| 00:22:41 | jc | This is sad. I've been building 2.2.12 for my application, with a kernel size of 772K. I decided, for grins, to see what 2.4.9 would come in at, with the same build. 927K. Sad sad sad. |
| 00:30:34 | Russ | probably what you decided to build in |
| 00:30:44 | Russ | 772k is pretty big for a kernel |
| 00:31:06 | Russ | 350k is about the lower limit for a usable embedded 2.4.x kernel |
| 00:31:45 | AeroSpace | 2.4 are bigger for the same settings, just like if you look at libc, they have grown also. |
| 00:32:13 | Russ | I haven't specifically found that with the kernel |
| 00:45:53 | Sammy | hi ... |
| 00:54:45 | jc | Well, I'm building what I consider a pretty limited kernel. |
| 00:55:10 | jc | 386, PPP, IDE, inet is basically it. Serial, serial console. |
| 00:55:27 | jc | Now those sizes are the image sizes, not the runtime footprint. |
| 00:56:04 | jc | No loadable modules, or anything like that. |
| 01:33:15 | AeroSpace | jc: I built one that was around 300K, I left about everything out including networking, naturally if you need it you can't exactly leave it out. What are you needing it for? |
| 01:38:10 | jc | PPP connection to an ISP. This thing has to upload data over a cellphone. |
| 01:39:43 | AeroSpace | Ok, just curious. |
| 01:40:47 | jc | Geez, looking through the symtable, the inet stuff does take a big honkin' chunk of code. |
| 01:54:47 | Aelhaz | hello from PLUG |
| 01:55:18 | Russ | P == ? |
| 01:57:06 | Aelhaz | the LUG that tim's presenting at |
| 01:57:11 | Aelhaz | finally found the wireless lan :) |
| 02:26:22 | RobFlynn | I know of PLUG. |
| 02:26:35 | RobFlynn | there were questions about gaim on the PLUG mailing list once. |
| 02:53:46 | | BZFlag was last seen on #tuxscreen 4 hours, 9 minutes and 56 seconds ago, saying: heh. plug tonight... [Wed Sep 19 23:43:50 2001] |
| 02:53:46 | vogon_jeltz | ibot, seen BZFlag |
| 02:53:51 | vogon_jeltz | hmm |
| 03:05:41 | lfj | Hello? |
| 03:07:23 | lfj2 | . |
| 03:25:06 | GoRK | whoot! fixed (read: worked around) my blitting problem in picogui |
| 03:30:12 | GoRK | whoot! got the aqua theme working. (found other bug) |
| 03:32:08 | lfj | Evening GoRK, I thought you gave up on the aqua theme because of the color palette |
| 03:33:46 | GoRK | nope turns out it was just a bug in the gradient drawing |
| 03:33:49 | GoRK | it looks quite nice |
| 03:33:53 | GoRK | about to post a screenshot |
| 03:34:15 | GoRK | after i get the kbd driver and terminal funcitoning, i'm gonig to release this bad boy and then integrate it with the build system |
| 03:34:48 | Russ | cool |
| 03:34:52 | lfj | Awsome. |
| 03:35:00 | lfj | I look forward to seeing it too. |
| 03:35:14 | Russ | I saw this girl on the way to class that had a upside down debian logo on her shirt |
| 03:35:47 | GoRK | i bet she didnt think it was an upside down debian logo :) |
| 03:35:53 | Russ | I decided against saying anything |
| 03:37:12 | GoRK | http://www.blurbco.com/~gork/tuxscreen/gork-picogui04.jpg <-- aqua |
| 03:48:05 | Aelhaz | hmm, can't reach it |
| 03:48:13 | Aelhaz | durn microsoft weenies screwing up the 'net |
| 03:48:19 | GoRK | hmm |
| 03:48:29 | GoRK | well it's on the picogui wiki page too if that helps |
| 03:48:36 | GoRK | maybe that's the wrong link |
| 03:49:36 | Aelhaz | I can't reach ebay, paypal . . . |
| 03:49:44 | GoRK | oh heh |
| 03:49:52 | Aelhaz | looks like i can't reach sourceforge |
| 03:53:11 | Aelhaz | huh, works now |
| 03:54:12 | Aelhaz | apple's gonna sue your ass :) |
| 03:57:24 | Russ | yah, I bought the tuxscreen instead of a G4 |
| 03:58:40 | jacques | or about 25 tuxscreens |
