gentoo-sources-5.1.0 ebuild question

Message

adcdam · Post by **adcdam** » Tue May 07, 2019 5:02 am

Hi, i installed kernel 5.1.0 but when i did menuconfig i saw that i can not choose zen cpu, i saw the ebuild
and it doesnt have experimental like previous kernels why is that?

Code: Select all

# Copyright 1999-2019 Gentoo Authors
# Distributed under the terms of the GNU General Public License v2

[code]EAPI="6"
ETYPE="sources"
K_WANT_GENPATCHES="base extras"
K_GENPATCHES_VER="1"

inherit kernel-2
detect_version
detect_arch

KEYWORDS="~alpha ~amd64 ~arm ~arm64 ~hppa ~ia64 ~mips ~ppc ~ppc64 ~s390 ~sh ~sparc ~x86"
HOMEPAGE="https://dev.gentoo.org/~mpagano/genpatches"
IUSE=""

DESCRIPTION="Full sources including the Gentoo patchset for the ${KV_MAJOR}.${KV_MINOR} kernel tree"
SRC_URI="${KERNEL_URI} ${GENPATCHES_URI} ${ARCH_URI}"

pkg_postinst() {
	kernel-2_pkg_postinst
	einfo "For more info on this patchset, and how to report problems, see:"
	einfo "${HOMEPAGE}"
}

pkg_postrm() {
	kernel-2_pkg_postrm
}

[/code]

i have the experimental use flag in package.use for gentoo-sources but the ebuild seems to not have it.

Post by **fedeliallalinea** » Tue May 07, 2019 5:56 am

Code: Select all

IUSE=""

I don't know why but in 5.1 experimental use flag is removed

commit b6ca40c8f992d5a66e29d442b183bfa571c19f34
Author: Mike Pagano <mpagano@gentoo.org>
Date: Mon May 6 13:06:00 2019 -0400

sys-kernel/gentoo-sources: Remove use variable not used

Signed-off-by: Mike Pagano <mpagano@gentoo.org>
Package-Manager: Portage-2.3.62, Repoman-2.3.11

The Main Man · Post by **The Main Man** » Thu May 09, 2019 9:57 am

I really miss this, any news on this?
Is it permanent ?

wrc1944 · Post by **wrc1944** » Thu May 09, 2019 10:18 am

I PM'd Mike about this and he sent me the following:

From: wrc1944
To: mpagano
Posted: Wed May 08, 2019 3:05 am
Subject: gentoo-sources experimental gcc opts for 5.1.0 question
Mike,
Just noticed there's no experimental gcc opts patches as of yet for kernel-5.1.0 in gentoo-sources, or even at the graysky2 web site.
I assume there will eventually be one, and maybe some work needed for a gcc-9.1.0 version, correct?

Thanks for your time, and maintaining the really nice gentoo-sources kernel series!

From: mpagano
To: wrc1944
Posted: Wed May 08, 2019 4:10 pm
Subject: Re: gentoo-sources experimental gcc opts for 5.1.0 question Quote message
I hope so. When it get's posted, I'll add it.

I have added the gcc opts sections manually to the 3 relevant kernel patch files in the past on versions where there was a delay, and they worked normally. If it's not forthcoming shortly, I might look into it again.
Can't operate without these opts on my zen systems.

Wish this would be mainlined into the kernel.

Related thread on this with what I did last time, but not the time I did the 3 patch files manually- still looking for that thread as i forgot exactly which files I edited in the gcc opts stanzas: https://forums.gentoo.org/viewtopic-p-8 ... ml#8311488

The Main Man · Post by **The Main Man** » Thu May 09, 2019 10:39 am

wrc1944 wrote:I have added the gcc opts sections manually to the 3 relevant kernel patch files in the past on versions where there was a delay, and they worked normally. If it's not forthcoming shortly, I might look into it again.
Can't operate without these opts on my zen systems. Wish this would be mainlined into the kernel.

It's just these two patches, right ?

