| 00:49.33 | WyrM | stefan^_: you can use some machid magic bytes just before the kernel to set machid. |
| 00:50.06 | WyrM | stefan^_: and you can add mem= to the builtin cmdline, and it will work flashed, without gen-blob. |
| 00:50.47 | WyrM | you can use http://people.openezx.org/wyrm/machid-magic as reference. |
| 00:51.57 | WyrM | it sets the machid as 867 (0x363), but its only 8 bytes, its easy to use a hexeditor to set any other machid you want. |
| 00:52.29 | WyrM | hexedit it and |
| 00:52.29 | WyrM | $ cat machid-magic zImage > zImage-to-be-flashed |
| 01:11.05 | stefan^_ | hia |
| 01:11.23 | stefan^_ | we tried it with the machid magic |
| 01:11.38 | stefan^_ | we have the problem that if we have gen-blob flashed |
| 01:12.20 | stefan^_ | and then flash the kernel on 0x20000 it seems to be that the genblob still gets executed, since the screen headder appears shortly |
| 01:26.48 | sabrod | Kernel flashed at 0x20000 overwrites genblob, lol. |
| 01:28.33 | sabrod | But since you don't reboot, genblob will leave in ram with no problem about that :) (like the kernel itself, before being "booted", it's copied on RAM) |
| 01:28.48 | sabrod | will live in RAM, not leave :) |
| 01:30.06 | sabrod | Yes, add theses 4 bytes and you have a generic way to boot a kernel on flash with a machid. (trust the inventor :) ) |
| 01:49.36 | WyrM | stefan^_: by screen headder you mean a motorola logo at startup? |
| 01:49.52 | WyrM | its drawn by MBM, not gen-blob. |
| 01:51.05 | WyrM | sabrod: its 8 bytes ;) |
| 01:51.16 | stefan^_ | it was after a reboot, by header i mean the title of gen blob (we print a name on the screen) |
| 01:51.38 | stefan^_ | still our kernel does not seem to start. |
| 01:52.01 | stefan^_ | we tried to load the kernel from different locations, we tried to hack gen-blob |
| 01:52.36 | WyrM | stefan^_: the screen may appear shortly if you turned off the phone with gen-blob on screen, because the data may still be on the FB memory when its initialized. |
| 01:53.05 | WyrM | stefan^_: have you added mem= to the builtin cmdline? |
| 01:53.10 | stefan^_ | what we did was comment out the read from mmc to ram and use memcpy from flash to KERNEL_RAM_BASE |
| 01:53.15 | stefan^_ | yes |
| 01:53.31 | stefan^_ | mem= statements are in the kernel |
| 01:53.41 | WyrM | ok. |
| 01:53.50 | WyrM | do you want to use gen-blob or not? |
| 01:53.56 | stefan^_ | we could try again to boot this kernel with mem statements and everything from mmc |
| 01:54.07 | stefan^_ | not necessarily |
| 01:54.14 | WyrM | ahh, I just remembered. |
| 01:54.21 | WyrM | you need it anyway. |
| 01:54.39 | WyrM | because the original bootloader will only copy 1MB of the kernel partition to RAM. |
| 01:54.40 | stefan^_ | we have some mainboards of a780 but we do not have the mmc module for them |
| 01:55.11 | stefan^_ | we thought we would not need them since we only want one fixed kernel and dont need 2.4 |
| 01:55.24 | WyrM | ok. |
| 01:55.39 | WyrM | let me see gen-blob code before I suggest you anything. |
| 01:55.41 | WyrM | just a min. |
| 01:55.48 | stefan^_ | we used MyMemCpy to copy 2 mb from flash to ram |
| 01:56.10 | stefan^_ | it seems its not the same as reading from mmc somehow |
| 01:56.59 | stefan^_ | we flashed the kernel to this 0xa something location that is defined as KERNEL_FLASH_BASE |
| 01:57.40 | stefan^_ | genblob compplained that it should be at memory adresses that are multiples of 20000 |
| 01:57.57 | stefan^_ | can you suggest a location that would work? |
| 01:58.19 | WyrM | you are working on a780, correct? |
| 01:58.23 | stefan^_ | yes |
| 01:58.42 | stefan^_ | we also tried once to hack the kernel to always use 1740 |
| 01:58.43 | WyrM | revert your changes on gen-blob, or svn checkout it again. |
