02:02:48 | Russ|werk | hey Sammy |
02:03:05 | Sammy | Yo Russ |
03:09:54 | | erikm was last seen on #blob 3 days, 14 hours, 29 minutes and 56 seconds ago, saying: bye Sammy [Fri Jan 25 12:39:58 2002] |
03:09:54 | BZFlag | ibot: seen erikm |
10:49:32 | seletz | hi all |
10:51:04 | erikm | hello seletz |
11:01:56 | akira|wor | hoi |
11:33:15 | seletz | akira|wor: yup |
11:35:12 | Sammy | why configure can't know my target board in 2.0.5-pre2 when I add puppy into configure.in ? |
11:35:31 | Sammy | something I miss ? |
11:35:41 | erikm | run tools/rebuild-gcc twice |
11:36:17 | Sammy | Oh... really something I miss , thanx erikm |
11:38:30 | erikm | BZFlag: are you there? |
12:02:46 | BZFlag | erikm: for a sec, headed to bed. |
12:02:49 | BZFlag | sup? |
12:03:05 | erikm | ok, I'll followup by mail. |
12:03:22 | BZFlag | hit me if it's quick |
12:03:23 | erikm | I just wanted to paste the last five lines of "od -x blob" |
12:03:34 | BZFlag | not all nulls? |
12:03:37 | erikm | no |
12:03:49 | erikm | 0060040 7003 32d2 2314 c020 7004 32d2 232c c020 |
12:03:49 | erikm | 0060060 7005 32d2 234c c020 |
12:03:49 | erikm | 0060070 |
12:03:50 | BZFlag | ok. hmm.. tried with all the options I had on? |
12:03:54 | erikm | (last three lines) |
12:03:56 | erikm | not yet |
12:04:00 | erikm | gimme a sec... |
12:04:40 | BZFlag | 00007E40 74 69 78 45 44 4B 20 C0 0A 00 00 00 01 70 D2 32 tixEDK ......p.2 |
12:04:40 | BZFlag | 00007E50 6C 20 20 C0 02 70 D2 32 90 20 20 C0 03 70 D2 32 l ..p.2. ..p.2 |
12:04:40 | BZFlag | 00007E60 AC 20 20 C0 04 70 D2 32 C4 20 20 C0 05 70 D2 32 . ..p.2. ..p.2 |
12:04:41 | BZFlag | 00007E70 E4 20 20 C0 00 00 00 00 00 00 00 00 00 00 00 00 . ............. |
12:05:10 | erikm | ok, recompiled |
12:05:13 | erikm | 0101040 6974 496e 492c c020 0028 0000 6974 4578 |
12:05:13 | erikm | 0101060 4e64 c020 000a 0000 7001 32d2 22d4 c020 |
12:05:13 | erikm | 0101100 7002 32d2 22f8 c020 7003 32d2 2314 c020 |
12:05:14 | erikm | 0101120 7004 32d2 232c c020 7005 32d2 234c c020 |
12:05:14 | erikm | 0101140 |
12:05:14 | BZFlag | and then the 00's continue for the next 13k |
12:05:25 | BZFlag | hmmm |
12:05:54 | BZFlag | guess I need to objdump or have the linker create a map file. |
12:06:19 | BZFlag | why does your compile stop again? some sa1111 symbol, right? |
12:06:27 | erikm | I'll add that to the Makefile.am, that could come in handy |
12:06:49 | BZFlag | add the map file? yeah, that's probably a good thing. |
12:07:08 | erikm | *nod* |
12:07:17 | erikm | yeah, some sa1111 symbol |
12:07:30 | erikm | I'm compiling against 2.4.17-rmk3 |
12:07:31 | BZFlag | cool. I cvs up tomorrow and see what the map file shows. |
12:07:42 | erikm | ok, have a good night |
12:07:45 | seletz | erikm: hmmm, why does the sa1111 not compile, strange. I thought that i fixed it |
12:08:00 | seletz | ugh |
12:08:13 | BZFlag | I'm on various... 2.4.16-rmk1 on the idr at the moment. |
12:08:16 | seletz | fears another change in SA-1111.h |
12:08:28 | erikm | seletz: /home/erik/src/sourceforge/blob/src/diag/system3.c:97: `SBI_SKCR' undeclared (first use in this function) |
12:08:42 | BZFlag | we should think/talk about copying all the kernel bits over. |
12:08:51 | erikm | BZFlag: I was thinking the same :) |
12:08:55 | seletz | BZFlag: hmmm. |
12:09:17 | seletz | Probably with a script for syncing, no? |
12:09:30 | erikm | seletz: that wouldn't really solve the problems |
12:09:32 | BZFlag | I think it's a good thing, but I also think the build should ask for an optional kernel tree and verify that things are at least close on bits like param blocks etc. |
12:09:56 | erikm | BZFlag: we could also provide a set of known good kernel header files |
12:10:00 | seletz | hmmm-hmm. Two sources for the same datea is asking for trouble. |
