Porteus Kernel Builder

[beny]

hi roadie, you have to remove the old kernel version patch before to apply the new one,just for the record
patch -R -p1 < ../patch-x.y.z

[fulalas]

@roadie, could you upload your 6.6.x kernel module so I can test it?

Thanks!

[roadie]

@roadie, could you upload your 6.6.x kernel module so I can test it?

Thanks!

@fulalas,

Here's the link, as you'll see, it's what might be called a sparse kernel. I think it may be of no use to you as it's built specifically for AMD machines, all Intel has been removed.

https://www.mediafire.com/file/xr6vhchn ... ar.gz/file

[fulalas]

@roadie, thanks but it didn't work since I have an Intel machine.

Anyway, I managed to make everything work with this code:

Code: Select all

rm ../aufs_sources/tmpfs-idr.patch # this patch isn't useful
cp -r ../aufs_sources/{fs,Documentation} .
cp ../aufs_sources/include/uapi/linux/aufs_type.h include/uapi/linux
for i in ../aufs_sources/*.patch; do
	patch -N -p1 < "$i" > /dev/null 2>&1 || { echo "Failed to add AUFS patch '${i}'."; exit 1; }
done

[Blaze]

[Updated kernel Examples]
==== AUFS Kernel only for Porteus ====
Simple package (vmlinuz, 000-kernel.xzm, 06-crippled_sources-NNN-XXbit.xzm, 64bit.config)

6.6.3 <-- NEW : "All patches" patching was done.
64bit-ALL-kernel6.6.3.tar (~156 M)
https://www.mediafire.com/file/x09ttty3 ... l6.6.3.tar
5c60f6528ecb309490530c47c4cb51e3 64bit-ALL-kernel6.6.3.tar

Note 1: Compiler that was used.
for kernel 6.5.1: 11.2.0-x86_64

Note 2: Compilated under Porteus 5.01 (Slackware 15.0 base. Note. glibc-2.33 replaced on glibc-2.35)

Note 3: A new cryptsetup (version 2.3.5: presented by @ncmprhnsbl)

Note 4: ZSTD compression

Thanks to neko for support.

[peebee]

There seems to be an aufs problem with kernel 6.6.4
https://github.com/sfjro/aufs-standalone/issues/32

[beny]

hi blaze the 6.6.4 kernel also with the kernel builder fail to add the devices to union, operation not permitted, but the build task is ok just for the record.
the message is mkdir can't create directory union operation not permitted

[beny]

hi, so thanks to roadie we have finally the aufs kernel work:
Linux porteux.example.net 6.6.5-porteux #1 SMP PREEMPT_DYNAMIC Fri Dec 8 19:31:49 CET 2023 x86_64 AMD Ryzen 5 3500X 6-Core Processor AuthenticAMD GNU/Linux

[roadie]

If anyone wants a bit easier method until the sfjro patch set is ready, here's a patch I made. Just copy, paste and name it whatever. I'm not sure how the Kernel Builder handles patches, I use my own building script so it's just included with the sfjro patches. Tested on kernel-6.6.4 and kernel-6.6.5......works on 6.7-rc4 too, but aufs6-mmap doesn't yet.

Code: Select all

--- a/fs/stat.c	2023-12-07 23:52:25.000000000 -0800
+++ b/fs/stat.c	2023-10-29 19:31:08.000000000 -0700
@@ -133,8 +133,7 @@
 	idmap = mnt_idmap(path->mnt);
 	if (inode->i_op->getattr)
 		return inode->i_op->getattr(idmap, path, stat,
-					    request_mask,
-					    query_flags | AT_GETATTR_NOSEC);
+					    request_mask, query_flags);
 
 	generic_fillattr(idmap, request_mask, inode, stat);
 	return 0;
@@ -167,9 +166,6 @@
 {
 	int retval;
 
-	if (WARN_ON_ONCE(query_flags & AT_GETATTR_NOSEC))
-		return -EPERM;
-
 	retval = security_inode_getattr(path);
 	if (retval)
 		return retval;

[rchase]

Blaze, could you spell out how to use your 6.6.3 kernel? I tried using it on my N100-equipped mini PC, but still couldn't boot with graphical support; I understand that this CPU was supported as of kernel 6.2 -- bluetooth and wireless have apparently been problematic. Where does the .config file go?

[Blaze]

Hi rchase.
What is CPU of your mini PC (Intel Alder Lake N100)?
Can you share these logs files

• Xorg.0.log
• dmesg

in /var/log ?

The config file 64bit.config I put to tarball.
If you have activated 06-crippled_sources module you can look at /usr/src/linux/.config

[beny]

hi blaze take a look into the web page there is the device listed:
https://cateee.net/lkddb/web-lkddb/DRM_I915.html

[roadie]

I've been playing with kernel-6.6.5 without patching fs/stat.c and I've managed to make it boot cleanly by adding

Code: Select all

udba=none

to the mount options in linuxrc in the "Setup aufs" section.

In this post EXIT cheatcode/multiple saves... fanthom mentions that this option is already used in Porteus-v2.0 I believe.

I don't know if there might be problems doing this, though I wonder why it was removed in later initrd's. I haven't seen problems so far.

[roadie]

I haven't seen any issues using the "udba=none" option in the linuxrc aufs mount. I tried "udba=reval", and it failed to boot, same as the unpatched kernel with stock initrd.xz. I also tried "udba=notify" which also failed to boot....unable to mount union.

Since the post I linked to mentioned "changes=EXIT:", I tried it, and it seems to work as designed. I have little experience with that form of saving. But, /mnt/live/memory/changes looks right, and changes are copied to the folder I specified. Shutdown and reboots are as normal, fsck shows clean partitions. It works in PorteuX as well, though I didn't expect anything else.

[fulalas]

@roadie, thanks for looking at this!

I'm testing udba=none on kernel 6.6.1 and it seems to be working quite well, indeed.

Only 2 things to consider:
1. I couldn't find this flag in any Porteus I tested (old ISOs here: http://ftp.riken.jp/Linux/porteus/x86_6 ... .0-x86_64/). I wonder why it didn't go to any final release.
2. the official documentation isn't exactly clear (to my limited knowledge) what are the practical drawbacks of using this flag:

Aufs trusts the dentry and the inode cache on the system, and never test about UDBA. With this option, aufs runs fastest, but it may show you incorrect data. Additionally, if you often modify a branch directly, aufs will not be able to trace the changes of inodes on the branch. It can be a cause of wrong behaviour, deadlock or anything else.

It is recommended to use this option only when you are sure that nobody access a file on a branch. It might be difficult for you to achieve real ’no UDBA’ world when you cannot stop your users doing ’find / −ls’ or something.