| 00:29:15 | file | the Zaurus is a nice PDA |
| 00:29:23 | file | what do you run? Opie? QTopia? |
| 00:30:04 | BZFlagZaurus | qtopia at present |
| 00:30:12 | file | Opie looks nice in my opinion |
| 00:30:24 | BZFlagZaurus | agreed |
| 00:30:45 | file | made $2.7 today by activating voicemail on his cellphone |
| 00:30:49 | file | er $2.97 |
| 00:31:44 | file | FOOD |
| 00:31:49 | file | offers BZFlag food |
| 00:33:51 | file | bbl |
| 00:34:08 | BZFlag | chomps it down. |
| 00:48:04 | file | pokes BZFlag |
| 01:36:51 | Sammy | -ning |
| 07:04:20 | TomW | Well, that was a good sleep, 11 hours! |
| 07:26:07 | sjm | waves goodnight |
| 08:16:14 | mallum | hi |
| 08:16:30 | TomW | morning |
| 08:16:47 | mallum | saw episode II last night |
| 08:16:59 | TomW | of.. |
| 08:16:59 | mallum | really liked it |
| 08:17:04 | mallum | star wars ! |
| 08:17:10 | mallum | ;) |
| 08:17:28 | mallum | much better than epsisode I |
| 08:17:48 | TomW | Ok, my friend is going to see it. I usually wait until stuff is on the premium TV channel before I see them. |
| 08:18:22 | TomW | he and his family are hyped to see it. |
| 08:18:25 | mallum | yeah, some stuff is good to see at the cinema though |
| 08:19:24 | TomW | We have a really new cinema in a city nearby, it is one of those with the hi-tech screens? Best Seat In The House is wherever you sit! |
| 08:19:37 | mallum | I have a video projector and surround sound here so I normally wait for things to come out on DVD. Things come out on DVD in the states before hand and you can get them on import |
| 08:20:02 | TomW | At first, I couldn't figure out why the movie looked out of focus, until I sat down! |
| 08:22:43 | TomW | .. trying to figure out how to use dpkg, or apt-get |
| 08:22:43 | mallum | oo I havn't seen that |
| 08:24:08 | mallum | TomW: apts pretty wierd heres a summary of what I use ( and seems to be enough for me ) |
| 08:24:22 | TomW | In that theater, you don't have a set of 200W speackers blasting sounds at you, where if you sit up front your ears get blasted. The sound is distributed throughout the theater, so that everyone can hear it at the same volume. |
| 08:24:22 | mallum | apt-get install XXXX - installs package xxxx |
| 08:24:47 | mallum | apt-get source XXXX - get the source for XXXX |
| 08:25:12 | mallum | apt-cache search XXXX - search for package XXX |
| 08:25:24 | mallum | apt-cache showpkg XXXX - show info about package XXXX |
| 08:25:33 | mallum | dpkg -l - list installed packages |
| 08:25:50 | mallum | dpkg -L XXX - list files owned by package XXX |
| 08:26:06 | mallum | that about gets me through |
| 08:26:19 | mallum | it an install breaks |
| 08:26:31 | mallum | apt-get -f install ( no other params ) |
| 08:26:36 | mallum | will normally fix it |
| 08:27:08 | TomW | ah, it told me to do an apt-get update |
| 08:27:15 | mallum | and every now and then you wanna do an 'apt-get update' to get the package list up to date |
| 08:27:59 | TomW | can you get the darned thing to install the dependant files as well? or do you have to manually do that? |
| 08:28:41 | mallum | apt-get install , should just do that for you |
| 08:29:18 | mallum | 'aptitude' is a nice curses front end to it all |
| 08:33:31 | TomW | grrrr! |
| 08:33:49 | TomW | trying to install perl.. |
| 08:36:02 | LoneTech | good luck |
| 08:38:41 | mallum | TomW: perl should already be installed ? Its part of the debian base ..? |
| 08:39:44 | TomW | dpkg and apt-get are so screwy. I try to use dpkg to install the perl_5.004.05-2.deb package and it insists that perl conflicts with it. |
| 08:40:29 | TomW | ok perl is in there, but why does autoconf complain about not finding perl! |
| 08:40:41 | TomW | this is BS. |
| 08:41:02 | TomW | broken broken broken |
| 08:41:44 | TomW | ok, tantrum over, I am going for a cup of tea... |
