Kernel 5.9

[xeno74]

Hi All,

The Linux 5.9 merge window is open now so I released the first alpha today.

New:

• Linux Git log
• Phoronix articles, reviews and news stories covering Linux 5.9

Download: linux-image-5.9-alpha1-X1000_X5000.tar.gz



Please test the kernels.

Thanks,
Christian

[xeno74]

Hi All,

I compiled the alpha2 (latest Git kernel) today.

New:

• dma-mapping updates
• Linux Git log
• Phoronix articles, reviews and news stories covering Linux 5.9

Unfortunately, Xorg doesn't start on some Linux distributions anymore. For example on Fienix (Debian Sid PowerPC 32-bit) and on ubuntu MATE 16.04.6 (PowerPC 32-bit). I tested these distributions on the X1000, X5000, and in a virtual e5500 QEMU machine with a virtio_gpu.

Error messages:

Code: Select all

systemd-journald[2238]: Failed to send WATCHDOG-1 notification message: Connection refused
systemd-journald[2238]: Failed to send WATCHDOG-1 notification message: Transport endpoint is not connected
systemd-journald[2238]: Failed to send WATCHDOG-1 notification message: Transport endpoint is not connected
systemd-journald[2238]: Failed to send WATCHDOG-1 notification message: Transport endpoint is not connected
systemd-journald[2238]: Failed to send WATCHDOG-1 notification message: Transport endpoint is not connected
systemd-journald[2238]: Failed to send WATCHDOG-1 notification message: Transport endpoint is not connected

But Xorg works on Ubuntu 10.04.4 (PowerPC 32-bit), openSUSE Tumbleweed 20190722 PPC64 and on Fedora 27 PPC64 with the alpha2 (latest Git kernel).

I was able to revert the dma-mapping updates with the following command:

Code: Select all

git revert 2ed90dbbf7be3b7cd2790fc6fa946c478ab496b8 -m 1

Code: Select all

[master 1196b8b16801] Revert "Merge tag 'dma-mapping-5.9' of git://git.infradead.org/users/hch/dma-mapping"
 26 files changed, 416 insertions(+), 415 deletions(-)

But the reverting of the dma-mapping updates doesn't solve the issue.

Cheers,
Christian

[xeno74]

Hi All,

I compiled the latest Git kernel today. Unfortunately the issue with Xorg still exists.

I bisected today.

1. Code: Select all

   git bisect start

2. Code: Select all

   git bisect good bcf876870b95592b52519ed4aafcf9d95999bc9c

   (Good: Linux 5.8)

3. Code: Select all

   git bisect bad 7b4ea9456dd3f73238408126ab00f1d906963d81

   (Bad: Revert "x86/mm/64: Do not sync vmalloc/ioremap mappings" -- 2020-08-06 12:02:58 -0700)
4. git bisect bad
5. git bisect good
6. git bisect bad
7. git bisect bad
8. git bisect good
9. git bisect good
10. git bisect bad
11. git bisect bad
12. git bisect good
13. git bisect good
14. git bisect bad
15. git bisect bad
16. git bisect good

Result: net/scm: Regularize compat handling of scm_detach_fds() (c0029de50982c1fb215330a5f9d433cec0cfd8cc) is the first bad commit.

Code: Select all

c0029de50982c1fb215330a5f9d433cec0cfd8cc is the first bad commit
commit c0029de50982c1fb215330a5f9d433cec0cfd8cc
Author: Kees Cook <keescook@chromium.org>
Date:   Tue Jun 9 16:11:29 2020 -0700

    net/scm: Regularize compat handling of scm_detach_fds()
    
    Duplicate the cleanups from commit 2618d530dd8b ("net/scm: cleanup
    scm_detach_fds") into the compat code.
    
    Replace open-coded __receive_sock() with a call to the helper.
    
    Move the check added in commit 1f466e1f15cf ("net: cleanly handle kernel
    vs user buffers for ->msg_control") to before the compat call, even
    though it should be impossible for an in-kernel call to also be compat.
    
    Correct the int "flags" argument to unsigned int to match fd_install()
    and similar APIs.
    
    Regularize any remaining differences, including a whitespace issue,
    a checkpatch warning, and add the check from commit 6900317f5eff ("net,
    scm: fix PaX detected msg_controllen overflow in scm_detach_fds") which
    fixed an overflow unique to 64-bit. To avoid confusion when comparing
    the compat handler to the native handler, just include the same check
    in the compat handler.
    
