[Addressed] Compiled Kernel lacking features

Hi,
A while ago I was playing around with kernel compilation for a system that needed a custom kernel to enable deprecated features: namely UMS instead of KMS. I got a tarball from Linux's github release page (mainline), ran some make commands incl. menuconfig and got a working binary that I put into /boot. This build continues to work fine, but I had no idea what I was doing so I don't think i have headers etc where they are.
AUR has linux-mainline I later discovered. Edited the PKGBUILD to use menuconfig which I am familiar with, then had no issues until I tried to load the kernel.
First issue: no ext4 support. I had to change it to built-in to the kernel to be able to mount my root partition. Odd, hadn't had to do that before but it's working now.
But, UMS doesn't seem to load at all, ACPI complains about features missing (thinkpad_acpi was compiled into kernel, I checked). dmesg and grepping for phrases just shows nothing, neither of these things seems to work or attempt to load at all. If I try and startx, it fails to find screens because KMS isn't going.
I will try linux-git which I've succesfully done before, but it takes even longer than mainline because of the checkout time (Thinkpad T60 is getting old).
Any advice on getting the linux-mainline AUR compilation working as expected?
Thanks
Last edited by deags (2014-01-19 15:28:38)

Good job, you've made a liar out of me
I went and checked everything, it seems it's now running with KMS successfully. I've had to run with nomodeset on all distros for then to boot, but noticed one day that dmesg had a line saying UMS couldn't be started hence I started using it and things worked so I kept with it.
Unless this was a bug in the kernel that has been fixed recently, this is extremely strange. If it returns I'll repost with more details, but it seems for now I can join the KMS 99%
Thanks

Similar Messages

Does keynote on iPad lacks features as compared to keynote on Mac ?

Does keynote on iPad lacks features as compared to keynote on Mac ?

s,
Here's the address of the iOS Keynote discussion where you may have more luck finding an answer: https://discussions.apple.com/community/app_store/iwork_for_ios#
A good percentage of us in the Mac iWork discussions don't have the iOS device version.
Jerry