| 08:45:45 | mallum | TomW: dont install local packages |
| 08:45:54 | mallum | TomW: get them via apt-get |
| 08:49:43 | TomW | mallum: ok, I am calmed down. Yes, I have been thinking in an RPM frame of mind. |
| 08:50:05 | LoneTech | revolutions per minute? |
| 08:51:07 | TomW | I was installing packages using dpkg-deb and that screwed things up. apt-get told me to 'apt-get -f install' (fix?), after that: I did an apt-get install gcc and it installed gcc plus its' dependancies. |
| 08:51:47 | TomW | mallum: so, I just compiled a "hello world" and runs fine. |
| 08:51:53 | mallum | cool |
| 08:51:55 | TomW | mallum: bbl, tea is on. |
| 08:52:01 | mallum | apt does take a little while is get used too |
| 08:52:06 | mallum | can I have a cup too ? |
| 08:52:36 | mallum | I think apt may be one of them apps that has been 'debugged' into existance |
| 08:52:48 | TomW | mallum: it is only Black Tea, but come on by. |
| 08:52:52 | mallum | where the opotions and stuff dont seem that logical |
| 08:53:02 | mallum | earl grey ? |
| 08:53:21 | TomW | mallum: last of that went this morning. I have to get more. |
| 08:53:51 | TomW | I still have some djarlening (sp?), but that isn't a morning tea. |
| 08:57:16 | mallum | yes I like that too |
| 09:03:19 | TomW | smoking my pipe and having a cup of hot tea, how good it is... |
| 09:07:38 | TomW | mallum: now I understand why the RPMs disappeared off of uk.arm.linux.org, the debian + arm combination is pretty powerful! |
| 09:30:04 | TomW | mallum: heh, midnight commander running on the Tux. won't run from the LCD, but it works via the serail console. |
| 09:35:11 | mallum | install aptitude |
| 09:35:27 | mallum | it'll give you a front end to all the packages you can install |
| 09:36:49 | TomW | mallum: ah!! THANKS!! |
| 09:39:11 | mallum | TomW: there a program called dselect too , but its _very_ wierd |
| 09:42:56 | ctr | yes it is |
| 09:43:02 | ctr | apt is your friend |
| 09:43:06 | ctr | :) |
| 10:36:00 | TomW | mallum: can you select, then install packages with aptitude, or is it just a package info tool? |
| 11:03:20 | mallum | TomW: yep, in stall with it too, afaik |
| 14:04:42 | TomW | this doens't look right! debian libc6 includes define O_LARGEFILE as 0 while my redhat & mandrake machines define it as 0100000 |
| 14:06:31 | TomW | the tux kernel headers define it as: 0400000, whom am I to belive? |
| 14:10:54 | LoneTech | asm/fcntl.h:#define O_LARGEFILE 0100000 |
| 14:10:54 | LoneTech | bits/fcntl.h:# define O_LARGEFILE 0100000 |
| 14:11:07 | LoneTech | Debian GNU/Linux x86 |
| 14:11:39 | CosmicPenguin | morning! |
| 14:11:58 | LoneTech | afternoon! |
| 14:12:12 | TomW | CosmicPenguin: morning |
| 14:12:20 | CosmicPenguin | LoneTech: its a small world on IRC, huh? |
| 14:12:24 | LoneTech | yep |
| 14:12:30 | TomW | CosmicPenguin: up to a question about a confict in kernel headers? |
| 14:12:38 | CosmicPenguin | TomW: I'll do my best |
| 14:13:46 | TomW | k, the asm/fcntl.h defines 0_LARGEFILES as 0100000, /usr/include/bits/fcntl.h on debian-arm defines it as 0, who do I believe? |
| 14:14:00 | TomW | I am compiling the reiserfs tools. |
| 14:14:35 | CosmicPenguin | My skiff toolchain has this: |
| 14:14:37 | TomW | CosmicPenguin: do I go with the libc6 as this app is not a "kernel" app, technically speaking. |
| 14:14:40 | CosmicPenguin | #ifdef __USE_LARGEFILE64 |
| 14:14:40 | CosmicPenguin | # define O_LARGEFILE 0 |
| 14:14:40 | CosmicPenguin | #endif |
| 14:15:03 | TomW | CosmicPenguin: so, you would go with the libc6 def over the kernel then? |
| 14:16:10 | CosmicPenguin | TomW: lemme check my 2.95.3 distro really quickly |
| 14:18:31 | CosmicPenguin | Ok, so this is weird - in my 2.95.3 toolchain, the value of O_LARGEFILE is now 0400000, but it is still contained by the #ifdef |