12:10:10 | erikm | BZFlag: as a separate CVS thingy, for example |
12:10:31 | BZFlag | perhaps a fallback include directory that one needs to give an option to use. |
12:11:04 | BZFlag | one more ? while you are here.. |
12:11:09 | erikm | ok |
12:11:09 | seletz | We just have to make sure that the symbols are named the same. |
12:11:32 | BZFlag | what's the plan for > 4M jffs2 images. should "erase ramdisk" erase all of the rootfs flash? |
12:11:45 | BZFlag | it should probably not load it all as a ramdisk. |
12:12:02 | BZFlag | I've got 16M ram and 16M flash at the moment. |
12:12:13 | BZFlag | I want to do blob+jffs2 and nothing else. |
12:12:38 | erikm | BZFlag: IIRC erase ramdisk should erase the rootfs flash |
12:12:56 | BZFlag | defining ramdisk size as 15M (ish) does not boot. |
12:13:01 | erikm | BZFlag: the partitioning code I'm working on will solve that |
12:13:20 | BZFlag | needs a param block? |
12:13:42 | erikm | no, I just want to include a default partition table |
12:14:04 | BZFlag | then what says not to load it as an initrd? |
12:14:27 | BZFlag | and how to handle uploads of the rootfs if there is not enough ram? (ugh) |
12:14:31 | erikm | that's dependent on the system (so it comes in src/blob/<arch>.c |
12:15:09 | erikm | first thought: split it in parts |
12:15:22 | BZFlag | yeah, but what would the commands be? |
12:15:37 | erikm | haven't thought about it yet :) |
12:16:19 | erikm | you could make it a two stage approach: first download a minimal part in blob, boot the kernel, and do the rest in linux |
12:16:19 | BZFlag | I wan thinking a memory range on *download as in "xdownload 0x00020000-0x000c0000" |
12:17:02 | BZFlag | which would use the larges ram block (in this case the ramdisk space) |
12:17:52 | erikm | we need some kind of memory manager |
12:17:57 | BZFlag | I will likely not need it for this board, as I should have 2x or 4x ram to flash on the shipping model, it's just the devel board I have here. |
12:18:49 | BZFlag | I would like to have backup routines though so I can save all of flash to a file on the host. |
12:19:23 | BZFlag | think about ide/cf updates too as I will likely merge that in for the vercel. |
12:19:32 | seletz | BZFlag: Yeah! I like that too. |
12:19:44 | BZFlag | duck has that working on his board. |
12:20:00 | BZFlag | seletz: which, backup or ide or both? ;-) |
12:20:10 | erikm | BZFlag: if he could make that a separate patch I'd like to integrate it |
12:20:31 | BZFlag | I'll ask him, but baring that, I'll create one. |
12:20:38 | erikm | BZFlag: (separate as in separated from the port to non-SA) |
12:20:47 | BZFlag | right. agreed. |
12:20:58 | seletz | BZFlag: the idea of mine was: Have a specially marked CF card. Plug it in: Boot. Blob detects: Its an update CF card, ok, lets copy all to the internal flash. Bingo. |
12:21:22 | erikm | seletz: Russ' param blocks would make that possible |
12:21:33 | seletz | BZFlag: Send those OS updates via CF card to customers. |
12:21:35 | BZFlag | seletz: for devices with some kind of key/switches/etc I'd like to also require holding down a button. |
12:21:50 | erikm | seletz: (and doing a default parameter block is easy as well) |
12:21:51 | seletz | BZFlag: cool too. |
12:22:15 | erikm | BZFlag: can also be done with Russ' parameter block stuff |
12:22:24 | seletz | erikm: i thought of a magic sig somewhere in a CF card. |
12:22:47 | seletz | erikm: the paramblock could then contain a cmd call to check this sig. |
12:23:00 | BZFlag | agreed. I want a default param block though to still allow blob+jffs2 as the only partitions. |
12:23:11 | erikm | seletz: the magic sig stuff should be implemented in a param block. I don't want that stuff in LART, for example |
12:23:29 | BZFlag | nods |