| 04:01:33 | GoRK | hey picogui runs on OS X!! |
| 04:07:25 | RobFlynn | wants to run rc5 on his tuxscreen |
| 04:07:32 | RobFlynn | eerr dnetc, I mean |
| 04:08:02 | RobFlynn | Russ: A few days ago I went into a greek restaurant... |
| 04:08:12 | RobFlynn | Russ: There was this cute girl there and there was a stuffed tux in a basket of stuffed animals. |
| 04:08:26 | RobFlynn | Russ: my friend played with the tux and she said "Put my penguin down or I will climb over this counter and beat you." |
| 04:08:40 | RobFlynn | Russ: I thought that was cool, so I asked her if she used linux. She said, "no, not really, but that is Tux over there." |
| 04:08:46 | RobFlynn | Russ: I was all happy because she knew who tux was. :-) |
| 04:08:55 | RobFlynn | I told her that she was my new best friend. |
| 04:16:22 | Russ | heh |
| 04:27:49 | cory | cool. |
| 04:49:30 | GoRK | andersee: have you perhaps integrated the busybox telnetd yet? |
| 04:49:37 | GoRK | btw hi :) |
| 04:50:00 | GoRK | i got my blitting problem worked around (at least) in picogui |
| 04:51:27 | andersee | GoRK: hi. Nope, not yet. I was going to work on that today, but ended up giving my wife a day off so I played with the kids all day. |
| 04:51:41 | GoRK | heh that's cool |
| 04:52:42 | GoRK | i dont know enough about arm vs x86 to be able to tell what exactly is wrong with these funcitons |
| 04:53:24 | andersee | whats the problem? |
| 04:55:46 | GoRK | well this one blitting function segfaults on arm and works fine on x86 and this other blitting function for some reason just doesnt work right. it seems to be getting memory addresses wrong when doing memcpy's |
| 04:56:16 | GoRK | sprites tend to blit to 32 bit boundaries and garbage gets left 16 bits at a time |
| 04:56:51 | GoRK | i originally assumed it was some kind of endianness issue, but it shouldnt be |
| 04:57:03 | andersee | Nope. Both are little endian. |
| 04:57:13 | andersee | Probably an alignment bug. |
| 04:57:28 | andersee | x86s allow all sorts of stupid alignments. |
| 04:57:56 | andersee | arm is quite strict about enforcing alignment... |
| 04:58:01 | GoRK | well there is a devine for uClinux that i set that says it fixes some alignment problems but that didnt change anything |
| 04:58:09 | GoRK | s/devine/define |
| 04:59:12 | andersee | Try running your app on the phone, then run 'cat /proc/sys/debug/alignment' |
| 04:59:23 | GoRK | hmm well i dont know anything about all that alignment stuff.. maybe someone else will fix it when i get the code out there. until then, at least i have a decent workaround. (if i dont set the blit functions in the struct it falls back to pixel-based functions) |
| 05:00:36 | GoRK | ooh a proc somethinorudder. cool ill check it out |
| 05:00:36 | andersee | If you see the following: |
| 05:00:36 | andersee | $ cat /proc/sys/debug/alignment |
| 05:00:36 | andersee | User: 0 |
| 05:00:36 | andersee | System: 0 |
| 05:00:36 | andersee | Skipped: 0 |
| 05:00:36 | andersee | Half: 0 |
| 05:00:36 | andersee | Word: 0 |
| 05:00:36 | andersee | Multi: 0 |
| 05:00:36 | andersee | Then you don't have an alignment problem... |
| 05:00:56 | andersee | If you see non-zero #s... That tells you for sure where the bug lies |
| 05:01:54 | GoRK | # cat /proc/sys/debug/alignment |
| 05:01:55 | GoRK | User: 8068 |
| 05:02:32 | GoRK | i guess that would be a definate 'yes' |
| 05:02:38 | andersee | Ha! Am I brilliant or what. |
| 05:02:49 | GoRK | dude you wrote a libc |
| 05:03:06 | andersee | Check the before and after to verify the # increases (it could be something else....) |
| 05:03:38 | andersee | libc isn't involved (except to the extent that it is being abused by picogui) |