A PKGBUILD that helps you compile kernel from local source tree

I don't know if someone did this beofore. Hours ago I wrote a PKGBUILD file for compiling kernel,
it is different than the one from abs. It allows you
compile a kernel from a exiting kernel source tree and leave it clean.
honor the Archway, this means you have a clean filesystem
It is acutally because I'm currently playing with The Eudyptula Challenge.
and I'm tied our compress/decompress a kernel tree all the time. If you are kernel developer, you
may also find it useful.
The PKGBUILD file worked on my machine, I will add headers and docs later.
Oh, almost forgot: here is my PKGBUILD:
#So we will have a clean src tree
pkgbase=linux-test
_kernel_bin=kernel_build
#the variable you have to provide
_builddir=kernel_build
kernel_src_dir='/home/developer/Courses/kernel-base'
_srcname=kernel_tree
#end the variable you have to provide
pkgver=3.8.1
pkgrel=1
pkgdesc="The Linux kernel and modules"
depends=('coreutils' 'linux-firmware' 'kmod' 'mkinitcpio>=0.7')
makedepends=('xmlto' 'docbook-xsl' 'kmod' 'inetutils' 'bc')
optdepends=('crda: to set the correct wireless channels of your country')
provides=("kernel26${_kernelname}=${pkgver}")
conflicts=("kernel26${_kernelname}")
replaces=("kernel26${_kernelname}")
arch=('i686' 'x86_64')
url="http://www.kernel.org/"
license=('GPL2')
source=(#if we provide this, means kernel compile progress is already done
"${_kernel_bin}.tar.xz"
'linux.preset'
sha256sums=('65847bc847344434657db729d2dde4a408e303ea29ae1409520cecee8da6fc3d'
'2c2e8428e2281babcaf542e246c2b63dea599abb7ae086fa482081580f108a98')
#this one strip the linux off
_kernelname=${pkgbase#linux}
prepare() {
#XXX:checked
#build dir has to be the same as kernel_bin files, then builddir is created
#automatically by tar
if [ "${kernel_src_dir}" == "" ];then
return 1
fi
#provide kernel source tree for compile and move modules
ln -s ${kernel_src_dir} ${srcdir}/${_srcname}
mkdir -p "${srcdir}/${_srcname}"
#we need to check here if there exist kernel bin files
if [ "${_kernel_bin}" == "" ]; then
make O="${srcdir}/${_builddir}" menuconfig
fi
build() {
#XXX:checked
cd "${srcdir}/${_srcname}"
#we need to check here if there exist kernel bin files
if [ "${_kernel_bin}" == "" ]; then
#return 1
make O="${srcdir}/${_builddir}" bzImage modules
fi
#otherwise this step is done already done
_package() {
#we dont need to worry about mkinitcpio, depmod thing, They are done by
#install script, we need to provide a preset and install file instead.
#we build kernel objs on _builddir, and install them in pkgdir
#install binary files, this means we have a compiled binary tree
cd "${srcdir}/${_srcname}"
#echo "$(pwd)"
KARCH=x86
install=linux.install
# get kernel version
_kernver="$(make O="${srcdir}/${_builddir}" kernelrelease)"
_kernver=$(echo "${_kernver}" | sed -n 2p -)
#strip the -dirty away
_kernver=${_kernver%-*}
_basekernel=${_kernver%%-*}
_basekernel=${_basekernel%.*}
mkdir -p "${pkgdir}"/{lib/modules,lib/firmware,boot}
make O="${srcdir}/${_builddir}" INSTALL_MOD_PATH="${pkgdir}" modules_install
cp "${srcdir}/${_builddir}"/arch/$KARCH/boot/bzImage "${pkgdir}/boot/vmlinuz-${pkgbase}"
# set correct depmod command for install
cp -f "${startdir}/${install}" "${startdir}/${install}.pkg"
true && install=${install}.pkg
sed -e "s/KERNEL_NAME=.*/KERNEL_NAME=${_kernelname}/" -i "${startdir}/${install}"
sed "s/KERNEL_VERSION=.*/KERNEL_VERSION=${_kernver}/" -i "${startdir}/${install}"
# install mkinitcpio preset file for kernel
install -D -m644 "${srcdir}/linux.preset" "${pkgdir}/etc/mkinitcpio.d/${pkgbase}.preset"
sed \
-e "1s|'linux.*'|'${pkgbase}'|" \
-e "s|ALL_kver=.*|ALL_kver=\"/boot/vmlinuz-${pkgbase}\"|" \
-e "s|default_image=.*|default_image=\"/boot/initramfs-${pkgbase}.img\"|" \
-i "${pkgdir}/etc/mkinitcpio.d/${pkgbase}.preset"
# remove build and source links
rm -f "${pkgdir}"/lib/modules/${_kernver}/{source,build}
# remove the firmware
rm -rf "${pkgdir}/lib/firmware"
# gzip -9 all modules to save 100MB of space
find "${pkgdir}" -name '*.ko' -exec gzip -9 {} \;
# make room for external modules
ln -s "../extramodules-${_basekernel}${_kernelname:--ARCH}" "${pkgdir}/lib/modules/${_kernver}/extramodules"
# add real version for building modules and running depmod from post_install/upgrade
mkdir -p "${pkgdir}/lib/modules/extramodules-${_basekernel}${_kernelname:--ARCH}"
echo "${_kernver}" > "${pkgdir}/lib/modules/extramodules-${_basekernel}${_kernelname:--ARCH}/version"
# Now we call depmod...
#echo "Call Depmod"
cp "${srcdir}/${_builddir}/System.map" System.map
depmod -b "${pkgdir}" -F System.map "${_kernver}"
#echo "Called Depmod"
# move module tree /lib -> /usr/lib
mkdir -p "${pkgdir}/usr"
mv "${pkgdir}/lib" "${pkgdir}/usr/"
# add vmlinux
install -D -m644 "${srcdir}/${_builddir}/"vmlinux "${pkgdir}/usr/lib/modules/${_kernver}/build/vmlinux"
pkgname=("${pkgbase}")
for _p in ${pkgname[@]}; do
eval "package_${_p}() {
_package${_p#${pkgbase}}
done
and here is the address of it on github
Last edited by xedchou (2014-12-23 12:41:55)

Based on the title alone I almost reflexively binned this thread. Please rename this thread to *something* relating to what you're actually posting.

How to edit GRUB for boot new compile kernel?

