n=argument@cpe-76-173-115-95.socal.res.rr.com) |
| 06:15:05 | | Quit jurrie (Read error: 110 (Connection timed out)) |
| 06:20:36 | | Quit midgey () |
| 06:23:04 | | Quit saratoga ("CGI:IRC (Ping timeout)") |
| 06:25:23 | | Quit toffe82 ("ChatZilla 0.9.81 [Firefox 2.0.0.12/2008020121]") |
| 06:30:41 | | Quit major_works ("ChatZilla 0.9.81 [Firefox 2.0.0.12/2008020121]") |
| 06:32:36 | | Join toffe82 [0] (n=chatzill@adsl-70-235-227-56.dsl.frs2ca.sbcglobal.net) |
| 06:37:42 | | Quit desowin () |
| 06:41:06 | | Join midgey [0] (n=tjross@westquad-188-46.reshall.umich.edu) |
| 06:43:12 | | Quit wisp_sh (Read error: 110 (Connection timed out)) |
| 06:45:11 | | Quit corevette (Remote closed the connection) |
| 06:47:05 | | Join toffe82_ [0] (n=chatzill@adsl-70-235-227-56.dsl.frs2ca.sbcglobal.net) |
| 06:50:36 | | Join ol_schoola [0] (n=meatwad@c-67-167-20-91.hsd1.il.comcast.net) |
| 06:51:33 | | Quit Xerion (Read error: 104 (Connection reset by peer)) |
| 06:51:39 | | Join Xerion_ [0] (i=xerion@cp198589-d.landg1.lb.home.nl) |
| 06:51:53 | | Nick Xerion_ is now known as Xerion (i=xerion@cp198589-d.landg1.lb.home.nl) |
| 06:52:04 | | Join Siku [0] (i=Siku@e81-197-70-46.elisa-laajakaista.fi) |
| 06:53:56 | | Quit toffe82 (Read error: 110 (Connection timed out)) |
| 06:57:18 | | Quit ChanServ (card.freenode.net irc.freenode.net) |
| 06:57:18 | NSplit | card.freenode.net irc.freenode.net |
| 06:57:18 | | Quit markun (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit kies (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit miepchen^schlaf (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit courtc (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit fyre^OS (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit amigan_ (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit crwll (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit sneakums (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit fxb__ (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit Klowner (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit gibbon_ (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit sslashes (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit hachi (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit Kohlrabi (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit rgould (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit Nevtus (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit jcollie (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit hcs (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit m0f0x_ (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit pabs (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit lastebil (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit rasher (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit preglow (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit ol_schoola (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit Bagder (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit jhMikeS (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit Rob2222 (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit midkay (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit daurn (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit wrtlprnft (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit maraz (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit SirFunk (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit midgey (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit BlakeJohnson86 (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit Bjoern-Erik (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit [omni] (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit Addicted (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit kslater (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit toffe82_ (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit jurrie_ (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit x86ia (card.freenode.net irc.freenode.net) |
| 06:57:18 | | Quit |
icon identifies that the person is a core developer (has commit access).