12:23:43 | seletz | erikm: ok, i'm not into paramblocks that far yet, but as long it works... |
12:24:04 | BZFlag | though it could be a --enable thing too. |
12:24:27 | erikm | seletz: it allows for a lot of flexibility. I don't want to make the mistakes bootloader made |
12:24:41 | erikm | s/bootloader/bootldr/ |
12:24:49 | seletz | Once we cleaned up diag/blob cmds to a cmd lib, the "blob size" issue will be solved, too. |
12:25:06 | seletz | erikm: agreed. |
12:25:07 | BZFlag | heh, well on that note... bootldr has eliminated parameter partitions as the norm. ;-0 |
12:25:42 | erikm | BZFlag: yeah, I've seen that. I still want paramter partition on LART becuase it has boot block flash |
12:25:49 | BZFlag | well. great stuff happening. look for some patches from me this week. |
12:25:55 | erikm | BZFlag: ok |
12:26:04 | BZFlag | gotta sleep. nite all! |
12:26:09 | erikm | BZFlag: night |
12:26:15 | seletz | BZFlag: good night.... |
12:27:08 | erikm | seletz: for example moving the commands in src/blob/debug.c to src/cmds/debug.c and linking against a libcmds.a? |
12:28:01 | seletz | erikm: yup. And creating a script which creates a commands.c for blob and diag based on configuration. This command.c has the __commandlist() stuff in. |
12:28:35 | seletz | erikm: so everyone can have its own blob: fulle blown-up fat, or tiny-weeny small. |
12:29:14 | erikm | for example: ./configure --with-commands=chgbit,reboot,reblob |
12:29:24 | seletz | erikm: yup. |
12:29:25 | erikm | or: ./configure --with-commands=all |
12:29:34 | seletz | erikm: yup |
12:30:10 | erikm | seletz: in that case you should also put each command in a separate file, but that's another issue |
12:30:12 | seletz | I have that in mind since our last talk 'bout it, but i first have to solve my PCMCIA/CF kernel problems. Grrr. |
12:30:21 | erikm | :( |
12:30:38 | seletz | erikm: yup, one command per file, or closely related commands, that is. |
12:31:29 | seletz | erikm: bitman.c, memchk.c, pcmcia.c, ide.c lcd.c and so on. |
12:31:44 | seletz | erikm: or even more broken up |
12:32:20 | erikm | seletz: no, one command per file. for example: foobar.c contains the commands foo and bar. if the user only wants the foo command, the code for bar still gets included simply because the linker can only link per object file, not per function |
12:32:44 | seletz | erikm: yes, thats exactly what i want. |
12:32:55 | erikm | ok, then we agree |
12:33:41 | seletz | erikm: I'll get that one working, as soon as I have my kernel problems solved. |
12:34:03 | seletz | erikm: Its not too complicated, i think. |
12:34:23 | akira|work | <- offline |
12:34:57 | erikm | I'll put the necessary changes in CVS |
12:35:08 | seletz | uh? |
12:35:48 | erikm | as in: make src/cmds/ |
12:36:07 | erikm | and make blob and diag link against that |
12:36:54 | seletz | ok, fine. |
12:37:00 | erikm | http://www.funnybunch.com/fun/chili.htm <-- lol :) |
12:38:21 | seletz | looking ... |
12:38:53 | seletz | lol |
12:40:11 | Sammy | Amm... erikm: is that doc/porting.txt need to be update ? |
12:41:44 | Sammy | because when compiler some error occur in diag part |
12:43:51 | Sammy | nothing big deal , just because like me , add a new platform , also need to add some stuff in diag part |
12:44:44 | Sammy | but in porting.txt don't wirte that , so I think maybe should add that on it |
12:52:16 | seletz | debugging |
12:52:16 | seletz | hmmm |
12:52:16 | seletz | network hiccups |
12:53:13 | erikm | Sammy: yes, it needs some info on serial driver selection, but that's only minor |
12:53:14 | erikm | Sammy: just look at how the other architectures do it |
12:55:16 | erikm | yeah, the uni's router over here went berzerk |
12:57:20 | seletz | give him tranquilizers ... |
12:57:39 | erikm | I think they already did |
