| 10:49:41 | erikm | cvs updates |
| 10:50:16 | erikm | seletz: can you make normal functions from SHOWADR() and SHOWRUN()? |
| 10:51:28 | erikm | seletz: if you make them functions you only have to make one function call for SHOWRUN() instead of three |
| 10:51:35 | erikm | seletz: which saves space |
| 10:53:17 | erikm | commit policy is easy: obvious bug fixes can always be comitted, larger changes should either be discussed over here, or on the lart mailing list |
| 11:06:30 | seletz | erikm: oops. I think i'd rather discussed my changes in chkmem.c before, right? |
| 11:07:49 | seletz | erikm: yes. Macros as SHOWRUN() considered as hacks by me, too. |
| 11:08:40 | seletz | erikm: should adhere to "do it right the first time", but it was late :^/ |
| 11:09:02 | ed__ | do right the 3rd or 4th time...it keeps it interesting |
| 11:09:20 | seletz | :^) |
| 11:10:05 | seletz | BTW: does one of you guys use lincvs or do you all cmd-line cvs? |
| 11:10:15 | erikm | just command line CVS |
| 11:10:40 | erikm | I'm fluent in CVS :) |
| 11:10:41 | seletz | guessed so. :) |
| 11:12:27 | seletz | I soon will add real support for our hw in blob. Something special to be done? |
| 11:13:19 | seletz | or is it just changing main.h, led.h and linux.h? |
| 11:15:34 | erikm | I documented in the README file |
| 11:15:39 | erikm | I think those are the files, yes |
| 11:15:49 | erikm | and the memory setup stuff |
| 11:20:02 | seletz | OK, i see nothing new in memsetup-sa1110.S. I'll add a couple of defines and ifdefs there, OK? |
| 11:41:23 | erikm | ok |
| 11:41:53 | erikm | I'll clean up later, this weekend was the "build small ramdisk" weekend |
| 11:45:06 | erikm | seletz: if at all possible, check if blob still compiles for all other architectures before you submit something |
| 11:45:28 | erikm | seletz: utils/build/ contains some scripts for that |
| 11:45:56 | erikm | seletz: which use the fact that builddir!=srcdir for autoconf'ed makefiles |
| 11:46:53 | seletz | well, yes. I shgould not have anything broken, tough. |
| 11:47:18 | erikm | better be sure.... |
| 11:47:26 | seletz | Sure. :) |
| 12:08:04 | seletz | erikm: is it ok to include SA-1100.h from linux kernel? Would make life much more comfortable. |
| 12:08:21 | seletz | erikm: in memsettp-sa1110.S |
| 12:31:48 | erikm | include "sa1100.h" |
| 12:31:59 | erikm | it automatically includes it in the correct way |
| 12:32:07 | erikm | ehm |
| 12:32:11 | erikm | OTOH, no, that doesn't work |
| 12:33:13 | erikm | not for assembly, that is |
| 12:33:34 | seletz | Ok, then i include SA-1100.h, ok? |
| 12:33:52 | erikm | also doesn't work |
| 12:34:02 | seletz | _aehm_ ? |
| 12:34:21 | seletz | #define _ASSEMBLY |
| 12:34:34 | erikm | I got funny errors last time I tried that. the assembler barfed on quite some symbols |
| 12:35:17 | seletz | well, i'll try it, at last. |
| 12:35:21 | seletz | ok? |
| 12:35:30 | erikm | go ahead |
| 12:36:22 | erikm | but remember that indexed register access (like in memsetup-sa1100.S) is a *lot* smaller |
| 12:38:09 | seletz | i currently hack it that way :) |
| 12:38:29 | seletz | i only need the constatns for MDREFR |
| 12:38:37 | seletz | and MSCxxx |
| 12:38:47 | seletz | _much_ more readable |
| 12:39:21 | erikm | SA-1100.h doesn't have those things indexed |
| 12:40:35 | seletz | no, i want to do mov r0, #(SLFRSH | K2DB2 | K1RUN) |
| 12:40:50 | seletz | or similar, even #define it before |
| 12:41:32 | erikm | better copy those from SA-1100.h, I think, because SA-1100.h will also define things like MDCNFG |
| 12:41:53 | seletz | well, ok. extra #include? |
| 12:42:16 | seletz | additional file, i mean |
| 12:42:25 | erikm | no, don't think that's necessary. it's only used in a single file, right? |
| 12:42:48 | seletz | well, perhaps memsetup-sa1100.S may use it also, no? |
| 12:43:20 | erikm | only the MDCNFG stuff, but not K2DB2 etc |
| 12:43:49 | erikm | if you wan to #define it for MDCNFG make it a separate header file, yes |
| 12:43:56 | erikm | (memsetup.h or so) |
| 12:44:04 | seletz | agreed. So #defines in memsetup-sa1110.h. can be changed easy afterwards. |
| 12:44:12 | seletz | .S i mean |
| 12:45:06 | erikm | yes, just cvs add include/memsetup.h and also add the file to include/Makefile.am |
| 13:09:45 | prpplague | morning all |
| 13:11:48 | seletz | hi |
| 13:37:09 | erikm | seletz: unless CheckMem/Poke/Peek gets called from somewhere else, there is no need to make them non-static functions |
