irclog2html for #nslu2-linux on 20041001

00:01.46kergothrwhitby-away: want to mention the nslu2 stuff on the oe success stories page in the wiki?
00:03.30*** join/#nslu2-linux Christopher (
00:06.22Tierstenka6sox: No
00:09.36ka6soxI was just curious...
00:09.49ka6soxespecially the upper address bits.
00:10.01TierstenSomebody said they were connected
00:10.38ka6soxtrust but verify
00:11.21Tierstentrying to find it in the logs now
00:12.16Tierstenrwhitby said so
00:12.21TierstenAround August 13th
00:13.41TierstenThat's when the post was
00:13.44TierstenTrying to find it at the moment
00:22.32TierstenTook ages to find that
00:22.47TierstenThe search archive feature in Yahoogroups is terrible
00:23.14ka6soxit is
00:23.47TierstenAllan has traced it out so it should be okay
00:25.05ka6soxya..I see that.
00:25.08CodeWhackerImportant Safety Tip:  initrd=0x01000000,16M does not work very well at all
00:25.41Tiersten0x1000000 = 16777216
00:26.03TierstenYou'll hit the end of RAM by doing that and that's a bad thing
00:26.39CodeWhackerrats!  well, the 10M setting appears a bit small for the ramdisk contents
00:26.53TierstenTry something in between
00:27.36TierstenYou can't change the 0x01000000 as that is where redboot loads the ramdisk
00:27.44TierstenIf you want to change it then you must change redboot
00:28.09CodeWhackernot trying to change the base address
00:43.08ka6soxTiersten: anything else before I reflash to uNLSUng?
00:43.57ka6soxI'll install 1.92 again unless you want me to test 1.94.
00:45.05Tiersten1.94 is still broken.  I've not had any time in the last few days to look at it.  Been busy with real work
00:46.22ka6soxreal life sometimes is like that :P
00:46.50TierstenYeah.  Sucks :)
00:47.02TierstenI have to document the process to build redboot as well
00:47.05TierstenNot that it's hard
00:53.41CodeWhackerdoubleplusungood results
00:54.01CodeWhackerinitrd=0x01000000,13M fails as well
00:54.18TierstenLet me have another look at the redboot code
00:54.38CodeWhackeryou want to see the error?
00:55.55TierstenThe redboot code should load anything that fits
00:56.15TierstenIt doesn't have any limits on how big it is apart from size of flash and size of RAM
00:56.16CodeWhackerchecking if image is isn't (no cpio magic); looks like an initrd
00:56.16CodeWhackerDivision by zero in kernel.                                                    
00:56.16CodeWhacker[<c0054bc4>] ($a+0x0/0xfa4) from [<c0055ef4>] (try_to_free_pages+0x100/0x1f0)
00:56.16CodeWhacker[<c0055df4>] (try_to_free_pages+0x0/0x1f0) from [<c004da38>] ($a+0x208/0x304)
00:56.16CodeWhacker[<c004d830>] ($a+0x0/0x304) from [<c004b8b4>] (generic_file_aio_write_nolock+0x5e4/0xa7c)
00:56.17CodeWhacker[<c004b2d0>] (generic_file_aio_write_nolock+0x0/0xa7c) from [<c004be84>] ($a+0x58/0x80)  
00:56.19CodeWhacker[<c004be2c>] ($a+0x0/0x80) from [<c004bf68>] (generic_file_write+0x54/0x80)            
00:56.23CodeWhacker[<c004bf14>] (generic_file_write+0x0/0x80) from [<c0066e74>] (vfs_write+0xe4/0x11c)
00:56.27CodeWhacker[<c0066d90>] (vfs_write+0x0/0x11c) from [<c0066f44>] (sys_write+0x3c/0x58)
00:56.29CodeWhacker[<c0066f08>] (sys_write+0x0/0x58) from [<c000bc6c>] ($a+0xb4/0xd0)        
00:56.33CodeWhacker[<c000bbb8>] ($a+0x0/0xd0) from [<c001a1ac>] ($a+0x34/0x174)
00:56.37CodeWhacker[<c001a178>] ($a+0x0/0x174) from [<c00333d0>] (do_exit+0x0/0xa04)
00:56.43CodeWhackerand that repeats forever
00:56.47TierstenYou really don't need to paste the entire stacktrace
00:56.58TierstenThere is probably something you hitting at the end of memory
00:57.02CodeWhackerwasn't sure how much was relevant
00:58.01CodeWhackerwell, the old value was 10M, and the current ramdisk unpacks to just over 12M
00:58.38TierstenYou're probably overwriting something at the end
01:03.54ka6soxcatch you on the other side.
01:05.23CodeWhackersomething doesn't make sense here
01:06.16CodeWhackerwhen I unpack and mount the ramdisk, it only shows 4.4M used up, vs 4.0M from the Sunday build
01:07.24CodeWhackerso why is the unpacked size so much larger?
01:07.29ka6sox(headed home)
01:08.13TierstenYou got lots of little files?
01:08.30Tiersten if anybody is interested
01:10.46CodeWhackerTiersten: doesn't look like it, and the larger version only has 11 more files than the smaller one
01:11.08TierstenHow did you create the ramdisk?
01:11.19CodeWhackeroemake bootstrap-image
01:11.27CodeWhackerlooking for the creation command now
01:12.11dyoung-web2whats your CONFIG_DEV_BLK_RAM_SIZE set to?
01:14.08CodeWhackerdyoung: where's that?
01:14.45dyoung-web2Kernel Config.
01:16.28CodeWhackerdf says:
01:17.40CodeWhackerreally just looks like the ramdisk was arbitrarily created larger
01:19.30dyoung-web2Well, the 12M ramdisk wont fit because of that CONFIG_BLK_DEV_RAM_SIZE paramater is set to exactly 10M.  I dunno about the smaller one though.
01:20.07CodeWhackerok, that makes sense
01:21.24TierstenIt's set to 10?
01:21.31TierstenThe stock one is 16MB
01:25.21CodeWhackeradjusted and rebuilding