Code: Select all

5010_enable-additional-cpu-optimizations-for-gcc-4.9.patch
5011_enable-cpu-optimizations-for-gcc8.patch

wrc1944 · Post by **wrc1944** » Thu May 09, 2019 10:58 am

Maybe it's still the same- here's the places i looked before, but I'm on gcc-9.1.0 now, so all this might not work like last time with gcc-8.x.x versions. I'm no expert, and i suspect I just lucked out doing this in the past.
mpagano is the expert, and I suspect it will take more work than just doing a simple c/p hack of adding the gcc-opts sections into those three kernel source files like i did before. Will take another look into the source files and see if it looks like last time. Still, there might be gaa-9 factors involved beyond my knowledge.

https://github.com/graysky2/kernel_gcc_ ... 3%2B.patch

https://github.com/graysky2/kernel_gcc_patch

The Main Man · Post by **The Main Man** » Thu May 09, 2019 11:21 am

I'm on gcc 9 as well, and zen

I guess it's best to wait then.
Until that happens I changed Generic cpu to AMD.. and IO scheduler to BFQ

wrc1944 · Post by **wrc1944** » Thu May 09, 2019 11:30 am

As far as I can tell, this is the latest version posted on the graysky2 website: https://github.com/graysky2/kernel_gcc_ ... .13+.patch

It might apply OK to kernel-5.1.0, or perhaps if not just editing the zen sections into the 3 relevant 5.1.0 files might work. We do know gcc-9 has zen support, so why not?

Only way to find out at this point is to try both options. Otherwise, I guess we just wait for mpagano to figure it out with a new graysky2 version.

The Main Man · Post by **The Main Man** » Thu May 09, 2019 11:37 am

It applies fine

Code: Select all

# ebuild /usr/portage/sys-kernel/gentoo-sources/gentoo-sources-5.1.0.ebuild configure                                                                                                                                 
 * linux-5.1.tar.xz BLAKE2B SHA512 size ;-) ...                                                                                                                                                                                         [ ok ]
 * genpatches-5.1-1.base.tar.xz BLAKE2B SHA512 size ;-) ...                                                                                                                                                                             [ ok ]
 * genpatches-5.1-1.extras.tar.xz BLAKE2B SHA512 size ;-) ...                                                                                                                                                                           [ ok ]
 * checking ebuild checksums ;-) ...                                                                                                                                                                                                    [ ok ]
 * checking miscfile checksums ;-) ...                                                                                                                                                                                                  [ ok ]
>>> Preparing to unpack ...
>>> Unpacking source...
>>> Unpacking linux-5.1.tar.xz to /var/tmp/portage/sys-kernel/gentoo-sources-5.1.0/work
>>> Unpacking genpatches-5.1-1.base.tar.xz to /var/tmp/portage/sys-kernel/gentoo-sources-5.1.0/work/patches
>>> Unpacking genpatches-5.1-1.extras.tar.xz to /var/tmp/portage/sys-kernel/gentoo-sources-5.1.0/work/patches
 * Excluding Patch #5010_enable-additional-cpu-optimizations-for-gcc.patch ...                                                                                                                                                          [ ok ]
 * Excluding Patch #5010_enable-additional-cpu-optimizations-for-gcc-4.9.patch ...                                                                                                                                                      [ ok ]
 * Excluding Patch #5010_enable-additional-cpu-optimizations-for-gcc.patch ...                                                                                                                                                          [ ok ]
 * Excluding Patch #5010_enable-additional-cpu-optimizations-for-gcc-4.9.patch ...                                                                                                                                                      [ ok ]
 * Applying 1500_XATTR_USER_PREFIX.patch (-p1) ...                                                                                                                                                                                      [ ok ]
 * Applying 1510_fs-enable-link-security-restrictions-by-default.patch (-p1) ...                                                                                                                                                        [ ok ]
 * Applying 2500_usb-storage-Disable-UAS-on-JMicron-SATA-enclosure.patch (-p1) ...                                                                                                                                                      [ ok ]
 * Applying 2600_enable-key-swapping-for-apple-mac.patch (-p1) ...                                                                                                                                                                      [ ok ]
 * Applying 4567_distro-Gentoo-Kconfig.patch (-p1) ...                                                                                                                                                                                  [ ok ]
>>> Source unpacked in /var/tmp/portage/sys-kernel/gentoo-sources-5.1.0/work
>>> Preparing source in /var/tmp/portage/sys-kernel/gentoo-sources-5.1.0/work/linux-5.1.0-gentoo ...
 * Applying enable_additional_cpu_optimizations_for_gcc_v8.1+_kernel_v4.13+.patch ...                                                                                                                                                   [ ok ]
 * User patches applied.
>>> Source prepared.
>>> Configuring source in /var/tmp/portage/sys-kernel/gentoo-sources-5.1.0/work/linux-5.1.0-gentoo ...
>>> Source configured.

