| Summary: | [Beagleboard] Beagleboard can not boot up with 20110506 build | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] BSPs | Reporter: | Yi Zhao <yi.zhao> | ||||
| Component: | bsps-configuration | Assignee: | Darren Hart <dvhart> | ||||
| Status: | VERIFIED FIXED | QA Contact: | |||||
| Severity: | critical | ||||||
| Priority: | High | CC: | dvhart, jefro, jiajun.xu, liang.li2, liangliang.wang, pokylinux, sgw, yi.zhao, yp.bsp.watcher, yp.watcher | ||||
| Version: | 1.0 | ||||||
| Target Milestone: | 1.1 M1 | ||||||
| Hardware: | Beagleboard | ||||||
| OS: | arm | ||||||
| Whiteboard: | Waiting for QA to validate (19-May-2011) | ||||||
| OS type for building Yocto: | --- | Type of Regression: | --- | ||||
| Verified: | Documentation change: | --- | |||||
| Attachments: |
|
||||||
|
Description
Yi Zhao
2011-05-10 02:18:55 UTC
I'll be working on uboot recipes this week and next anyway, so I'll have a look at this. I believe that's due to the fact that what's in the latest version of git is a much older version than the one which worked before. Here is what I found out http://article.gmane.org/gmane.linux.embedded.poky/6114 http://article.gmane.org/gmane.linux.embedded.poky/6125 BTW you can find the version of u-boot without being able to run it like this: strings u-boot-beagleboard.bin | grep U-Boot U-Boot 2011.03-00038-g8d4addc-dirty (May 13 2011 - 00:01:02) ** Invalid partition type "%.32s" (expect "U-Boot") ** Invalid partition type "%.32s" (expect "U-Boot") U-Boot BUG at %s:%d! Das U-Boot U-Boot 2011.03-00038-g8d4addc-dirty BTW don't get fooled by the name: That's the working one for me: u-boot-beagleboard-v2010.12+r1+git2+0e0c0892a1fab6f93ba09d89ec3315f08f1b6ef4-r1.bin it's __not__ really v2010.12 In my case it's: strings u-boot-beagleboard-v2010.12+r1+git2+0e0c0892a1fab6f93ba09d89ec3315f08f1b6ef4-r1.bin | grep U-Boot U-Boot 2011.03-00038-g8d4addc-dirty (May 13 2011 - 00:01:02) ** Invalid partition type "%.32s" (expect "U-Boot") ** Invalid partition type "%.32s" (expect "U-Boot") U-Boot BUG at %s:%d! Das U-Boot U-Boot 2011.03-00038-g8d4addc-dirty here the naming is better, but that's the non-working one: strings u-boot-beagleboard-v2010.12+r2+git1+d0e3a3f4574f1f51ae3d7c5605b7ea5c6df929d4-r2.bin | grep U-Boot U-Boot 2010.12-00038-g8d4addc-dirty (May 13 2011 - 08:44:08) ** Invalid partition type "%.32s" (expect "U-Boot") U-Boot BUG at %s:%d! Das U-Boot U-Boot 2010.12-00038-g8d4addc-dirty This issue also happened on 1.0.1 build 20110513 Branch: Bernard git commit: 1b159ff35d9f484e72afb8b19a9e69a2770d8379 darren indicated he was looking into this, so I haven't been following. And then I noticed that I was still the Assignee. So I'm updating that field to reflect reality. Created attachment 157 [details]
v2011.03 master build
Confirmed that 2010.12 does not work on BBxM and that 2011.03 does. Patches tested on master, Bernard currently under test. I need someone to test on the C4. Please copy the attached file to the Beagleboard /boot filesystem as u-boot.bin, replacing the existing one.
I am rebuilding and testing with the latest (starting from a fresh git clone). I should have just left you my rev C board while I was in town! I can send it up if you like. I had to do this before it would read the first partition using the images you sent: - mmc init - mmc rescan 0 Then the boot script worked fine with this: - fatload mmc 0 0x80000000 boot.scr - source That last command just runs whatever is in memory at 80000000. Note that not everyone may have this problem, but I had to do that with two different cards, so I think it is an x-loader/u-boot thing. However, it is running now and appears to be stable. Fixed in Bernard: 53b15f2732cef26473ff651f86da3e4808616989 And in Master: c6dde3245eea43252c5bc1a9de616187a797909f (In reply to comment #8) > I had to do this before it would read the first partition using the images you > sent: > > - mmc init > - mmc rescan 0 > > Then the boot script worked fine with this: > > - fatload mmc 0 0x80000000 boot.scr > - source > > That last command just runs whatever is in memory at 80000000. > > Note that not everyone may have this problem, but I had to do that with two > different cards, so I think it is an x-loader/u-boot thing. However, it is > running now and appears to be stable. Hi Jefro, On a beagle-xm I see something very similar. You need to mmc init mmc rescan first. That's what I do in order to autoboot: http://thread.gmane.org/gmane.linux.embedded.poky/6141/match=autoboot With the Yocto 1.0.1 20100519 build, it still can not boot up. Image location: http://autobuilder.pokylinux.org/nightly/20110519-2/machines/beagleboard/arm/ Branch: bernard Commit info: 7d9e1bda1d539a2ee158e9657fd8a3ac1ead889f Although the file name is "u-boot-beagleboard-v2011.03+git1+d0e3a3f4574f1f51ae3d7c5605b7ea5c6df929d4-r3.bin", actually the version is "2010.12" ######################## $ strings u-boot-beagleboard-v2011.03+git1+d0e3a3f4574f1f51ae3d7c5605b7ea5c6df929d4-r3.bin |grep U-Boot U-Boot 2010.12-00038-g8d4addc-dirty (May 19 2011 - 20:26:23) ** Invalid partition type "%.32s" (expect "U-Boot") U-Boot BUG at %s:%d! Das U-Boot U-Boot 2010.12-00038-g8d4addc-dirty ######################## When I use the u-boot from attachment in comment 6, it can boot up. And the version is "2011.03". ########################## $ strings u-boot-beagleboard-v2011.03+r3+git1+b29fbb347698286935bfc401c08499a6f63479de-r3.bin |grep U-Boot U-Boot 2011.03-00038-g8d4addc-dirty (May 17 2011 - 11:31:29) ** Invalid partition type "%.32s" (expect "U-Boot") U-Boot BUG at %s:%d! Das U-Boot U-Boot 2011.03-00038-g8d4addc-dirty ########################## Fixed in oe-core http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=c6dde3245eea43252c5bc1a9de616187a797909f Fixed in Bernard http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=72f06800bcf6f3d188b99bd0c359f5ef74879621 and http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?h=bernard&id=53b15f2732cef26473ff651f86da3e4808616989 This bug had been fixed. Verified it. |