| 01:58.55 | stefan^_ | ok we can do that tommorrow |
| 01:59.06 | WyrM | ok, then log what im going to say :) |
| 01:59.14 | stefan^_ | yes |
| 02:00.48 | WyrM | on include/blob/arch/lubbock.h: |
| 02:00.48 | WyrM | change KERN_MAX_SIZE to (2048 * 1024) - 4 |
| 02:01.19 | WyrM | on src/blob/boot_modes.c |
| 02:01.38 | WyrM | change "int machid = 0;" to 1740 |
| 02:02.29 | WyrM | build gen-blob, flash it to 0x20000 |
| 02:03.04 | WyrM | download http://people.openezx.org/wyrm/gen-blob/magic-boot-from-flash |
| 02:03.32 | WyrM | cat magic-boot-from-flash zImage > zImage-to-be-flashed |
| 02:03.47 | WyrM | flash zImage-to-be-flashed to 0x40000 |
| 02:04.15 | WyrM | it will overwrite the rootfs partition, so you will need to place your rootfs somewhere else.. |
| 02:04.29 | tmzt | stefan^_: would the sl5500 one work? do you mean sd module for 2.4? |
| 02:04.32 | stefan^_ | yes we will boot rootfs from a usb device |
| 02:04.38 | *** join/#openezx arne (i=nobody@dslc-082-082-084-123.pools.arcor-ip.net) |
| 02:04.43 | arne | re |
| 02:04.49 | WyrM | hi arne |
| 02:04.59 | arne | WyrM: what if gen-blob is larger than the 0x20000 in between? |
| 02:05.10 | arne | relating to stefan's query |
| 02:05.16 | arne | the thing is, our's was 300k or so |
| 02:05.33 | WyrM | why? what have you added to it? I want it too! :) lol |
| 02:05.54 | stefan^_ | some boot menu options , like usb mode poweroff and stuff |
| 02:06.05 | stefan^_ | i think |
| 02:06.10 | stefan^_ | that correct, arne? |
| 02:06.28 | stefan^_ | but we could just make it smaller again... |
| 02:06.35 | stefan^_ | using // |
| 02:06.38 | stefan^_ | ;) |
| 02:07.01 | WyrM | arne: then you need to change |
| 02:07.01 | WyrM | GEN_BLOB_MAX_SIZE on include/blob/arch/lubbock.h |
| 02:07.15 | arne | maybe it was because we enabled the debug feature in the meantime |
| 02:07.21 | WyrM | always to a 128k multiple. |
| 02:07.31 | arne | let me recheck |
| 02:07.56 | arne | <PROTECTED> |
| 02:07.57 | arne | hmm |
| 02:08.41 | arne | but what i didnt underswtand was when we flash the zImage into 0x20000 |
| 02:08.44 | arne | why would it still boot gen-blob? |
| 02:08.52 | arne | if it says flashing was okay, can there be any other problem? |
| 02:08.57 | arne | i mean how can it still contain gen-blob |
| 02:14.57 | arne | ok maybe we contaminated gen-blob :) |
| 02:17.01 | *** join/#openezx WyrM1 (n=wyrm@20158149197.user.veloxzone.com.br) |
| 02:17.15 | WyrM1 | last message I saw: arne: 303K Mar 9 13:33 src/blob/blob-a780 |
| 02:17.30 | arne | arne> hmm |
| 02:17.30 | arne | <arne> but what i didnt underswtand was when we flash the zImage into 0x20000 |
| 02:17.30 | arne | <arne> why would it still boot gen-blob? |
| 02:17.31 | arne | <arne> if it says flashing was okay, can there be any other problem? |
| 02:17.31 | arne | <arne> i mean how can it still contain gen-blob |
| 02:17.31 | arne | <arne> ok maybe we contaminated gen-blob :) |
| 02:18.33 | WyrM1 | arne: what may happen is that on the first boot after flashing, the FB memory still contains the last screen drawn by gen-blob. |
| 02:18.57 | WyrM1 | when you boot the kernel this screen may appear when the lcd is first turned on. |
| 02:19.14 | WyrM1 | but if you flash the kernel to 0x20000 it would not work because of the size. |
| 02:19.28 | arne | ok |
| 02:19.33 | arne | but if i flash the kernel in 0x20000 |
| 02:19.36 | WyrM1 | its too big for the original bootloader (only loads 1024k to ram). |
| 02:19.36 | arne | then switch the phone off |
| 02:19.37 | arne | no power |
| 02:19.41 | arne | then switch back on |
| 02:19.49 | arne | how can there be a live interactive menu of gen-blob |
| 02:20.05 | WyrM1 | the ram is preserved even after some minutes. |