wrc1944 · Post by **wrc1944** » Thu May 09, 2019 11:47 am

Yeah- I just did this:

Code: Select all

gentoo-audio ~ # cd /usr/src/linux-5.1.0-gentoo
gentoo-audio /usr/src/linux-5.1.0-gentoo # patch -p1 < 5011_enable-cpu-optimizations-for-gcc8.patch
patching file arch/x86/Makefile_32.cpu
patching file arch/x86/Kconfig.cpu
patching file arch/x86/Makefile
patching file arch/x86/include/asm/module.h
gentoo-audio /usr/src/linux-5.1.0-gentoo #

Will now compile 5.1.0 using make xconfig to load my regular config file and see if the gcc-opts with zen are there. I expect they will be, and will work. will post back shortly.

EDIT: zen and "native are there with all the rest- I always use native instead of zen. Compiling normally right now

so the current 5011 patch also applies cleanly to 5.1.0, and compiles with gcc-9.1.0!

wrc1944 · Post by **wrc1944** » Thu May 09, 2019 12:00 pm

kajzer,
The extras patch excludes the gcc-opts. Use the 5011. Are you on ~amd64 or just amd64?

The Main Man · Post by **The Main Man** » Thu May 09, 2019 12:08 pm

~amd64

What do you mean 5011..... how? where is it?
I was trying something else prior to your post where you linked the latest patch.
I made custom ebuild in local overlay, but that's deleted now, so never mind that.

I patched the kernel with that patch like you did in /usr/src/linux, selected cpu option and compiled the kernel.
I missed something ?

edit:
I guess you mean the same patch but you named it 5011

wrc1944 · Post by **wrc1944** » Thu May 09, 2019 2:24 pm

Sorry- since you mentioned it I thought you must have already had it.
I got it here at bottom of page.
https://dev.gentoo.org/~mpagano/genpatc ... .0-14.html just place it in usr/src/linux-5.1.0-gentoo, and do a:

Code: Select all

gentoo-audio ~ # cd /usr/src/linux-5.1.0-gentoo
gentoo-audio /usr/src/linux-5.1.0-gentoo # patch -p1 < 5011_enable-cpu-optimizations-for-gcc8.patch

and compile kernel as usual.

OOPS-actually it's at top of page. download the experimental link genpatches-5.0-14.experimental.tar.xz somewhere and extract it, and you'll have the older 5010 and the 5011.
5.0-14 Patch List

Patches 0000-2999 are available in genpatches-5.0-14.base.tar.xz
Patches 3000-4999 are available in genpatches-5.0-14.extras.tar.xz
Patches 5000-5099 are available in genpatches-5.0-14.experimental.tar.xz

The Main Man · Post by **The Main Man** » Thu May 09, 2019 2:48 pm

I applied this one :
https://github.com/graysky2/kernel_gcc_ ... .13+.patch

I think it's the same as the one you're talking about located in genpatches-5.0-14.experimental.tar.xz
At least I think it is, you said that one was the latest on the graysky2 website

Anon-E-moose · Post by **Anon-E-moose** » Thu May 09, 2019 5:01 pm

I'm running 5.0.13 (a true zen kernel) w/gcc 8.2 and when I compile w/PILEDRIVER or Native (bdver2) it sends out tons of warning messages from objtool.

Has anyone compiled the 5.0/5.1 kernel with gcc 9 and selected PILEDRIVER or NATIVE and had it compile clean?

wrc1944 · Post by **wrc1944** » Thu May 09, 2019 5:59 pm

kajzer,
Yes, I too think it's the same one, and apparently mpagano renames it to 5011 in the Gentoo genpatches listing.

Anon-E-moose,
With gcc-9.1.0 the few 5.0 kernels I've compiled function fine, and I don't recall any warnings but wasn't really watching the entire compile.