I compile the newest Linux kernel.
I read Archlinux wiki about how to edit GRUB for boot new compile kernel,
but Archlinux wiki do not reference to it.
Who can help me, baby?

bangkok_manouel wrote:
how did you build your kernel? traditional way or arch way?
if it's the arch way (TM), here's an example of a custom kernel with -mm patch.
what I have in /boot:
-rw-r--r-- 1 root root 4.3M 2008-12-23 09:52 kernel26mm-fallback.img
-rw-r--r-- 1 root root 668K 2008-12-23 09:52 kernel26mm.img
-rw-r--r-- 1 root root 764K 2008-12-22 18:44 System.map26mm
-rw-r--r-- 1 root root 1.7M 2008-12-22 18:44 vmlinuz26mm
now here's the /boot/grub/menu.lst related entry (you may add the fallback one):
# (2) Arch Linux
title kernel26mm
root (hd0,0)
kernel /vmlinuz26mm root=/dev/sda5 ro vga=773
initrd /kernel26mm.img
Should be traditional way.

[SOLVED]Compiling kernel fails with segfault

Hi guys,
I'm out of answers.
The situation :
Compiling kernel randomly stops with : "segmentation fault" and never runs through.
Reissuing make will make it continue until it will fail at some other point.
Background:
The system feels completely unstable to say the least, but nothing 100 % reproduceable ( vuze(java) crash, firefox/chromuim flash sites crash every 120 seconds, X crashes once a day , for instance when moving tvtime around, kernel oops daily, sometimes hard lockups, logs of yesterdays oops : http://pastie.org/private/imrqdzdbejdb3jxn4czuw )
But only the kernel compile is 100 % reproduceable so this must be fixed asap.
Specs:
intel e8400,nvidia gtx260, 2 G ddr2,Gigabyte EP45-UD3R, "gigs of space", off. arch kernel 2.6.33-ARCH , x86_64, "bigtime" cooling :
cpu temps 50 C° max during compiling
What configs/packages are important ? I got paranoid and ditched my old makgepkg.conf and used a newer *.pacnew file , I used to have :MAKEFLAGS="-j3" but not in this config :
http://pastie.org/private/4ryshivi2hgyqenjlkicg
local packages:
local/gcc 4.5.0-1 (base-devel)
The GNU Compiler Collection
local/gcc-libs 4.5.0-1 (base)
Runtime libraries shipped by GCC for C and C++ languages
Anything else updated daily.
The blowup:
http://pastie.org/private/hoamev0sli7zapbr0xvpbq
What I tried:
fsck -f: it fixed two inodes, but still the same behavior
badblocks -nvs : three hours of badblock testing, 0 hits
memtest : two passed runs, no errors
stress from AUR, 15 minutes :
stress --cpu 8 --io 4 --vm 2 --vm-bytes 128M --timeout 15m
stress: info: [26841] dispatching hogs: 8 cpu, 4 io, 2 vm, 0 hdd
stress: info: [26841] successful run completed in 900s
Max coretemp was 52 C° during stress.
The usual footnote statement ^_^ :
On w7 _everything_ runs stable.
And not only stable , I'm able to OC this box to 4.5 GHz and game under w7 without a prop for hours, well this is what it was build for at some point.
Conclusion:
So I'd say hardware wise this box kicks ass and now software wise my Arch kicks me in the a**.
I doubt there is a faulty hardware part, but what else could it be ?
Last edited by tuxfusion (2010-05-17 22:48:06)

Update:
Trying to compile another big package just to have another result, gcc itself now :
- segfault after 4 m18 s
- segfault after 5 m50 s
Update²:
I think I found a faulty BIOS setting in my non-OC profile concerning DRAM termination, which said "normal" where "auto" would be correct.
My OC-profile has this setting correct and since I run this profile in Arch only it might explain that w7 did not fail.
Still have to verify this but i makes perfect sense, only gcc seems to trigger this. Compiling gcc still running with over 10 minutes in the game now.
After hitting 4.5 GHz with OC I must have reset the non OC-profile wrong.
*slams head on desk*
Last edited by tuxfusion (2010-05-17 13:10:04)

[SOLVED] error compiling kernel 2.6.29

Hi,
Due to having the DVD-S card mentioned here
http://www.linuxtv.org/pipermail/linux- … 23265.html
I am trying to compile kernel 2.6.29.6 with the above mentioned patch but get an error:
include/sound/soc-dai.h: At top level:
include/sound/soc-dai.h:224:25: error: duplicate member ‘codec’
sound/soc/soc-core.c: In function ‘snd_soc_init_card’:
sound/soc/soc-core.c:1360:18: warning: variable ‘ac97’ set but not used
[-Wunused-but-set-variable]
make[2]: *** [sound/soc/soc-core.o] Error 1
make[1]: *** [sound/soc] Error 2
make: *** [sound] Error 2
bash: vim include/sound/soc-dai.h
/* DAI runtime info */
struct snd_pcm_runtime *runtime;
struct snd_soc_codec *codec;
unsigned int active;
unsigned char pop_wait:1;
void *dma_data;
/* DAI private data */
void *private_data;
/* parent codec/platform */
union {
struct snd_soc_codec *codec;
struct snd_soc_platform *platform;
Note:
I found a solution:
http://mailman.alsa-project.org/piperma … 16949.html
Last edited by casacristo (2011-10-19 07:46:01)

I've solved the issue, infact there was a problem with the patch. Now I am able to compile the kernel.
Thanks anyway!