| 02:20.18 | WyrM1 | I used this "feature" to debug suspend resume :) |
| 02:20.25 | arne | ok :) |
| 02:20.48 | arne | but what happened i thought that the initial bootstrap thing that boots the 0x20000 place would try to boot the kernel and fail |
| 02:21.02 | WyrM1 | it tries to boot a partial kernel. |
| 02:21.11 | WyrM1 | and the kernel fails. :) |
| 02:25.00 | arne | ok |
| 02:25.06 | arne | but it would not go into gen-blob right? |
| 02:25.40 | WyrM1 | ahh, you mean a live gen-blob? |
| 02:25.44 | arne | yes |
| 02:25.52 | WyrM1 | this shouldn't happen! |
| 02:26.18 | WyrM1 | the flash failed, this really can't happen. |
| 02:26.20 | arne | is there any way that the host-side boot_usb thinks that its flashing but is not? |
| 02:26.36 | WyrM | not really. |
| 02:27.03 | WyrM | and it also checks for the gen-blob reply that the flash succeded. |
| 02:28.27 | arne | ok |
| 02:28.31 | arne | thats strange |
| 02:28.39 | WyrM | are you sure that the address is correct? |
| 02:29.06 | WyrM | 0x20000 not 0x200000. |
| 02:29.46 | WyrM | small errors like these leads to very weird black magic.. ;) |
| 02:30.56 | arne | yes |
| 02:31.01 | arne | but i'm sure :) |
| 02:31.05 | arne | i know its unbelievable |
| 02:31.10 | arne | i will recheck it tomorrow or so |
| 02:31.20 | arne | there has to be an explanation anyway.. |
| 02:31.31 | WyrM | also, check the lcd when you are flashing. |
| 02:31.41 | WyrM | any flash layer errors will show up on the lcd. |
| 02:31.46 | arne | but basically we have to have a gen-blob + zimage in flash to get it to work right |
| 02:31.54 | arne | what again would you propose as a flash-addr for the zimage? |
| 02:31.56 | WyrM | and just a generic FAIL will appear for boot_usb. |
| 02:31.58 | arne | 0x40000? |
| 02:32.24 | WyrM | with a gen-blob smaller than 128k, 0x40000. |
| 02:32.32 | arne | ok |
| 02:34.53 | arne | if its bigger the next 0x20000 block right? |
| 02:35.00 | arne | like for example 0x80000 |
| 02:35.35 | arne | <WyrM> GEN_BLOB_MAX_SIZE on include/blob/arch/lubbock.h |
| 02:35.37 | arne | i dont have this parameter |
| 02:36.35 | arne | ok we need to get the newest gen-blob |
| 02:36.38 | arne | i think ours is messed up |
| 02:38.24 | WyrM | arne: the latest gen-blob has code to boot the 2.4 kernel from flash. |
| 02:39.02 | WyrM | its working on all my phones, I did it because gen-blob can't use the a910 mmc slot, and because I want to be able to charge phones without a card :) |
| 02:39.43 | WyrM | it uses a magic-boot-from-flash flag placed at gen-blob-addr + gen-blob-max-size. |
| 02:40.27 | arne | ok |
| 02:40.43 | WyrM | if this flag exists, then it loads kernel-max-size bytes from (gen-blob-addr+gen-blob-max-size+4) and jumps to it. |
| 02:40.44 | arne | so the magic-boot-from-flash pushes the machid into the stack and starts the kernel? |
| 02:40.56 | WyrM | no. |
| 02:41.14 | WyrM | the magic-boot-from-flash (4 bytes) is just a flag. |
| 02:41.16 | arne | i mean in gen-blob there was this function pointer with two int-parameters that was being called ("theKernel") |
| 02:41.28 | arne | so i guessed that the machid needs to be on the stack in order to run the kernel |
| 02:41.56 | WyrM | http://people.openezx.org/wyrm/machid-magic |
| 02:41.57 | WyrM | ^ this one sets the machid if you put it before the kernel. |
| 02:42.42 | WyrM | it is set to machid 867 (0x363), but you can edit it to any other machid you want. |
| 02:42.48 | arne | yes i know |
| 02:42.52 | arne | its 8-byte binary |
| 02:43.05 | arne | we tried it (hex-edited to 1740) |
| 02:43.08 | arne | i mean decimal 1740 |
| 02:43.19 | arne | but obviously we had some problems with our gen-blob :) |