My first 5.1.0 kernel has 4 warnings randomly spaced out. Haven't rebooted yet, but don't expect any problems. Even though the patch seemed to apply cleanly, I'm still wondering if unreported offsets could be in play here.

Code: Select all

In file included from arch/x86/kernel/head64.c:35:
In function ‘sanitize_boot_params’,
    inlined from ‘copy_bootdata’ at arch/x86/kernel/head64.c:389:2:
./arch/x86/include/asm/bootparam_utils.h:40:3: warning: ‘memset’ offset [197, 448] from the object at ‘boot_params’ is out of the bounds of referenced subobject ‘ext_ramdisk_image’ with type ‘unsigned int’ at offset 192 [-Warray-bounds]
   40 |   memset(&boot_params->ext_ramdisk_image, 0,
      |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   41 |          (char *)&boot_params->efi_info -
      |          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   42 |    (char *)&boot_params->ext_ramdisk_image);
      |    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./arch/x86/include/asm/bootparam_utils.h:43:3: warning: ‘memset’ offset [493, 497] from the object at ‘boot_params’ is out of the bounds of referenced subobject ‘kbd_status’ with type ‘unsigned char’ at offset 491 [-Warray-bounds]
   43 |   memset(&boot_params->kbd_status, 0,
      |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   44 |          (char *)&boot_params->hdr -
      |          ~~~~~~~~~~~~~~~~~~~~~~~~~~~
   45 |          (char *)&boot_params->kbd_status);
      |          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


In function ‘ftrace_dump’,
    inlined from ‘ftrace_dump’ at kernel/trace/trace.c:8553:6:
kernel/trace/trace.c:8631:3: warning: ‘memset’ offset [8377, 8432] from the object at ‘iter’ is out of the bounds of referenced subobject ‘seq’ with type ‘struct trace_seq’ at offset 4240 [-Warray-bounds]
 8631 |   memset(&iter.seq, 0,
      |   ^~~~~~~~~~~~~~~~~~~~
 8632 |          sizeof(struct trace_iterator) -
      |          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 8633 |          offsetof(struct trace_iterator, seq));
      |          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


awk: ./arch/x86/tools/gen-insn-attr-x86.awk:260: warning: regexp escape sequence `\:' is not a known regexp operator
awk: ./arch/x86/tools/gen-insn-attr-x86.awk:350: (FILENAME=./arch/x86/lib/x86-opcode-map.txt FNR=41) warning: regexp escape sequence `\&' is not a known regexp operator


ld: arch/x86/boot/compressed/head_64.o: warning: relocation in read-only section `.head.text'
ld: warning: creating a DT_TEXTREL in object

The Main Man · Post by **The Main Man** » Thu May 09, 2019 6:22 pm

Well, since 4.12 BFQ is in the kernel, experimental patches only include additional CPU options.
I still don't understand why the USE flag was removed from 5.1.0
Thinking about disabling it anyway and placing the patch in /etc/portage/patches, patch seems to be working fine for over a year now.

As for warnings I have '--quiet-build' so I don't see warning and compile messages.
Only when I compile the kernel, but that appears to be clean and w/o warnings.

wrc1944 · Post by **wrc1944** » Thu May 09, 2019 8:53 pm

As in previous versions, the removal of the USE flag is (was) temporary, as mpagano tends to wait until the graysky2 kernel patches people issue a new updated version for gcc and/or kernel updates.

I recall there were times when the older version wouldn't apply cleanly and had too many offsets making it difficult to manually adjust.

UPDATE: Rebooted into my new 5.1.0-gentoo compiled with gcc-9.1.0 and using the 5011 gcc-opts patch with MNATIVE set for my Ryzen 7 1700. All is normally functioning.

The Main Man · Post by **The Main Man** » Sat May 11, 2019 10:10 pm

experimental is back with 5.1.1

edit: just tried it and there's no additional cpu options !?
Looking the ebuild... just IUSE got back the flag but there's no 'experimental' in K_WANT_GENPATCHES nor there's genpatches-5.1.1-experimental file
Looks like the person maintaining this is having a great fun, good for him, I'm giving up on this flag most definitely now.
Ridiculous!

wrc1944 · Post by **wrc1944** » Sat May 11, 2019 11:22 pm

Yeah- I just got same result. When I went here https://dev.gentoo.org/~mpagano/genpatc ... 5.1-1.html ane tried to download the experimntal patch is says not found on this server, but the others were on the server.

