mkinitrd_command_generator.sh: broken with 3.17.3 kernel
Hello,
today I've compiled a 3.17.3 kernel with the newly added generic config in -current. What I noticed was that it did not find the modules directory, complaining that on line 417 a binary operator was expected or so. This command: Code:
strings /boot/vmlinuz-huge-3.14.24 | grep '(.*@.*) #' Code:
3.14.24 (root@hive64) #2 SMP Fri Nov 14 13:08:43 CST 2014 Code:
3.17.3 (root@aristipp) #1 SMP PREEMPT Mon Nov 17 11:37:31 CET 2014 Code:
strings /boot/vmlinuz-generic-3.17.3 | grep '(.*@.*) #' | cut -f1 -d' ' |
Hi,
I've done the same kernel building as you for my brand new laptop. Except that I've used the .config that Pat made for the huge kernel. (Yeah, I know... but it works) Perhaps appending a custom string to the kernel name would have prevented mkinitrd_command_generator.sh to fail. Also, if you don't use some specific configuration like LUKS and/or LVM on your /boot and /root partitions, then you can use the mkinird command manually. Here the command I've used : mkinitrd -c -k 3.17.3 -m ext4 That's my first kernel compile so far. :D I had to do it in order to have the following recognized : - Proper Pstate CPU governor driver for my Haswell CPU - Recognition of my ALPS touchpad - Recognition of my Logitech Unifying dongle |
If you use huge config, then you don't need initrd anymore
|
Is that strictly true? I seem to recall that when I've used huge in the past I still needed an initrd if I used LVM. Maybe that's no longer the case, but I'm pretty sure it was as of 13 or so.
|
It seemed to be a problem just on my machine, at least the 32 bit kernel I compiled on my T60 (generic-smp 3.17.3) didn't have this problem. As for the example (kernel huge): I don't have kernel-generic installed, instead I always build a newer generic and leave kernel-huge installed just in case. However, I use UUIDs with lilo, and I couldn't boot the huge kernel without an initrd.
|
Quote:
|
All times are GMT -5. The time now is 07:29 AM. |