| 02:43.38 | arne | but this binary thing does not push the machid on the stack? |
| 02:43.52 | WyrM | and you can |
| 02:43.52 | WyrM | cat magic-boot-from-flash machid-magic zImage > kernel-to-be-flashed, it will work too ;) |
| 02:44.02 | arne | ok sorry |
| 02:44.05 | arne | :) |
| 02:44.12 | arne | i thought machid-magic = magic-boot-from-flash |
| 02:44.13 | arne | :) |
| 02:44.24 | WyrM | the machid magic puts the machid on r0. |
| 02:45.08 | arne | so the magic-boot-from-flash is a memcpy + jump? |
| 02:45.58 | WyrM | arm-none-eabi-objdump -D -b binary -m arm machid-magic |
| 02:45.58 | WyrM | <PROTECTED> |
| 02:45.58 | WyrM | <PROTECTED> |
| 02:46.10 | WyrM | no, magic-boot-from-flash is only 4 bytes. |
| 02:46.32 | WyrM | and its just a flag so gen-blob can recognize that there is a flash image on flash. |
| 02:46.46 | WyrM | s/flash image/kernel image/ |
| 02:48.53 | WyrM | magic-boot-from-flash is KERN_ON_FLASH_FLAG on lubbock.h |
| 02:49.10 | WyrM | 0xDEADB007 |
| 02:51.35 | arne | ok |
| 03:00.54 | arne | thanks for all the info :) |
| 03:00.58 | arne | i'll try tomorrow |
| 03:01.08 | WyrM | You are welcome :) |
| 03:01.57 | arne | by the way, if you are interested, we can get a780 mainboards from china :) |
| 03:02.16 | arne | around 15eu per piece |
| 03:02.27 | arne | the camera module is ~2eu per piece |
| 03:02.36 | arne | and display is also available |
| 03:03.05 | arne | i said some month ago in the channel we could get used a780s in larger quantities but its not so easy, but the mainboards seem to be more available |
| 03:03.10 | arne | we got 10 of them last week |
| 03:03.46 | arne | and using the usbhost patch and power-pin mod, we can boot from an usb-stick :) |
| 03:04.13 | arne | but the problem is the kernel, we successfully tried booting gen-blob from flash, then kernel frmo mmc and rootfs from usb-flash |
| 03:04.39 | arne | but it should be possible as you described to boot the kernel frmo flash, then the rootfs from usb-flash, of course one could use a flash-based rootfs, but we prefer usb-flash |
| 03:05.08 | WyrM | good to know! :) I will talk to you if need any parts. Thanks! |
| 03:05.13 | arne | no problem |
| 03:05.25 | WyrM | Yes, with usb flash you have more freedom. |
| 03:05.33 | WyrM | the builtin flash on a780 is only 32mb |
| 03:05.49 | WyrM | there is a nand chip with 64mb, but we still don't have a driver for it. |
| 03:05.58 | WyrM | (and probably we will never have). |
| 03:06.00 | arne | yes and you cannot easily change it :) with usbflash you can deploy the device and then change the usb-flash without reflashing etc |
| 03:07.00 | WyrM | well, with the usb host patch in theory you can use one device to flash another ;) |
| 03:07.19 | WyrM | so you can just use a phone to flash your devices when you need to update it. |
| 03:07.40 | WyrM | as easy as updating all your flash disks ;) |
| 03:08.21 | arne | depending where the device is :) |
| 03:08.44 | arne | if you need to send out a mail or so to a remote area where theres only one a780, then i guess usb-flash wins :) |
| 03:09.09 | WyrM | are you using wifi on these devices? |
| 03:09.12 | arne | yes |
| 03:09.16 | arne | using d-link dwl-122 |
| 03:09.22 | arne | we can get it very cheap from china |
| 03:09.22 | WyrM | then only software wins ;) |
| 03:09.37 | arne | by the way |
| 03:09.43 | arne | do you think it's hard to get gprs working? |
| 03:09.50 | WyrM | no. |
| 03:09.56 | arne | what is missing? |
| 03:10.01 | WyrM | it will be probably easy. |
| 03:10.36 | arne | once you sent/received all the AT stuff to enter a gprs session, does it behave like a regular modem? |
| 03:10.39 | WyrM | there is a driver on the 2.4 code, for gprs, using the kernel net layer directly. |
| 03:10.39 | arne | i mean does it do PPP? |