Preferred way to compile kernel?

After the discussion here:
http://bbs.archlinux.org/viewtopic.php?t=4057&start=30
I'm wondering which way most people use for compiling their kernel. I've used /usr/src because its simpler and makes me feel more in control. Having an Arch package for the custom kernel has no real value, according to apeiro. Plus, using /usr/src has the kernel headers in place automatically, useful for programs that need it. (This offset by the fact that the kernel package copies the necessary files into place anyway). However, using ABS has the advantage of automating a lot of the task using Arch systems, rather than, for example, a custom shell script. (I'm trying an ABS version right now, for comparison).
Both methods are described in the wiki. I'm going to be cleaning up and expanding both shortly.
I'm asking this question in order to semi-standardize the process. Some people might say that the Arch philosophy is "There is more than one way to do everything", but I believe we could also append "but there is a best way to do it and Arch suggests that way". Now I want to find out which way is best in this case.
I'd be most interested in "I use xyz method because... [valid reason]" statements, rather than just "I use xyz".
Dusty

The 'abs' way was basically my idea - or at least the first wiki entry about it, I don't know if anyone tried it that way before. I thougt it could be a nice automatism, to have a proper kernel configuration once, tweak the PGKBUILD file, and after all is set up, all I had to change was the version number of the package/kernel source or change the config, and run makepkg then. I once made it switching from kernel24 to 26, using my own config. It worked fine. So, I thought it could be helpfull writing it to the wiki.
But this idea has some disadvantages, too. F.e., if you compile a minor upgrade, you may construct a modules mess, as a lot of files are already existing. Therefore, all cases must be thoroughly considered and configurated. Kernel addicts would prefer to patch the kernel instead of fully refetch the source.
After playing around with the idea for a while, I decided the work of configurating all possible eventuallities in compiling kernels by using abs was to difficult. Using the standard procedures seemed straight and easy to me.
So I removed my 'arch way' section in the kernel compiling wiki. Nevertheless, another user started an abs kernel wiki, redevelopping and enhancing my idea.
The idea is still attractive.

MAC Address-Table Move Update Feature

Hi guys
Does 6500 SUP720/2T support MAC Address-Table Move Update Feature?
I cannot find it in anywhere..
Thanks very much!
QXZ

Hi,
Please refer following link :
http://www.cisco.com/en/US/products/hw/switches/ps708/products_tech_note09186a00807347ab.shtml
CAM
—All Catalyst switch models use a CAM table for Layer 2 switching. As frames arrive on switch ports, the source MAC addresses are learned and recorded in the CAM table. The port of arrival and the VLAN are both recorded in the table, along with a timestamp. If a MAC address learned on one switch port has moved to a different port, the MAC address and timestamp are recorded for the most recent arrival port. Then, the previous entry is deleted. If a MAC address is found already present in the table for the correct arrival port, only its timestamp is updated.
regards,
Ajay Kumar

Update modules with compiling kernel from soure issue.

Compiling kernel from abs or aur take me much time, so i decide compile kernel from source, compilation and installation successfully but make initial RAM disk doesn't update full modules (crypto,lvm...). I specified -c option to /etc/mkinitcpio.conf but the result is the sam, non-autodectec doesn't help too. When i look into /lib/modules directory, 3.6.6-1-ARCH folder contain "kernel" folder which has a lot of things (arch, crypto, net..) and i think this is my extra modules were declared in HOOK array's mkinitcpio.conf. The another one only contain net folder so i can't boot into this kernel without having loaded others.
any suggestion for this problem? tks for reply.

am sorry, i re-tried with HOOK array without autodectect and it can load modules but there are two error:
==> ERROR: module not found: `dm-crypt`
==> ERROR: modules not found: `dm-snapshot`
ok i will try to load these modules by manually.
EDIT: i tried to load dm-mod, dm-crypt, dm-snapshot but it isn't helpful. Errors still remain. any suggestion?
Last edited by angelfalls (2012-11-12 23:33:53)