    Cc: Christoph Hellwig <hch@lst.de>
    Cc: Sargun Dhillon <sargun@sargun.me>
    Cc: Jakub Kicinski <kuba@kernel.org>
    Cc: netdev@vger.kernel.org
    Cc: linux-kernel@vger.kernel.org
    Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
    Signed-off-by: Kees Cook <keescook@chromium.org>

:040000 040000 b309c4be7d13b31a56f770d2eb6c76fe66b7eb19 70f6594345d1cf1aae088f9b5da99fab56d63f52 M	include
:040000 040000 8abcdca26e3d17c6dccc3fcfd445adc692e58965 eb8157a8cff232956d71ca37a115aa6ee73e51f0 M	net

This commit has been merged with the seccomp updates v5.9-rc1 on 2020-08-04 14:11:08 -0700. Since these updates, Xorg doesn't start anymore on some Linux distributions.

Unfortunately I wasn't able to revert the first bad commit.

The first bad commit depends on many other commits, which unfortunately I don't know. I tried to remove the modifications of the files from the first bad commit but without any success. There are just too many dependencies.

Christian

[xeno74]

Hi All,

I compiled a linux-next kernel because of the issue with the latest Git kernel.

Sources:

Code: Select all

git clone git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next a

This kernel includes all available updates. Unfortunately this kernel doesn't boot. It can't initialize the graphics card.

I think we have to wait for some updates.

Cheers,
Christian

[xeno74]

I reported the issues to the kernel developers today.

Link: [Latest Git kernel/Linux-next kernel] Xorg doesn't start after the seccomp updates v5.9-rc1

[xeno74]

Hi All,

Kees has released a patch because of the Xorg issue. I think his patch works because I can patch the Git source code but the kernel doesn’t boot. In my point of view his modifications aren’t responsible for this second issue. The kernel can’t initialize the graphics card anymore. I think the latest DRM updates are responsible for the second issue. Because of this second issue I can’t test his patch.

Cheers,
Christian

[xeno74]

Kees’ patch is in the Git kernel now.

seccomp-v5.9-rc1-fix1

I tried to revert the DRM updates today but without any success because of dependencies so I can’t test Kees’ patch.

[xeno74]

FYI:

PowerPC updates 5.9-1

selftests/powerpc: Fix pkey syscall redefinitions

DRM updates

[xeno74]

Hi All,

I compiled the latest Git kernel today. The issue with Xorg is solved. The X5000 boots and works with the latest Git kernel. Unfortunately the X1000 doesn't boot anymore. It can't initialize the graphics card anymore.

I bisected today.

1. Code: Select all

   git bisect start

2. Code: Select all

   git bisect good bcf876870b95592b52519ed4aafcf9d95999bc9c

   (Good: Linux 5.8)

3. Code: Select all

   git bisect bad 7b9de97711225559af213dc52b6ea883ef1ea7a8

   (Bad: powerpc/ptrace: Fix build error in pkey_get() -- 2020-08-07 18:27:26 -0700)
4. git bisect good
5. git bisect good
6. git bisect good
7. git bisect bad
8. git bisect bad
9. git bisect bad
10. git bisect good
11. git bisect bad
12. git bisect good
13. git bisect bad
14. git bisect good
15. git bisect good
16. git bisect good

Result: powerpc/book3s64/pkeys: Simplify pkey disable branch (a4678d4b477c3d2901f101986ca01406f3b7eaea) is the first bad commit.

Code: Select all

a4678d4b477c3d2901f101986ca01406f3b7eaea is the first bad commit
commit a4678d4b477c3d2901f101986ca01406f3b7eaea
Author: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Date:   Thu Jul 9 08:59:32 2020 +0530

    powerpc/book3s64/pkeys: Simplify pkey disable branch
    
    Make the default value FALSE (pkey enabled) and set to TRUE when we
    find the total number of keys supported to be zero.
    
    Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
    Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
    Link: https://lore.kernel.org/r/20200709032946.881753-10-aneesh.kumar@linux.ibm.com

:040000 040000 ba5ec72c45ced2ec3059d47db0ab493e63c3d07c 7d11c97b991b194aa834ef0ea0967c07710dbc06 M	arch

Unfortunately I wasn't able to revert the first bad commit.

The first bad commit depends on many other commits, which unfortunately I don't know. I tried to remove the modifications of the files from the first bad commit but without any success. There are just too many dependencies.

I reverted the commit selftests/powerpc: Fix pkey syscall redefinitions and compiled a new kernel but without any success.

Christian

[xeno74]

I reported the X1000 boot issue to the kernel developers.

Link: [PASEMI] Nemo board doesn't boot anymore after the commit "powerpc/book3s64/pkeys: Simplify pkey disable branch"

I will create a test kernel for the X5000 tomorrow because the kernel boots on the X5000.