| 13:37:59 | seletz | erikm: hmm. Its called by the command line parser, tough. |
| 13:38:11 | erikm | seletz: no |
| 13:38:27 | seletz | erik. "static: only used in this module" i thought. |
| 13:38:34 | erikm | seletz: the command line parser doesn't know about them, they are just function pointers exported by the __command() directive |
| 13:39:13 | seletz | well, then _every_ command func should be static. |
| 13:40:00 | seletz | erikm: they get called by the parser, are'nt they? |
| 13:40:51 | erikm | seletz: every command function should be, and most are, yes |
| 13:41:25 | erikm | seletz: the nice thing about the new command structure is that you can add commands without having to export a function |
| 13:42:29 | seletz | erikm: yes. fully understand that. IMHO every function that is called outside its defining module should be non-static. otherwise static for functions would be useless. |
| 13:42:58 | seletz | erikm: its not too much an issue, tough |
| 13:43:38 | erikm | seletz: compare with linux: most module have init_module() function to initialise them, but it doesn't clash because those functions are static |
| 13:45:15 | seletz | erikm: good point. But OTOH command funcs should be named unique. bad things happen when link-order is important. |
| 13:46:34 | prpplague | my_really_bad_init_module() |
| 13:46:57 | erikm | yes, I think the linker will complain about multiple defined modules |
| 13:47:37 | erikm | though I think I can also encode the filename in the command list entry |
| 13:48:13 | prpplague | erikm: i should have my mods for blob documented and a patch by friday |
| 13:49:44 | erikm | prpplague: cool |
| 13:58:10 | erikm | goes home |
| 14:21:11 | seletz | someone here who knows arm asm? |
| 14:31:38 | prpplague | seletz: only a little, not enough probably to help... |
| 14:33:39 | seletz | prpplague: give it a try. only a simple loop: i want to copy one table of longs to a specified address |
| 14:34:04 | seletz | prpplague: with auto-index |
| 14:34:39 | seletz | prpplague: i use "ldr r2, [r1], #4" to load data to r2 |
| 14:35:03 | seletz | prpplague: and suppose that r1 is post-incremented with 4. right? |
| 14:35:26 | seletz | prpplague: r1 contains the base adr of the source table |
| 14:36:03 | prpplague | sounds like you might be able to use something like this: |
| 14:36:05 | prpplague | /* r0 = source address |
| 14:36:05 | prpplague | * r1 = target address |
| 14:36:05 | prpplague | * r2 = source end address |
| 14:36:05 | prpplague | */ |
| 14:36:05 | prpplague | copy_loop: |
| 14:36:06 | prpplague | ldmiar0!, {r3-r10} |
| 14:36:07 | prpplague | stmiar1!, {r3-r10} |
| 14:36:09 | prpplague | cmpr0, r2 |
| 14:36:11 | prpplague | blecopy_loop |
| 14:37:12 | seletz | prpplague: well, this is block-copy, is it not? Copies in units of 8 longs. |
| 14:37:51 | prpplague | seletz: oops your right, didn't look close enough... |
| 14:38:12 | seletz | prpplague: what i really want is copy one long after each other. possibly odd number |
| 14:38:52 | seletz | prpplague: is above inexed load right? "ldr r2, [r1], #4" i mean? |
| 14:39:17 | seletz | prpplague: does it post-increment r1 |
| 14:40:33 | prpplague | seletz: that looks right, but again, i'm a newbie to arm asm... |
| 14:41:26 | seletz | prpplague: well, i'll keep on trying :^), but thank you nevertheless! |
| 14:42:13 | prpplague | seletz: sorry i couldn't be of more help, i know how it is doing dev work... |
| 14:44:07 | seletz | prpplague: you helped me! i now keep investigating those ldmia xx!, {} mnemonics! |
| 14:49:59 | prpplague | seletz: great! |
| 15:57:39 | prpplague | welcome back seletz... |
| 17:43:30 | sammy | ......back to bed <@_@> |
| 18:41:22 | prpplague | is away: lunch |
| 19:23:20 | prpplague | is back (gone 00:41:59) |
| 20:40:48 | EsSeNcE | hullo |
| 20:40:55 | EsSeNcE | sup ppls |
| 20:41:03 | EsSeNcE | really |
| 20:41:35 | EsSeNcE | y'all not tiddlie peeps are ya |
| 20:41:53 | EsSeNcE | kewl |
| 20:42:08 | EsSeNcE | blobby blobby blobby |
| 20:42:25 | prpplague | howdy |
| 20:42:31 | EsSeNcE | ahhhhhhhhhhhhhhhhh life |
| 20:42:35 | EsSeNcE | hiya |
| 20:42:44 | EsSeNcE | sup |
| 20:42:57 | prpplague | my wife might disagree about me being alive... |
| 20:43:09 | EsSeNcE | i know the feeling |
| 20:43:40 | EsSeNcE | feels avive from the toes down |
| 20:43:47 | EsSeNcE | alive* |
| 20:44:16 | EsSeNcE | brb |
| 20:51:31 | prpplague | BZFlag: hows biz? |
| 22:43:19 | prpplague | night guys |
| 22:43:26 | prpplague | happy blobing.... |