I'm running with iMX6 custom hardware, and have been trying to understand why my system isn't
coming up the way it should. I tracked this down to find that the line...
for i in /etc/init.d/S??* ;do
isn't substituting the 'S??*' with the files in /etc/init.d. Then I just tried
ls /bin/m* to see if it would list all files that start with 'm' in /bin. Even that
fails:
# ls /bin/m*
ls: /bin/m* : No such file or directory
I'm guessing there is some misconfiguration in busybox, but I haven't found anything.
-- EDIT on Jan 13 --
I managed to get this to work but have no clue why...
First, note that I am using buildroot to create my rootfs (kernel and uboot are mainline)...
In my buildroot configuration, I had "Target Architecture Variant" set to cortex-A9 (since I'm working with iMX6).
When I backed this setting off to A8, for some bizarre reason that fixed this issue.
What is it about A8/A9 that would cause regular expression wildcard processing in the shell not to work?
I sure would like to understand this one.
-- End EDIT --
Anyone ever have to deal with this ?
Tx
Ed
Message was edited by: Ed Sutter
解決済! 解決策の投稿を見る。
Ok, I think I've got this figured out, but its shakey...
I use buildroot to create my rootfs, and it uses the command line option "-m tune cortex-A9" when building.
For both u-boot (denx) and kernel (freescale git) that option is not used, so I tried manually removing it from
the buildroot scripts...
That fixed it. I can now configure buildroot with cortex-A9 target architecture and the above issue is gone.
Not 100% sure why... is there a bug in "-m tune cortex-A9" with the codesourcery tools?
Or is it possibly an option that must be used for kernel as well?
Has anyone ever used this option?
BTW, I'm using Sourcery CodeBench ARM 2013.05.
Ok, I think I've got this figured out, but its shakey...
I use buildroot to create my rootfs, and it uses the command line option "-m tune cortex-A9" when building.
For both u-boot (denx) and kernel (freescale git) that option is not used, so I tried manually removing it from
the buildroot scripts...
That fixed it. I can now configure buildroot with cortex-A9 target architecture and the above issue is gone.
Not 100% sure why... is there a bug in "-m tune cortex-A9" with the codesourcery tools?
Or is it possibly an option that must be used for kernel as well?
Has anyone ever used this option?
BTW, I'm using Sourcery CodeBench ARM 2013.05.
Ed
Had your issue got resolved? If yes, we are going to close the discussion in 3 days. If you still need help please feel
free to contact Freescale.
Thanks,
Yixing
Yixing,
I did resolve this, but as has been the case with a few other questions I post to the forum, I answered it myself (see above).
I've done it a few times with other questions that I've self-answered; but I hesitate to mark my own response as the "correct answer".
I guess I'll just do that so you can manage the state of the question better...
Ed
which shell are you running? (echo $0)
Leo
Leo,
(resending this worded differently to see if my recent failures were caused by the forum's spam filter...)
"echo $0" just shows -sh. According to busybox menuconfig, it appears to be ash.
Ed