Compiling kernel module

Hello. I tryed to compile kernel module
/home/das/job/C/foo.c:
#include <linux/module.h>
#include <linux/kernel.h>
int init_module(void)
printk(KERN_INFO "Hello world 1.\n");
return 0;
void cleanup_module(void)
printk(KERN_INFO "Goodbye world 1.\n");
I created /home/das/job/C/Makefile:
obj-m += foo.o
all:
make -C /lib/modules/$(shell uname -r)/build M=$(PWD) modules
clean:
make -C /lib/modules/$(shell uname -r)/build M=$(PWD) clean
But when i tryed to compile it with command make I take the error:
make -C /lib/modules/2.6.33-ARCH/build M=/home/das/job/C modules
make: *** /lib/modules/2.6.33-ARCH/build: No such file or directory. Stop.
make: *** [all] Error 2
My question is what is wrong and how i can compile module?
Last edited by F1sher (2011-11-01 15:37:51)

tomk wrote:Install it somewhere under /lib/modules/2.6.32-lts, run depmod, and modprobe it.
After
depmod -a
modprobe module
I can't start my system. GDM start but keyboard and mice don't work

Compile Ctrl-F7 feature unavailble for Driver projects in Visual Studio 2013

Is there a way to compile individual files in Driver projects?
I'm running Visual Studio Professional 2013 Version 12.0.21005.1 REL on Windows 7 SP1 x64
This happens with any driver project but the easiest way to repro is to look at one of the samples.
Open up Windows Driver Kit (WDK) 8.1 Samples.zip\IOCTL\C++\ioctl.sln
In the Solution Explorer pane - Right click on Exe\ioctlapp\Source Files\testapp.c - The Compile Ctrl-F7 feature is available
If you right click on Sys\sioctl\Source Files\sioctl.c - The Compile Ctrl-F7 feature is greyed out

Thanks for Viorel’s help.
Hi vvx123,
Not the real expert about the Driver projects, maybe you could share me the requirement tools or the detailed steps, so I could repro this issue in my VS2013.
But if you mean that other projects works well, and all driver projects have this issue, I doubt that it has this limitation for this specific project. Maybe you could check it in other machine.
To get the issue confirmed and diagnose by product team, if possible, you could create connect report for it. You will get email notification for update.http://connect.microsoft.com/VisualStudio/feedback/CreateFeedback.aspx.
If you submit it, you could share us the link here, so could get the latest information from the product team expert.
I have the VS2013 Environment in Windows 8, but it seems that I couldn’t get the driver project template in VS 2013 IDE, do I need to install any tools like SDK or others? Or how I could create a driver project in my side. Could you share me the detailed
steps to repro this issue?
If there's any concern, please feel free to let me know.
Best Regards,
We are trying to better understand customer views on social support experience, so your participation in this interview project would be greatly appreciated if you have time. Thanks for helping make community forums a great place.
Click
HERE to participate the survey.

Cross compile kernel

I have compiled a few kernels in an environment in which it's designed to run on - but cross compiling is completely new to me. However - I will persevere because I have wanted to learn this for a while now.
The target system is an arm based board for a NAS. I am using QEMU to install and configure a Debian system but it requires a working kernel in order to boot.
I have installed the arm-elf-gcc-base package (which I assume is the toolchain - am I wrong on this?) but I don't know where to go from there.
How do I invoke this particular toolchain to compile a kernel for the target arch?
Any other pointers or 'gotchas' would be greatly appreciated.
Thank you.

Which board is it?
Even if you manage to cross compile, kernel will need some extra configuration or patching to boot in qemu.
I have Raspberry Pi and qemu needs custom kernel to boot RPi images, but it's almost useless since there is
no support for network adapter. I have never cross compiled anything for it, but you might want to read on RPi
kernel cross compilation since there is a lot of documentation and you probably need just a different toolchain.
What I'm doing is distributed cross compiling via distcc. That way most of work gets done on my laptop, but it's
still quite slow because makepkg doesn't support distcc pump for distributing pre-processing.
I'm using toolchain provided by Arch Linux ARM project because I run Arch on RPi. If you can find crosstool config for
your board, making toolchain shouldn't be too difficult. This should get you started.