12:57:47 | Sammy | what's berzerk ? |
12:58:02 | erikm | go completely crazy |
12:59:08 | erikm | this is what gdict says: |
12:59:30 | erikm | Berserk Ber"serk, Berserker Ber"serk*er |
12:59:32 | erikm | (Scand. Myth.) One of a class of legendary heroes, who |
12:59:32 | erikm | fought frenzied by intoxicating liquors, and naked, |
12:59:32 | erikm | regardless of wounds. --Longfellow. |
12:59:42 | erikm | 2. One who fights as if frenzied, like a Berserker. |
13:00:27 | seletz | man, ide.c is _horror_ |
13:00:31 | Sammy | ok , write it :) |
13:00:33 | erikm | *nod* |
13:01:03 | seletz | This_is_tha_func_name_ind_its_long() |
13:01:07 | seletz | argh |
13:01:27 | erikm | seletz: IMHO it needs a real cleanup. rmk wanted to do that, but he scared from the details |
13:02:14 | seletz | sigh. |
13:03:00 | seletz | I only wand a CF -ide card. damned. I seek some point where i can dump out th last cmd before my board goes, erm, berzerk. |
13:03:20 | seletz | afk, bbl |
13:05:37 | Sammy | I am getting berzerk too , lost 5 CD from CD burner, |
13:05:51 | erikm | heh :) |
13:05:56 | Sammy | maybe I need new one ....think |
13:06:17 | erikm | or better software |
13:06:52 | erikm | I got my first coaster with a CD burner last week when I tried a HP USB CD writer |
13:07:17 | Sammy | but I use nero 5.5.16 |
13:07:52 | erikm | I always use xcdroast, cdrecord, or cdrdao (all under linux) |
13:10:06 | Sammy | don't hit me , and if any one see this still don't hit me too , in here I still use Wx2xxx to be my platform , |
13:10:30 | seletz | back |
13:10:44 | Sammy | but I must say , my workstation use RedHat 6.22 with cle 0.9 |
13:11:09 | erikm | Sammy: I won't hit you. if you feel comfortable with it, or if it's company policy, go ahead |
13:11:36 | erikm | Sammy: (you also might want to look at vmware, btw. it allows you to run win2k within linux) |
13:12:07 | seletz | Sammy: yup, actually windoze makes up a goot client os. Just install cygwin to have a real shell (it saves much work). |
13:12:50 | Sammy | no because the Wx2x support more S/W for chinese |
13:13:16 | erikm | I thought turbolinux had good asian language support |
13:13:44 | seletz | SUSE will go on the asian market, too. |
13:13:53 | Sammy | but S/W ? |
13:14:04 | | SW is probably shock wave |
13:14:04 | seletz | WHAT SW? |
13:14:08 | erikm | Sammy: like what? |
13:14:10 | seletz | Word processing? |
13:14:15 | seletz | Get TeX |
13:14:19 | erikm | lol :) |
13:14:32 | seletz | I mean it. |
13:14:36 | Sammy | Office , |
13:14:53 | Sammy | because , our office lady use that , |
13:15:03 | erikm | don't know if StarOffice/OpenOffice has good I18N support |
13:15:09 | seletz | Ugh. |
13:15:26 | seletz | Create PDF formatted docs. |
13:15:45 | erikm | I must say that I can handle the word and excel documents I get with abiword and gnumeric |
13:16:37 | Sammy | so we need to tie in with her to go out doc ... |
13:16:50 | Sammy | I have try Koffice , |
13:16:54 | seletz | well, i know this too good. "Linux is notcompatibele because it cant run office". Do a s/Linux/Mac/ and you'll get the same. Sigh. |
13:17:29 | seletz | forces everyone to send plain ASCII or PDF |
13:17:38 | erikm | tomorrow we'll get a change to get the uni out of the MS claws. |
13:18:12 | seletz | erikm: really? Got a big "license costs" argument i guess? |
13:18:23 | erikm | yeah. the yearly MS tax got quadrupled, so I proposed to switch to linux/staroffice |
13:18:25 | seletz | erikm: NT-> M$2000 |
13:18:32 | Sammy | but this doesn't matter , because this is for company use , in my home I use RH 7.0 + cle1.0 and feel good too |
13:19:00 | seletz | Sammy: yeah, keep it that way. Saves trouble. |
13:19:35 | Sammy | yeah , |