| 05:04:22 | GoRK | yeah but i mean re: your knowledge of things such as alignment of which i cannot comprehend |
| 05:04:51 | GoRK | besides i already tested it against glibc just in case :) |
| 05:05:29 | GoRK | # cat /proc/sys/debug/alignment |
| 05:05:30 | GoRK | User: 8333 |
| 05:05:50 | GoRK | that is absolutely it |
| 05:05:58 | GoRK | which begs the question of what am i looking for to fix it |
| 05:06:41 | GoRK | some sort of offset value? |
| 05:06:46 | andersee | Lets say you have some memory. You can't just access it at any old offset on ARM. http://www.netwinder.org/~brianbr/alignment.html |
| 05:07:10 | andersee | This FAQ goes into the details of whats illegal |
| 05:07:18 | GoRK | so it is some problem with what gets passed to memcpy, right? |
| 05:07:40 | andersee | you mentioned a struct, right? |
| 05:08:10 | andersee | structs are often the cause of alignment problems.. |
| 05:08:22 | GoRK | well i dont know there. there is a struct that holds function names and capabilities of the video drivers. |
| 05:08:45 | andersee | It could be a lot of thing, really. |
| 05:08:48 | GoRK | basically i was working around the problem by not having picogui use the 'accelerated' blit function in the 8bpp framebuffer driver |
| 05:10:34 | andersee | could you run 'dmesg' on the phone and see if the offset at whch the alignent trap hit is printed? |
| 05:11:55 | GoRK | it isnt |
| 05:11:55 | GoRK | i had a page fault |
| 05:11:55 | GoRK | pgserver: unhandled page fault at pc=0x0002b9ac, lr=0x0000011b (bad address=0x0) |
| 05:11:55 | GoRK | i think that was my bad though |
| 05:13:27 | GoRK | the last time i got this low level it was writing z80 asm |
| 05:13:40 | GoRK | heh |
| 05:13:57 | andersee | There was a patch put out by np about February on arm-linux-kernel the allowed StrongArms to print the binary name and offset at which alignment problems occured... |
| 05:14:54 | andersee | I used that ~March to help a company track down an alignment problem in their closed source bluetooth stack... |
| 05:15:52 | andersee | They had spent over 3 weeks looking for the problem |
| 05:27:43 | GoRK | blah it's like i'm looking at chicken scratches :) |
| 05:28:00 | GoRK | maybe ill mail the picogui list. the author told me there are some arm users out there |
| 05:28:02 | GoRK | heh |
| 05:30:59 | GoRK | there is a dude named Eric Christianson who is working on PicoGUI with ARM that has found some alignment problems in the theme interpreter, but i cant find his email addy |
| 05:33:03 | andersee | Probably Eric Christianson <echristi@ridgerun.com> |
| 05:33:59 | GoRK | hmm what does he do? |
| 05:34:55 | andersee | dunno. |
| 05:35:00 | andersee | I asked google |
| 05:35:20 | andersee | Was answering stuff about directfb... |
| 05:39:21 | GoRK | oh |
| 05:40:31 | GoRK | yeah that's probably him. here he is talking about uclinux/arm |
| 05:42:18 | andersee | figured as much, since I know ridgerun does uClinux stuff |
| 05:42:39 | GoRK | i know nothing about the embedded world :) |
| 05:44:26 | andersee | I just know because I've been looking for work... |
| 06:40:52 | andersee | GoRK: k. See ya in a few days then |
| 13:23:19 | pattieja_ | is away: I'm busy |
| 13:23:38 | pattieja_ | is back (gone 00:00:10) |
| 13:29:16 | mallum | Hi CosmicPenguin |
| 13:30:05 | ctr | mallum, how long did you units take to arrive? |
| 13:33:19 | CosmicPenguin | mallum: hello there.... |
| 13:34:49 | CosmicPenguin | I kinda sorta integrated Microwindows into the buildroot last night..... |
| 13:35:25 | ctr | what is this buildrooot thingy? |