| 14:19:19 | TomW | my tux kernel sources do not qualify the #def |
| 14:19:32 | CosmicPenguin | I don't have an up to date RMK kernel, but the vanilla kernel reports O_LARGEFILE to be 0400000 in include/asm-arm/fcntl.h too.... |
| 14:20:02 | CosmicPenguin | TomW: I would go with whatever the kernel says.... |
| 14:20:23 | TomW | ok, it seems that that def only applies to files > 2 gig in size, I'll ignore it for now on the Tux, I only have a 10G in it. |
| 14:20:34 | TomW | CosmicPenguin: thanks. |
| 14:21:01 | TomW | CosmicPenguin: just didn't want to experiment with stuff when building the reiser tools! |
| 14:21:02 | CosmicPenguin | TomW: sorry I couldn't be of more help - I really hate the how the system include files are so hard to follow |
| 14:21:17 | CosmicPenguin | TomW: At least you're not playing with a critical system... :) |
| 14:21:41 | TomW | yup, heh, can replace the contents of the harddrive in minutes via nfs. |
| 14:21:41 | file | TomW: How's your hardrive? |
| 14:21:50 | TomW | file: works fine. |
| 14:21:57 | file | TomW: Good, good |
| 14:22:03 | CosmicPenguin | Thats such a cool hack, by the way... :) |
| 14:22:34 | TomW | file: I have debian-arm installed on it with the 2.4 tux kernel. I am running total debian apps right now and compiling the reiser tools on it now. |
| 14:22:48 | TomW | CosmicPenguin: thanks, makes the Tux so much more useful. |
| 14:23:03 | file | marvelous |
| 14:23:33 | TomW | nice to do a 'set -o vi' at the command prompt and have it work! |
| 14:24:34 | TomW | transfer rate on the IDE is only 1.7MB/s, but then, I will look later to see why 'hdparm -P 2 /dev/hda' doesn't work. |
| 14:25:21 | TomW | goes back to coding. |
| 14:31:08 | file | prpplague! |
| 14:32:59 | | Good morning! |
| 14:32:59 | prpplague | file: morning? |
| 14:33:08 | | prpplague: what? |
| 14:33:08 | prpplague | ibot: drop dead |
| 14:33:18 | file | kinda morning |
| 14:44:57 | prpplague | anything fun happening? |
| 14:56:52 | TomW | CosmicPenguin: uhoh, the reiserfs tools compiled ok on the Tux. I replaced the debian-arm /usr/include/asm & /usr/include/linux headers with those of the Tux kernel. |
| 14:57:09 | TomW | Now, to see how badly they mangle the drive! heh. |
| 15:12:33 | CosmicPenguin | TomW: As long as the kernel and the apps are on the same page, you should be fine |
| 15:13:19 | TomW | heh, yeah, well. Debian has a 2.2 kernel and just thought it prudent to change the kernel headers.. |
| 15:22:20 | TomW | hmm, no joy: |
| 15:22:23 | TomW | Tux:~# mkreiserfs /dev/hda1 |
| 15:22:23 | TomW | |
| 15:22:23 | TomW | <-------------mkreiserfs, 2001-------------> |
| 15:22:23 | TomW | reiserfsprogs 3.x.0j-esafe |
| 15:22:23 | TomW | [155] mkreiserfs: obsolete system call 000000c3. |
| 15:22:23 | TomW | Code: e1a03001 e1a00002 e1a01003 ef0000c3 (e1a04000) |
| 15:22:25 | TomW | Illegal instruction |
| 17:05:02 | Russ|werk | TomW: oops |
| 17:07:16 | TomW | yeah. debian uses 2.2 kernel, although they have a libc6, I wonder if it is a downgraded version to work with 2.2 and 2.4? I just found something that don't work. |
| 19:12:01 | as_ | Evnin all. |
| 19:12:11 | prpplague_food | as_:evening |
| 19:12:56 | file | prpplague_food? any progress? |
| 19:15:06 | as_ | Hes probably prplague_digesting by now. |
| 19:15:19 | as_ | Which is progress, of a sort. |
| 19:21:04 | as_ | :] |
| 19:23:44 | | somebody said prpplague was mailto:dave123@abcsinc.com |
| 19:23:44 | file | prpplague!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! |
| 19:23:46 | | somebody said prpplague was mailto:dave123@abcsinc.com |
| 19:23:46 | file | prpplague!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! |
| 19:23:46 | | prpplague is probably mailto:dave123@abcsinc.com |
| 19:23:46 | file | prpplague!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! |