| 03:10.43 | arne | ok |
| 03:10.52 | WyrM | but it should also be possible to use pppd on a mux device. |
| 03:11.19 | WyrM | as the phone original firmware can export ppp from the carrier when used in modem mode. |
| 03:11.51 | arne | ok |
| 03:12.01 | WyrM | I don't know what exactly is needed to get this mode working, but some strace on the original firmware can point us the commands to get ppp on a mux. |
| 03:12.36 | arne | ok |
| 03:59.55 | *** join/#openezx WyrM (n=wyrm@20158130029.user.veloxzone.com.br) |
| 05:28.03 | *** join/#openezx uwe_ (n=uwe_@dslb-084-056-024-012.pools.arcor-ip.net) |
| 05:52.02 | *** join/#openezx toi (n=toi@d54C2AAB7.access.telenet.be) |
| 06:11.22 | *** join/#openezx sabrod (n=sab@lns-bzn-33-82-252-28-176.adsl.proxad.net) |
| 06:48.31 | *** join/#openezx stefan_schmidt (n=stefan@w1506.wlan.rz.tu-bs.de) |
| 07:51.22 | *** join/#openezx l403 (n=l@85.132.159.239) |
| 08:04.23 | *** join/#openezx WyrM (n=wyrm@20158134019.user.veloxzone.com.br) |
| 08:05.10 | Hopscotch | good morning |
| 09:12.58 | *** join/#openezx gDD (i=io@125.116.32.50) |
| 09:18.04 | *** join/#openezx florian_kc (n=fuchs@port-217-146-132-69.static.qsc.de) |
| 10:35.27 | *** join/#openezx mickey|OpenExpo (n=mickey@zux190-231.adsl.green.ch) |
| 10:35.42 | mickey|OpenExpo | anyone found time to look at the ts on gen1 yet? |
| 10:49.18 | *** join/#openezx tombhadAC (n=tombhad@pD95DDB9A.dip.t-dialin.net) |
| 11:01.26 | *** join/#openezx WyrM1 (n=wyrm@20158173072.user.veloxzone.com.br) |
| 12:27.45 | *** join/#openezx gmzhuo (n=gmzhuo@125.69.10.123) |
| 12:28.17 | *** join/#openezx flyhorse|linux (n=flyhorse@116.18.154.56) |
| 12:33.26 | *** join/#openezx tombhadAC (n=tombhad@pD95DDB9A.dip.t-dialin.net) |
| 12:39.09 | *** join/#openezx flyhorse|linux (n=flyhorse@116.18.154.56) |
| 12:59.34 | *** join/#openezx trod (n=trod@59.163.114.3) |
| 13:44.01 | gmzhuo | anybody tested http://ezxdev.org/qtopia/testing/ |
| 13:45.09 | gmzhuo | I can't get ssh connect use rootfs http://ezxdev.org/qtopia/testing/Angstrom-console-image-glibc-ipk-2009.X-test-20090331-a1200.rootfs.tar.bz2 |
| 14:59.21 | *** join/#openezx sabrod (n=sab@lns-bzn-38-82-253-78-94.adsl.proxad.net) |
| 15:01.06 | *** join/#openezx johanbr (n=j@JBrannlund2.MathStat.Dal.Ca) |
| 15:28.09 | *** join/#openezx pleemans (n=toi@d51A49D31.access.telenet.be) |
| 15:43.18 | *** join/#openezx tombhadAC (n=tombhad@pD95DD129.dip.t-dialin.net) |
| 15:50.11 | gmzhuo | no ts respons on qtopia |
| 16:08.58 | *** join/#openezx gmzhuo- (n=gmzhuo@125.69.10.123) |
| 17:04.29 | *** join/#openezx ao2 (n=none0@190-197-dsl.kielnet.net) |
| 17:04.44 | ao2 | hi |
| 17:08.48 | gmzhuo- | hi |
| 17:55.10 | *** join/#openezx pleemans (n=toi@81.164.157.49) |
| 18:40.44 | *** join/#openezx stefan_schmidt (n=stefan@p5B03504A.dip.t-dialin.net) |
| 18:41.16 | *** join/#openezx tombhadAC (n=tombhad@pD95DE397.dip.t-dialin.net) |
| 19:32.13 | *** join/#openezx stefan_schmidt_ (n=stefan@p5B037C13.dip.t-dialin.net) |
| 19:49.12 | *** join/#openezx florian (n=fuchs@g227204067.adsl.alicedsl.de) |
| 20:11.51 | *** join/#openezx pleemans (n=toi@d54C2AAB7.access.telenet.be) |
| 21:02.46 | yang | hey ao2 still awake ? |
| 21:02.54 | yang | WyrM1: ?* |
| 21:03.21 | sabrod | Hi yang! |
| 21:03.50 | yang | I would like to know about the web-gui which would let m e track the code progress of the project git/svn/trac , what should I choose ? |
| 21:03.54 | yang | hi sabrod ! |
| 21:05.31 | sabrod | Well, svn is now used for the tools , and git for the kernel... |
| 21:06.05 | sabrod | Both have a web interface and probably a sort of "feed" to track them... |
| 21:06.12 | yang | I am talking about a project outside openezx |