| 13:36:13 | CosmicPenguin | ctr: Its a bad ass script from andersee that automatically downloads and builds all the needed packages for TuxScreen |
| 13:36:27 | CosmicPenguin | basically, you type make world and out spits a kernel and a jffs2 iamge |
| 13:36:32 | CosmicPenguin | s/iamge/image/ |
| 13:37:24 | ctr | ahh cool |
| 13:39:01 | CosmicPenguin | Anyway, I can build everything except the browser... |
| 13:39:58 | CosmicPenguin | I think I'll just include the binary for that - Until we can fix the C++/uclibc problems, that will be a pain in the ass to compile.... :( |
| 13:40:27 | mallum | ctr: 10-14 days I think |
| 13:40:57 | ctr | so whats this issue with c++/uclibc? |
| 13:41:12 | ctr | mallum, did you go for the 4-7 delivery schedule? |
| 13:41:21 | CosmicPenguin | You can't compile / run c++ apps against uclibc.... |
| 13:42:11 | ctr | CosmicPenguin, how come? |
| 13:42:48 | mallum | ctr: yep I think so ... parcel force I believe |
| 13:42:55 | CosmicPenguin | [OT]: 9450 Nimda hits so far..... Thats a violent little bug.... |
| 13:42:57 | ctr | k |
| 13:43:02 | ctr | hrmm |
| 13:43:11 | ctr | ordered monday |
| 13:43:21 | ctr | so should be here this time next week hopefully |
| 13:43:35 | CosmicPenguin | ctr: It has something to do with the strongarm loader..... Andersee has it working on the m68k, so there's hope.... |
| 13:43:50 | ctr | cool |
| 14:07:06 | prpplague | can anyone here explain about the head-armv.S file where the function calls to verify architecture type? |
| 14:21:57 | CosmicPenguin | GoRK: How goes the Picogui stuff? |
| 14:55:02 | prpplague | CosmicPenguin: you still around? |
| 14:55:39 | CosmicPenguin | prpplague: I am here |
| 14:56:49 | prpplague | CosmicPenguin: had any problems with the patch-2.4.7-rmk on the sa1100fb stuff? |
| 14:57:39 | CosmicPenguin | not so far.... which specific patch are you using? |
| 14:59:01 | prpplague | CosmicPenguin:patch-2.4.7-rmk3 with diff-2.4.7-rmk3-np1 |
| 14:59:58 | prpplague | arm-linux-gcc -D__KERNEL__ -I/usr/src/arm-2.4.7/linux/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fno-common -pipe -mapcs-32 -march=armv4 -mtune=strongarm1100 -mshort-load-bytes -msoft-float -DEXPORT_SYMTAB -c sa1100fb.c |
| 14:59:59 | prpplague | sa1100fb.c:753: #error FIXME: fix pixclock, ACBsDiv |
| 14:59:59 | prpplague | sa1100fb.c:1147: #error Do we have to do this here? We already do it at init time. |
| 14:59:59 | prpplague | make[3]: *** [sa1100fb.o] Error 1 |
| 14:59:59 | prpplague | make[3]: Leaving directory `/usr/src/arm-2.4.7/linux/drivers/video' |
| 14:59:59 | CosmicPenguin | prpplague: Hmmm -- The last known working patch that I am using is rmk2-np1.... |
| 15:00:00 | prpplague | make[2]: *** [first_rule] Error 2 |
| 15:00:02 | prpplague | make[2]: Leaving directory `/usr/src/arm-2.4.7/linux/drivers/video' |
| 15:00:05 | prpplague | make[1]: *** [_subdir_video] Error 2 |
| 15:00:06 | prpplague | make[1]: Leaving directory `/usr/src/arm-2.4.7/linux/drivers' |
| 15:00:08 | prpplague | make: *** [_dir_drivers] Error 2 |
| 15:01:44 | prpplague | CosmicPenguin: i pulled them down from russell's site friday |
| 15:01:51 | CosmicPenguin | prpplague: Got it.... you have #CONFIG_SA1100_OMNIMETER selected.... |
| 15:01:57 | CosmicPenguin | Is your config screwy? |
| 15:02:03 | prpplague | maybe |
| 15:02:27 | CosmicPenguin | You need to select CONFIG_SA1100_SHANNON.... |
| 15:03:13 | prpplague | CosmicPenguin: sorry, i'm working on a omnimeter, i was just checking to see if you guys had found any problems with the fb stuff |
| 15:03:32 | prpplague | CosmicPenguin: my tuxscreen is still on order |