13:19:55 | erikm | seletz: there are a couple of groups over here that completely run with linux/staroffice and they can do with only a single sysadmin for 50 people |
13:20:25 | seletz | erikm: Yup. And no license fees from M$. |
13:20:32 | erikm | seletz: exactly |
13:20:32 | Sammy | but you know, in here still have lot of people can use computer but don't know what's PDF |
13:20:33 | seletz | erikm: Save money. |
13:20:58 | erikm | seletz: well, they only three MS licenses to run windows in VMware |
13:21:19 | Sammy | so poor ... |
13:21:24 | seletz | erikm: Oh, really? Got VMWare Server (whats that thingy actually called?) |
13:21:49 | erikm | seletz: I don't know exactly how the sysadmin solved it |
13:21:51 | seletz | erikm: VMware got horribly expensive |
13:22:20 | Sammy | yeap that's right |
13:23:02 | Sammy | so I have two computer her , and I use telnet , and it's free |
13:23:14 | Sammy | and save $ again |
13:24:44 | Sammy | but tell you all guys one more things , if you have free $ , still don't try use WxXP , |
13:25:13 | erikm | looks for the link... |
13:25:37 | erikm | here it is: http://www.theregister.co.uk/content/4/23830.html |
13:25:38 | Sammy | because , some of my colleague use that |
13:26:05 | Sammy | first made him into memory trouble ... |
13:26:48 | erikm | (FYI: teletubbies are a popular TV program for little children) |
13:27:16 | Sammy | and after add 266MB , now get into second server connect trouble |
13:27:21 | Sammy | er 256 |
13:29:01 | erikm | hmm, blob-2.0.5-pre2 already got 71 times downloaded from sourceforge. good. |
13:29:23 | erikm | (2.0.5-pre1 got downloaded only 110 times in its whole lifetime) |
13:34:55 | Sammy | see this http://www.atnzone.com/moviezone/reviews2001/antitrust.html |
13:35:07 | Sammy | http://www.preview-online.com/jan_feb2001/feature_articles/antitrust/ |
13:35:24 | seletz | gives pcmcia/cf another try |
13:37:11 | Sammy | time to go home |
13:37:16 | Sammy | night all |
13:37:41 | seletz | Sammy: bye ... |
13:38:09 | Sammy | bye seletz |
13:56:43 | seletz | erikm: do you know the reason behind the sa1100_mask/unmask_*_irq() function magic? |
13:57:02 | erikm | seletz: no, not really. sorry |
13:57:09 | seletz | in arch/arm/mach-sa1100/irq.c |
13:57:14 | seletz | ok, np. |
15:27:32 | seletz | prpplague: hi |
15:36:45 | prpplague | seletz: howdy |
15:36:50 | prpplague | seletz: hows biz? |
15:38:10 | erikm | seletz: regarding the CDB89712 question on linux-arm-kernel: duck is working on a non-StrongARM blob port. as soon as that's done, it really gets easy to port blob |
15:54:19 | seletz | erikm: ok. |
15:54:55 | seletz | prpplague: well, varies. Got to work a lot, but i expected that. |
16:20:07 | erikm | makes a "src/commands/ directory |
16:25:01 | erikm | first commits his outstanding diffs to configure.in and acconfig.h |
16:33:03 | erikm | seletz: could you look at abraham's memory test questions in blob-cvs-commit? |
16:35:27 | sammy_wms | ho ho Sammy_wms return |
16:35:52 | erikm | seletz: I'll answer his third question |
16:35:56 | erikm | re, sammy_wms |
16:37:08 | sammy_wms | just save my HD , and get my RH7.0 back , |
16:39:00 | sammy_wms | but still need to buy one new HardDisk and one CD burner ...(bad block still there) |
16:39:59 | sammy_wms | thinks backup is important ..... |
16:55:06 | Russ | morning |
16:55:49 | erikm | hello Russ |
16:56:11 | erikm | adds a map file to the blob link command line |
17:07:41 | erikm | right, link map support is in, that would help debug BZFlag's problem |
17:07:52 | erikm | s/would/should/ |
17:08:16 | erikm | starts moving the commands |
17:45:41 | erikm | seletz: ok, we now have src/commands/ in CVS |
21:53:38 | erikm | hiya, BZFlah |
21:53:43 | erikm | hiya, BZFlag, even |
21:53:58 | erikm | BZFlag: the code map file change is in |
21:54:10 | erikm | s/code// |