| 21:06.18 | sabrod | CIA stats is also pretty readable :) |
| 21:06.58 | yang | which tools should I use for tracking this ? |
| 21:07.23 | sabrod | Heh, I use "tootella" :) |
| 21:07.38 | yang | hm |
| 21:07.58 | yang | must be inside debian |
| 21:08.18 | sabrod | Nah, tootlella is a french website ;) |
| 21:08.52 | sabrod | liferea works! |
| 21:09.18 | sabrod | But you want to have a gui for your own project ? |
| 21:09.34 | sabrod | github provide a goo service! |
| 21:09.36 | sabrod | good |
| 21:10.38 | yang | git-core - fast, scalable, distributed revision control system |
| 21:11.27 | yang | is there any difference between svn and git ? |
| 21:18.37 | sabrod | Well, they are differents, yes. |
| 21:18.56 | yang | yes |
| 21:19.10 | yang | but they do basically the same thing, just differnt interface?* |
| 21:21.10 | l403 | Id say they are kinda rivals |
| 21:21.28 | yang | ok |
| 21:21.38 | yang | http://en.wikipedia.org/wiki/Git_(software) |
| 21:21.57 | yang | really many projects use this |
| 21:22.09 | yang | .o0 (must be good) |
| 21:24.06 | sabrod | yang a guy make some very good git's tutorials >> gitcasts at http://www.gitcasts.com/ |
| 21:25.13 | yang | thanks |
| 21:25.25 | yang | sabrod: do you have any git experience? |
| 21:26.06 | sabrod | Brr, very little! :) |
| 21:26.19 | sabrod | Why? |
| 21:26.28 | yang | I gotta learn it from the start |
| 21:29.22 | sabrod | Ok, so , forgot about gitcasts. You'll go to that, later! :) |
| 21:30.18 | sabrod | Well, in fact it's easy, if you already know how to handle a revision system... |
| 21:30.58 | yang | sabrod: are you that french guy who also collects computers? |
| 21:31.55 | yang | I remember him from #gnupg, but I cannot recall if you are him |
| 21:31.55 | sabrod | Hum, I don't think so... ( but I have to admit: I collect portables computer, in fact :) ) |
| 21:57.32 | stefan^_ | hi there |
| 21:58.15 | stefan^_ | one question: we are right now in the process of flashing the kernel into the flash at 0x80000, however it seems that it is not updating the bytes at 0x80000 after "successful" flashing |
| 21:58.49 | stefan^_ | we modded the bytes using a heeditor and put in code in gen-blob to dump the flash at this address, and after changing with different bytes it stays the same after flashing |
| 21:59.00 | stefan^_ | is it possible that the internal flash (motorola a780) is somehow locked? |
| 21:59.04 | stefan^_ | at 0x80000 that is |
| 21:59.43 | stefan^_ | (gen-blob is at 0x20000) and we could start-up a 2.4 (at least its logo) by flashing it to 0x80000 |
| 22:10.11 | ao2 | stefan^_, have you checked inside gen-blob code yet? gen-blob can be used to flash stuff at kernel address as you know, but I have never tried flashing where rootfs is |
| 22:10.48 | stefan^_ | no, rootfs is not in flash |
| 22:10.59 | stefan^_ | and we did adapt the gen-blob addresses |
| 22:11.19 | stefan^_ | because gen-blob is ~300k so we had to increase the gen-blob-max-size to 384*1024 |
| 22:11.29 | stefan^_ | therefore 0x20000 + 0x60000 -> 0x80000 |
| 22:12.19 | stefan^_ | i just created a /dev/zero filled file and flashed it into 0x80000 (1.9mb) |
| 22:12.33 | stefan^_ | upon rebooting and reading from this part of the flash it gives me the same values as before |
| 22:12.34 | stefan^_ | how can this be? |
| 22:22.06 | ao2 | I don't have any theory right now, sorry |
| 22:22.16 | stefan^_ | no problem |
| 22:22.22 | stefan^_ | but do you know anything about a locked flash or so? |
| 22:22.27 | stefan^_ | or too large to flash or something |
| 22:22.30 | stefan^_ | our kernel is 1.9mb |
| 22:22.42 | stefan^_ | anything we might try |
| 22:23.55 | ao2 | I do know that moto kernel locks flash by software, but I am not aware of any other limitation |