| 15:03:45 | CosmicPenguin | prpplague: Ahh, I see.... no problems here, but there are definately lots of #error statements throughout the omnimeter stuff.... |
| 15:04:21 | prpplague | CosmicPenguin: ok thanks, i'll just turn fb off for now, until i can get the code fixed |
| 15:04:35 | CosmicPenguin | prpplague: obviously, RMK didn't like the code that was there before |
| 15:04:59 | prpplague | CosmicPenguin: its my understanding it was done at MIT by m$ gurus |
| 15:06:31 | CosmicPenguin | MIT and M$ are not two words that go together very often... :) |
| 15:07:20 | prpplague | CosmicPenguin: true, but this was an odd project, very little documentation |
| 15:07:35 | prpplague | CosmicPenguin: i hope to change that |
| 15:09:08 | CosmicPenguin | prpplague: good luck with all of that... :) |
| 15:09:46 | prpplague | CosmicPenguin: thanks, i'll be playing with my tuxscreen soon as well |
| 15:11:26 | CosmicPenguin | prpplague: IMHO the tuxscreen is the best strongarm platform we have yet.... |
| 15:13:03 | CosmicPenguin | mallum: How goes it with the virtual keyboard rewrite? |
| 15:15:36 | mallum | CosmicPenguin: getting there ..... |
| 15:15:59 | CosmicPenguin | mallum: cool.... have you thought about adding word completion? |
| 15:16:31 | mallum | CosmicPenguin: yeah, I thought of that as a seperate app though maybe |
| 15:16:53 | mallum | CosmicPenguin: it just 'listens' to keypresses sent to the focussed window |
| 15:17:00 | mallum | CosmicPenguin: and make suggestions |
| 15:17:19 | mallum | CosmicPenguin: that way it could be used with xstroke too |
| 15:17:19 | CosmicPenguin | mallum: Just wondering... I was doing my twice yearly hard drive clean up and I found some code that I wrote up to compress the dictionary.... |
| 15:17:28 | CosmicPenguin | mallum: Thats a good idea! |
| 15:17:38 | mallum | CosmicPenguin: mail it over to me :) |
| 15:17:55 | mallum | CosmicPenguin: also it could be sensitive to the current ficussed app |
| 15:18:12 | CosmicPenguin | mallum: Ok, its been a while, so lemme clean it up a bit.... |
| 15:18:23 | mallum | CosmicPenguin: so if it had 'term' in its name, it would offer up common commands |
| 15:18:56 | mallum | CosmicPenguin: cool, with some dictionary code it shouldn't be too hard to write as I already have the xtest stuff sussed |
| 15:19:34 | CosmicPenguin | mallum: Ok... I failed you on the TS driver, but I will deliver this time... |
| 15:20:07 | mallum | heh :) dont worry about that .... |
| 15:20:14 | mallum | CosmicPenguin: you have my email addr ? |
| 15:20:53 | CosmicPenguin | I dunno... /msg it to me to make sure.... |
| 16:04:20 | prpplague | BZFlag: i sent off a money order yesterday for tuxscreen, time to get your wife boxing those tuxscreens up again... |
| 16:05:41 | BZFlag | heh, I didn't ship any this morning, cause I'm not feeling well, but I normally ship a few every day. Yesterday's orders will go out tomorrow. |
| 16:07:09 | prpplague | BZFlag: as long as i get it before christmas |
| 16:07:43 | prpplague | BZFlag: will this be your first young'n? |
| 16:14:50 | pattieja_ | is away: lunch |
| 16:17:23 | prpplague | is away: lunch |
| 16:30:51 | BZFlag | prpplague: 2nd I have a 2 yr old now (loves to play tuxracer) |
| 22:21:38 | BZFlag | jacques: happy to help. ;-) |
| 22:21:59 | jacques | BZFlag: thanks! :-D |
| 22:22:25 | jacques | BZFlag: he's just such a useful little guy |
| 22:22:27 | | It's great to be back. I missed this place! |
| 22:22:27 | BZFlag | ibot: wb |
| 22:22:39 | | jacques: bugger all, i dunno |
| 22:22:39 | jacques | ibot: where were you? |
| 22:54:33 | BZFlag | home |
| 23:20:53 | zachh | GoRK: Were you the one with the 5V TRENDnet cards? |