| 19:28:30 | lyakh | Hi! TomW, you're there? |
| 19:31:13 | TomW | lyakh: you here, or working? |
| 19:31:45 | TomW | lyakh_working: or, you over here? |
| 19:31:50 | lyakh | here:-) half past 21 - too late to work:-) |
| 19:31:55 | TomW | lyakh: heh |
| 19:32:13 | lyakh | Was leaving work to fast to change lyakh_working to lyakh_gone:-) |
| 19:33:01 | lyakh | I just sent some micro-suggestions to your assembly function to the list and - I believe - to you. Have a look, but maybe just rubbish:-) |
| 19:33:02 | TomW | I tried to compile the reiserfs tools, natively, it worked. I had replaced the debian /usr/include/{linux,asm} contents with that of my 2.4 kernel |
| 19:33:29 | TomW | kaboom! |
| 19:33:58 | lyakh | Normally (for user-space apps) you shouldn't do that (replace headers). They should match libc, not your running kernel. |
| 19:34:42 | TomW | yes, but if they ask for <asm/blah.h> or <linux/blah.h>, then you would have to. |
| 19:34:59 | lyakh | You only need kernel-headers when compiling modules, maybe some other kernel-critical stuff:-) (like cardmgr, etc.), but not sure. |
| 19:35:19 | TomW | mmmm, no.. |
| 19:35:21 | lyakh | Why? There should be same blah.h's there coming with glibc. |
| 19:36:03 | TomW | right, no prob if they do <blah.h>, but if they direct the include into the kernel headers, then I would want the headers to match the running kernel. |
| 19:36:22 | TomW | nicht war? |
| 19:36:23 | lyakh | yes.:-) Sure. For user-space apps - use same headers, against which your libc was compiled - and this is what's supplied with distros. |
| 19:36:46 | TomW | AHH! Ok. |
| 19:36:54 | lyakh | What do you mean? "direct the include into the kernel headers"? |
| 19:37:04 | TomW | Even if they do syscall()s? |
| 19:37:27 | TomW | yes, <adm/blah.h> would go into the kernel headers. |
| 19:37:30 | lyakh | All progs do that:-) (well, _practically_ all). |
| 19:37:35 | TomW | s/adm/asm/ |
| 19:38:20 | lyakh | No. /usr/include/asm, linux - this is what should be used. Otherwise why is it there?:-) |
| 19:39:26 | TomW | Ok, since glibc was compiled against the older headers, I need my app to use the same bit flags etc, even if they are not currently defined that way in the latest kernel headers. The glibc is important as it only knows &-ing & |-ing masks as the original headers had it? |
| 19:41:37 | lyakh | Yeah, something like that:-) Not sure I understood you quite right, but I am sure for most "normal" compiles provided headers should be used. |
| 19:42:07 | lyakh | (was it English or German or...?:-)) |
| 19:43:47 | prpplague | file: man, i think you need a chill-pill |
| 19:43:56 | TomW | well, it blew up the first time sawing "outdated syscall()" , or similar, so it actually sounded like the libc6 was making syscalls in a version 2.2 manner, and the 2.4 kernel did not like it. |
| 19:45:01 | TomW | english, the error message came from arch/asm/kernel/trap.c of the kernel |
| 19:45:38 | lyakh | ok, then try the kernel headers:-)) I am still sure I was right - as a general rule, but sometimes it doesn't work:-) |
| 19:46:12 | lyakh | Did you get my email? |
| 19:49:16 | TomW | lyakh: yeah, no data loss. |
| 19:49:36 | lyakh | good:-) what do you think? |
| 19:51:40 | TomW | It runs ok, I was looking to tighten up the code some more. I don't know ARM assembler, so I was sort of guessing at the ops and looking at what others had written. I was hoping to do something like: |
| 19:52:00 | TomW | ldrb r1, [r0], r1, #lsl 8 |
| 19:52:04 | TomW | or something. |
| 19:52:19 | TomW | anything to loose a few clock cycles! |
| 19:53:41 | TomW | if I adjust the timing of the CS3, by one cound of the MSC1 setting, it makes a marked difference in thruput. maybe a couple of percent. |
| 19:53:52 | TomW | s/cound/count |
| 19:54:50 | DarkDream | Hi :) |