| 22:24.33 | ao2 | have to go now, I hope we will find a solution soon stefan^_ Bye |
| 22:58.35 | *** join/#openezx sabrod (n=sab@lns-bzn-45-82-65-191-8.adsl.proxad.net) |
| 23:02.10 | *** join/#openezx Fosforo (n=Fosforo@20151018197.user.veloxzone.com.br) |
| 23:19.46 | WyrM1 | stefan^_: I think that your gen-blob is buggy somehow. |
| 23:20.10 | WyrM | because I really don't see how it could be 300k long. :/ |
| 23:20.31 | WyrM | unless you added a lot of stuff to it (not just a few menu entries for usb mode). |
| 23:21.33 | stefan^_ | no we did not |
| 23:21.40 | stefan^_ | actually its 300k from the regular source (svn) |
| 23:21.50 | stefan^_ | maybe compiler wrong? |
| 23:21.58 | WyrM | stefan^_: if you send me your source code (a .tgz is ok), I can compile it here and try to determine why it got so big with your toolchain. |
| 23:22.11 | WyrM | yes, I suspect that something is wrong with your toolchain. |
| 23:22.14 | stefan^_ | ok wait |
| 23:22.19 | WyrM | its less than 64k here. |
| 23:22.30 | WyrM | send it to my email. |
| 23:22.57 | stefan^_ | how large is your kernel image? |
| 23:22.59 | stefan^_ | ours is 1.9mb |
| 23:23.23 | WyrM | here too. |
| 23:23.39 | WyrM | but the zImage is compressed, not comparable to gen-blob. |
| 23:24.00 | WyrM | the uncompressed kernel image is 4.6 MB here. |
| 23:24.00 | stefan^_ | how do you keep off the logs? by adding a space-prefix? |
| 23:24.04 | WyrM | with the default .config |
| 23:24.05 | stefan^_ | (irc channel) |
| 23:24.14 | WyrM | yes. |
| 23:24.22 | WyrM | <space>message. |
| 23:24.54 | stefan^_ | do you have any idea why flashing at address 0x8000 with 1.9mb full of 0x00 does not empty the flash at this address? |
| 23:24.58 | stefan^_ | err |
| 23:25.00 | stefan^_ | 0x80000 |
| 23:25.39 | WyrM | stefan^_: have you tried flashin with my gen-blob on http://people.openezx.org/wyrm/gen-blob? |
| 23:25.41 | stefan^_ | i dropped the flash-flag detection from gen-blob, moved the read-in to 0x80000 and only did magic-machid + zimage |
| 23:25.58 | stefan^_ | if i use this kernel just for booting into through usb, it works |
| 23:26.05 | stefan^_ | but flashde on the device it wont |
| 23:26.22 | stefan^_ | the point is, somehow the flash process does not really put the bytes into their destination |
| 23:26.37 | stefan^_ | as i said, a file full of 0x00 does not yield an empty flash at 0x80000 |
| 23:26.39 | WyrM | I think that your gen-blob version is not working, your version somehow is not flashing the data. |
| 23:26.40 | stefan^_ | i'll try your gen-blob |
| 23:26.45 | stefan^_ | do you have a boot-usb precompiled too? |
| 23:27.06 | WyrM | no, but I can upload one for you, just a min. |
| 23:27.11 | stefan^_ | thank you :) |
| 23:28.19 | WyrM | stefan^_: http://people.openezx.org/wyrm/boot_usb |
| 23:28.26 | stefan^_ | thanks :) |
| 23:28.33 | WyrM | 0d3ace9ef1d298cfbd6caac1b04d274b boot_usb |
| 23:28.45 | stefan^_ | 302bf512fe5eb8c5ebe66bbf8558ae9c gen-blob |
| 23:29.05 | stefan^_ | ./boot_usb: /lib/libc.so.6: version `GLIBC_2.7' not found (required by ./boot_usb) |
| 23:29.06 | stefan^_ | damnit |
| 23:29.16 | stefan^_ | 0d3ace9ef1d298cfbd6caac1b04d274b boot_usb |
| 23:29.23 | stefan^_ | but its the same file :) i guess different libcs |
| 23:29.34 | WyrM | gen-blob md5sum is ok. |
| 23:29.40 | stefan^_ | <PROTECTED> |
| 23:29.48 | WyrM | yes, its ubuntu here. |
| 23:31.00 | stefan^_ | so with your gen-blob i need to: cat magic-boot-from-flash machid-magic zimage > kernel |
| 23:31.04 | stefan^_ | and then to 0x40000 ? |
| 23:31.10 | stefan^_ | machid-magic: |
| 23:31.17 | stefan^_ | 0000000 1004 e51f 06cc 0000 |
| 23:31.18 | stefan^_ | 0000008 |
| 23:31.39 | stefan^_ | (hexdump output) |