Either Mike hasn't just gotten to it yet, or found some problems that needed work on 5011. Guess we can still try patching in 5011 ourselves, but I did have those 4 warnings with 5.1.0, even though 5.1.0 does seem to function OK on 2 of my Gentoo systems with gcc-9.1.0 and 5011.

EmTeeEmEl · Post by **EmTeeEmEl** » Sun May 12, 2019 1:42 am

It's back in 5.1.1

Code: Select all

5.1.1(5.1.1)^bs(08:52:39 05/12/19)(experimental -build -symlink)

Jaglover · Post by **Jaglover** » Sun May 12, 2019 1:50 am

There should have been a news item or ebuild message about this new experimental flag not doing what expected.

wrc1944 · Post by **wrc1944** » Sun May 12, 2019 4:52 am

It's still not there. All that is there in distfiles and ebuild are extras and base, or listed at https://dev.gentoo.org/~mpagano/genpatc ... 5.1-1.html

Just synced at 12:51 AM EDT USA

Still get:

Code: Select all

amd64 ~ # emerge gentoo-sources -pv

These are the packages that would be merged, in order:

Calculating dependencies... done!
[ebuild   R    ] sys-kernel/gentoo-sources-5.1.0:5.1.0::gentoo  USE="-build -symlink (-experimental%*)" 103,755 KiB

(That's a MINUS experimental FLAG)

EmTeeEmEl · Post by **EmTeeEmEl** » Sun May 12, 2019 9:43 am

@wrc1944

You still have 5.1.0

This is from my emerge:

Code: Select all

emerge gentoo-sources -pv 

These are the packages that would be merged, in order:

Calculating dependencies      ... done!                            
[ebuild   R   ~] sys-kernel/gentoo-sources-5.1.1:5.1.1::gentoo  USE="experimental -build -symlink" 0 KiB

The Main Man · Post by **The Main Man** » Sun May 12, 2019 10:52 am

wrc1944 wrote:Guess we can still try patching in 5011 ourselves, but I did have those 4 warnings with 5.1.0, even though 5.1.0 does seem to function OK on 2 of my Gentoo systems with gcc-9.1.0 and 5011.

Warnings are not related to experimental patching.
Wasn't lazy and I just tried to compile with and without cpu optimizations patch.
In both cases I get the same warning, and it's the only one in my case :

Code: Select all

In file included from arch/x86/kernel/head64.c:35:
In function ‘sanitize_boot_params’,
    inlined from ‘copy_bootdata’ at arch/x86/kernel/head64.c:389:2:
./arch/x86/include/asm/bootparam_utils.h:40:3: warning: ‘memset’ offset [197, 448] from the object at ‘boot_params’ is out of the bounds of referenced subobject ‘ext_ramdisk_image’ with type ‘unsigned int’ at offset 192 [-Warray-bounds]
   40 |   memset(&boot_params->ext_ramdisk_image, 0,
      |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   41 |          (char *)&boot_params->efi_info -
      |          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   42 |    (char *)&boot_params->ext_ramdisk_image);
      |    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./arch/x86/include/asm/bootparam_utils.h:43:3: warning: ‘memset’ offset [493, 497] from the object at ‘boot_params’ is out of the bounds of referenced subobject ‘kbd_status’ with type ‘unsigned char’ at offset 491 [-Warray-bounds]
   43 |   memset(&boot_params->kbd_status, 0,
      |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   44 |          (char *)&boot_params->hdr -
      |          ~~~~~~~~~~~~~~~~~~~~~~~~~~~
   45 |          (char *)&boot_params->kbd_status);
      |          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

I placed the patch in /etc/portage/patches/sys-kernel/gentoo-sources/
Tried that as well, patches fine, no problems, I'll leave it there from now on and if there's some error about it in the future I'll deal with it.
Removed the experimental USE flag, that's it.

IMHO that USE flag should be removed and the additional CPU optimizations patch should be included with the rest of the patches.
Or.. leave it as it is but don't play with it for no apparent reason, this isn't the first time it's happening.