Cant boot a compiled kernel

Hi,
I want to use Arch to mess with the source of linux 3.2.7. Unfortunately, I cannot even get the regular 3.2.7 kernel to boot.
I basically followed the guide at:
https://wiki.archlinux.org/index.php/Ke … raditional
I downloaded the source, did make mrproper, make menuconfig, make all and make modules_install. I copied arch/x86/bzImage to /boot as well as System.map and created a cpio ramdisk using mkinitcpio using
mkinitcpio -g /boot/initramfs-test.img -k 3.2.7-test
In the end my /boot contained:
initramfs-test.img
vmlinuz-test
System.map-test
I tried booting with Syslinux, GRUB and GRUB2 modifying the options for all of them appropriately.
For Syslinux and GRUB if use a UUID in the root field for their respective config files I get the error that the root device cannot be found during boot. If I use root=/dev/sda4 (which is my root partition, checked it using stat /dev/sda4) I successfully mount the root partition but get an error that /sbin/init does not exist. That's weird since /sbin/init is there. I believe both of these errors have something to do with ramdisk but not sure what.
With GRUB2 booting just freezes at "Loading init ramdisk".
I read a ton on these errors online and tried all kinds of stuff and could not get it to work. Any help will be greatly appreciated!
Thanks

Thanks tomk and sorry for the ambiguous post.
I actually managed to get it finally running, I basically screwed up my .config file. This was not obvious to me at all (total noob at compiling kernels). I just used my current arch config file instead of messing with menuconfig myself and it worked. I still don't know why it works, so if you could point me to a source of enlightenment I would appreciate that. Right now compilation time is atrocious even with the -j option for make.
Regarding your question, I want to screw around with the latest pax patch and add some hooks of my own to security.h which is why I need 3.2.7.
Thanks again

FF4 sort of works in Win 98SE via Kernel Ex but slow and lacks features

I had last version of FF2 working fine with Win 98SE but could not access features using java, Flash, etc.; I installed Kernel EX and FF4, after which it seems to work but is slow as molasses...takes more than a minute to load home page or other link, after which it won't scroll for 10 to 30 seconds. There is no history or bookmarks and the back/forward buttons do not work consistently.
It works well enough to be promising, but frustrating; is there any hope? Perhaps missing system files from Win XP? Any help greatly appreciated...I really like Firefox!

Sorry, your operating system has been officially obsolete since July 11, 2006. It's likely that websites require more recent versions of Flash Player and Java than you can run.
* [http://kb2.adobe.com/cps/406/kb406791.html Flash Player 9 for older operating systems]
* [http://java.com/en/download/faq/other_jreversions.xml Where can I get older versions of Java and related documentation?]
You can try starting Firefox in a new profile and see if that helps any. But it's likely that these problems are caused by running Firefox on an unsupported operating system with the aid of third-party software. Having an overclocked CPU doesn't help matters either.
* [[Managing profiles]]
If things aren't any better in a new profile, you can try Firefox 3.6.28 instead. This version is now obsolete and insecure, but that seems a bit of a moot point in your case.
* [http://www.mozilla.org/firefox/all-older.html www.mozilla.org/firefox/all-older.html]
If you're unable to upgrade your hardware, you may want to consider switching from Windows to Linux. You would have the benefit of an up-to-date and secure operating system, browser and plug-ins at no cost.
* [http://lubuntu.net/ http://lubuntu.net/]

Recompile a compiled kernel with ABS

I had been compiling my own kernel with the traditional method for ages. Sometimes ago, I decided to try the ABS method, modifying properly a PKGBUILD for the kernels provided by arch. I now appreciate the provided automation and the ease in combining arch patches with the patches I need or like.
After a deep personalization of the PKGBUILD, there is now only a feature of the old method I continue to miss. It is when I decide to change something in the configuration of a kernel I have already compiled, and I want to recompile it. If I do so with the traditional method, the previous compilation is somewaht "cached" and only what is actually different is compiled again: thus, the recompilation time is apparently shorter than the time needed by the first compilation.
I am not able to do the same with the ABS method. It seems that the directory with the kernel source into src/ is rewritten every time (i.e., the kernel package is untarred also when is already there). Moreover, I also need to delete completely src/ or part of his content, otherwise makepkg complains about patches already being there. Is there a way to avoid this behaviour, so to shorten the recompilation time?

Yes, ccache drastically reduces recompilation times (for my kernel, from 30 min to 5 min). However, when you change just a couple of peripheric things in config, 5 minutes are too much. Anyway, thanks a lot.

[Addressed] Compiled Kernel lacking features

Similar Messages

Maybe you are looking for