| 23:32.34 | sabrod | Stefan, you can use mine http://people.openezx.org/sabrod/boot_usb_svnr2421 |
| 23:32.47 | sabrod | I flashed severals stuff with it :) |
| 23:32.53 | WyrM | yes, machid magic is correct. |
| 23:33.05 | stefan^_ | WyrM: at 0x40000 |
| 23:33.13 | stefan^_ | sabrod: thanks, i'll try if this setup doesnt work |
| 23:33.23 | WyrM | yes. |
| 23:33.26 | stefan^_ | one try with the old one :) |
| 23:33.33 | stefan^_ | ok flashing |
| 23:34.34 | stefan^_ | ok what do i do to make it boot from flash? any special key combo? |
| 23:34.39 | stefan^_ | it tells me it wont find mmc |
| 23:35.00 | stefan^_ | last time i removed some source so it would boot from flash :) but there was one check for some buttons before it would check the flash flag |
| 23:35.01 | WyrM | you need the magic-boot-from-flash at 0x40000. |
| 23:35.06 | stefan^_ | it is there |
| 23:35.13 | WyrM | and the kernel at 0x40000 + 4 |
| 23:35.24 | stefan^_ | i thought machid at 0x40004 |
| 23:35.31 | stefan^_ | and kernel after that |
| 23:35.35 | WyrM | yes, its ok too. |
| 23:35.38 | stefan^_ | ok |
| 23:36.11 | stefan^_ | WyrM: can you please try to compile my gen-blob and see the size |
| 23:36.28 | WyrM | magic-boot-from-flash at 0x40000, machid-magic at 0x40000 and the the kernel at 0x40000 + 12. |
| 23:36.42 | stefan^_ | second item + 4 right |
| 23:36.49 | WyrM | s/0 and/0 + 4 and/ |
| 23:37.16 | stefan^_ | yes i think that is what i flashed in it |
| 23:37.24 | WyrM | im building your source now. |
| 23:37.24 | stefan^_ | <PROTECTED> |
| 23:37.37 | stefan^_ | i'm not sure what this checks for |
| 23:37.53 | WyrM | ! (not enterering the menu) |
| 23:38.01 | WyrM | if you are not pressing vol+ |
| 23:38.08 | stefan^_ | ok |
| 23:38.14 | WyrM | err, vol- |
| 23:38.28 | stefan^_ | so if i dont press anything (except maybe power key) then it should boot from flash |
| 23:38.38 | WyrM | yes |
| 23:40.19 | WyrM | -rwxr-xr-x 1 wyrm users 65536 2009-04-01 17:37 src/blob/blob |
| 23:40.24 | stefan^_ | hehe |
| 23:40.24 | WyrM | ^ with your sources. |
| 23:40.25 | stefan^_ | ok |
| 23:40.46 | stefan^_ | can you please give me the binary? :) |
| 23:41.04 | WyrM | arm-angstrom-linux-gnueabi-gcc --version |
| 23:41.05 | WyrM | arm-angstrom-linux-gnueabi-gcc (GCC) 4.1.2 |
| 23:41.10 | WyrM | sure |
| 23:41.19 | stefan^_ | gcc version 4.2.4 |
| 23:41.28 | stefan^_ | arm-angstrom-linux-gnueabi-gcc |
| 23:41.35 | stefan^_ | Target: arm-angstrom-linux-gnueabi |
| 23:42.11 | WyrM | http://people.openezx.org/wyrm/stefan-gen-blob |
| 23:42.23 | WyrM | no changes to your code. |
| 23:42.31 | stefan^_ | ok thanks let me try |
| 23:43.01 | stefan^_ | actually first i'm going to try and flash the 0x00 file in |
| 23:43.06 | stefan^_ | theres nothing preventing this right? |
| 23:43.17 | WyrM | right. |
| 23:43.26 | stefan^_ | (just making sure) |
| 23:47.56 | stefan^_ | hmm doesnt work either |
| 23:48.26 | stefan^_ | not with my setup and also not with your gen-blob + the 0x40000 with flag and machid and kernel setup |
| 23:48.47 | stefan^_ | and i checked by flashing an empty file into 0x80000 with your gen-blob and then using my gen-blob to read it out |
| 23:48.49 | stefan^_ | and it was not empty |
| 23:49.32 | stefan^_ | am i missing something or is the flash mounted 'as-is' (0x20000 flash == 0x20000 in the AP memory space) |
| 23:49.58 | stefan^_ | and we already switched hardware once tonight |
| 23:52.18 | sabrod | stefan^_: this http://pastebin.com/m34c34bca method works 100% here... |
| 23:54.52 | stefan^_ | ok but it writes to 0xa0000 |
| 23:55.30 | stefan^_ | and only 1mb |
| 23:55.35 | stefan^_ | so its for the original kernel? |
| 23:59.03 | stefan^_ | ok we have another laptop with glibc 2.7 and will try the boot_usb |