| 19:54:52 | lyakh | Don't see how you can squeeze the code any more (but I am no ARM-asm guru myself), so, do you think putting 2 reads from the drives back to back is ok? |
| 19:54:57 | DarkDream | Just a little question ;) |
| 19:55:14 | DarkDream | Is someone has tested Gentoo linux ? |
| 19:55:30 | lyakh | Also, how is MECR set on Tux?... |
| 19:55:58 | TomW | MECR? let me look. |
| 19:57:02 | russ|werk | its in blob |
| 19:57:14 | russ|werk | but you can set it in the kernel as well |
| 19:57:32 | TomW | lyakh: it doesn't look like they set it on the tux, it is set for the SA1110, but not for the SA1100. |
| 19:57:34 | Russ|werk | cs3 is most likely at the slowest setting by default |
| 19:57:54 | lyakh | not _can_, _should_, I believe... |
| 19:58:18 | TomW | Russ: CS3 is set at a 600ns cycle, the drive is running in Pio mode 0 right now. |
| 19:59:16 | TomW | Russ: I would like to get it up to PIO 2, but hdparm complains when I try to set that mode. |
| 19:59:46 | TomW | Russ: PIO 3 & 4 are not possible with my hardware (8 bit). |
| 20:00:34 | TomW | hmmm, maybe I can play around with the Read Recovery timing? |
| 20:01:00 | TomW | I don't think that Read Next is involved here.. |
| 20:01:33 | TomW | when I drop Read first below 600ns, the drive starts to disappear. |
| 20:04:53 | TomW | lyakh: /* configure the CS3 chip select timings. */ |
| 20:04:53 | TomW | MSC1 = (MSC1 & 0xffff) | \ |
| 20:04:53 | TomW | ((MSC_NonBrst | MSC_16BitStMem | MSC_RdAcc(31) | \ |
| 20:04:53 | TomW | MSC_NxtRdAcc(31) | MSC_Rec(3)) << 16); |
| 20:06:17 | TomW | lyakh: I tweaked the MSC_RdAcc() down, but it I go to 30, then every other boot, the drive is not found. |
| 20:08:26 | lyakh | Sorry, not my field:-) I just remember we tried tweaking MECR when we had problems with our flash disk. |
| 20:09:23 | lyakh | BTW, we do still from time to time get "unable to mount the route filesystem, unrecognised partition table" stuff, but then after a few recompiles it magically disappears... |
| 20:09:39 | lyakh | s/route/root/ |
| 20:10:29 | TomW | lyakh: the interface can only run so fast. AFAICT, I have to remove cycles from the ARM asm loop, or get the drive into PIO mode 2 so I can drop the CS3 cycle timing down to 240ns from 600ns. |
| 20:11:28 | lyakh | hm, well, you can compute that - you know from the table how many cpu-cycles every asm-opcode takes...:-) |
| 20:11:40 | TomW | lyakh: I am on some new ground here myself, this is the deepest that I have gone into IDE drive technology! |
| 20:12:14 | TomW | yeah, yeah: x number of S cycles with some amount of N cycles. :P |
| 20:12:45 | TomW | ...only if the moon is blue this evening, you don't need an N cycles. |
| 20:12:54 | TomW | s/an/any/ |
| 20:14:33 | TomW | that ARM assembly PDF from arm.org is filled with a lot of gibberish. And I know assembler, been reading code for years, and working with processors down to the microcode level. IMO, they deliberatly obfucsicate the ARM7 manual to make you pay money for the "real" one! |
| 20:15:30 | TomW | lyakh: ok, so I am upset, sorry, it IS a lousy manual. |
| 20:16:45 | lyakh | no prob, I didn't write it:-) but, I guess, I almost got used to it (DDI0100E_ARM_ARM.pdf?) |
| 20:19:19 | TomW | reiserfsprogs 3.x.0j-esafe |
| 20:19:19 | TomW | [154] mkreiserfs: obsolete system call 000000c3. |
| 20:19:19 | TomW | Code: e1a03001 e1a00002 e1a01003 ef0000c3 (e1a04000) |
| 20:19:19 | TomW | Illegal instruction |
| 20:20:09 | TomW | same thing with both the debian headers and my own, the prob is in the libc6 library, it is making an obsolete call, I will have to get the source to the libc6 and track it down. |
| 20:24:58 | lyakh | ok, best of luck |
| 20:25:41 | as3 | Tom, not sure what you mean by ldrb r1, [r0], r1, #lsl 8. |
| 20:26:13 | as3 | Should be able to do ldrb r1,[r0 + r1 lsl #8]. |