Subject: Reiser4 SFRN 4.0.2 issues with kernel 5.15

Niltze Ed-


On Thu, Dec 23, 2021 at 10:47 AM Metztli Information Technology <[email protected]> wrote:
>
> Niltze [Hello]-
>
> Attached is a Reiser4 -enabled Debian packaging hack for Linux 5.14.21-
> 1 EOL: metztli-reiser4-gcc10-debian-packaging-for-5.14.21-1EOL.patch
>
> For a one off, we can fetch the Debian packaging for 5.14.16-1, the
> last one that Debian released, as:
>
> git clone --depth 1 -b debian/5.14.16-1 --single-branch
> https://anonscm.debian.org/git/kernel/linux.git
>
> Customize your branch:
> git switch -c <new-branch-name>
>
> fetch the 5.14.21 EOL kernel source, etc.
> wget https://cdn.kernel.org/pub/linux/kernel/v5.x/linux-5.14.21.tar.xz
> https://cdn.kernel.org/pub/linux/kernel/v5.x/linux-5.14.21.tar.sign
>
> Verify:
> xz -dc linux-5.14.21.tar.xz | gpg --verify  linux-5.14.21.tar.sign -
>
> Make a sym link to feed it to debian:
> ln -s linux-5.14.21.tar.xz linux_5.14.21.orig.tar.xz
>
> Change directory:
> cd linux
>
> Apply metztli-reiser4-gcc10-debian-packaging-for-5.14.21-1EOL.patch
>
> Debianize your kernel source tree:
> debian/rules orig
>
> Apply either of Ed's modified patch for 5.14:
> metztli-reiser4-stable-for-5.14.patch.gz
> OR
> metztli-reiser4-unstable-for-5.14.patch.gz
>
> Of course, we assume -- a priori -- development environment.
>
> Snapshot of the hack in an AMD Ryzen bare metal server:
> https://metztli.it/bullseye-reiser5/reiser4-sfrn4-5.14.20-1-amd.png
>
> Hack is currently working in a couple of physical hardware instances
> but there are **no implicit nor explicit warranties.**
[]

I have been stuck with upgrading your reiser4 SFRN4 stable patch to make it work with kernel 5.15.xy;
although I am able to build the patched kernel successfully, once I install it under a local Debian Bullseye AMD64 testing instance,
I experience considerable latency during certain operations, most noticeably running firefox 96.0.2 and/or its offshoot suite Seamonkey
< https://www.seamonkey-project.org/start/ >, etc..

Of course, it is because kernel code changes from 5.14 to 5.15; more specifically the reiser4 -patched files below seem to be an issue:

fs/fs-writeback.c
fs/read_write.c
include/linux/fs.h
mm/filemap.c
mm/page-writeback.c
mm/vmscan.c

I have used vim -Rd to compare the relevant files in a reiser4 sfrn4 -patched kernel 5.14.21 EOL, executing where I am writing this email,
and the reiser4 sfrn4 -patched kernel 5.15.6, where considerable latency issues are being experienced.

Attached, please find gzip archives with the files of reference from 5.14.21 and 5.15.6:
reiser4-sfrn4-linux-5.14.21-vs-5.15.6.tar.gz

as well as the patch that I applied onto both kernels:
metztli-reiser4-stable-for-5.14.patch.gz

If you have some free time, I would appreciate it if you could take a look at those.


Best Professional Regards.

--
--
Jose R R
http://metztli.it
-----------------------------------------------------------------------
----------------------
Download Metztli Reiser4: Debian Bullseye w/ Linux 5.14.21 AMD64
-----------------------------------------------------------------------
----------------------
feats ZSTD compression https://sf.net/projects/metztli-reiser4/
-----------------------------------------------------------------------
----------------------
or SFRN 5.1.3, Metztli Reiser5 https://sf.net/projects/debian-reiser4/
-----------------------------------------------------------------------
--------------------
Official current Reiser4 resources: https://reiser4.wiki.kernel.org/


Attachments:
metztli-reiser4-stable-for-5.14.patch.gz (608.90 kB)
reiser4-sfrn4-linux-5.14.21-vs-5.15.6.tar.gz (307.55 kB)
Download all attachments