Mirror of the lttng-dev mailing list
 help / color / mirror / Atom feed
From: Michael Jeanson via lttng-dev <lttng-dev@lists.lttng.org>
To: Roxana Nicolescu <roxana.nicolescu@canonical.com>,
	Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
	lttng-dev@lists.lttng.org
Cc: "Stéphane Graber" <stgraber@stgraber.org>
Subject: Re: [lttng-dev] [PATCH lttng-modules 0/1] Introduce configure script to describe changes in linux kernel interface
Date: Fri, 7 Jul 2023 14:37:42 -0400	[thread overview]
Message-ID: <9b8f3bca-61f9-998d-3215-48b974a04bd7@efficios.com> (raw)
In-Reply-To: <ee4ab50c-9847-7bd6-b0a9-479f6d06fe32@canonical.com>

On 2023-07-06 17:10, Roxana Nicolescu wrote:
> Hi,
> 
> I gave 2 derivatives as an example, but in fact there are more than 10. We 
> will end up with a bigger if condition that the actual function 
> implementation. And this is also not a sustainable solution in case there are 
> new derivatives coming after the fix. It will be hard to maintain. Plus, to 
> achieve this we have to release a new kernel version to expose the suffixes in 
> utsrelease.h and getting the fix on the user side would require a reboot to 
> upgrade their kernel version.
> 
> While the current solution based on versioning may be easier to understand 
> from the developer perspective, it adds technical depth to the implementation 
> and in fact it requires more maintenance than expected, also on the 
> distribution side. When there is a change on the kernel side, you need to 
> check if it landed in older kernel versions and based on that you need to add 
> a new interval in the if condition.
> Now, looking at the configure script approach, while maybe not as clear as an 
> ifdef, only one change would be enough to solve this problem. The LTTng module 
> does not have to care about each distribution's way of versioning and the 
> distributions don't have to deal with issues on their side if they use a 
> version that the LTTng module did not take into account because you were not 
> aware of it.
> I think maintaining it won't be that hard if it's readable and documented 
> properly. I am more than happy to improve it and help with maintaining it if 
> you agree on this.
> 
> All the best,
> Roxana

Hi Roxana,

We are still not convinced of the benefits of a configure script over the 
added complexity since the issue is specific to Ubuntu where you choose to 
maintain 10s of different kernel branches with diverging code and overlapping 
versions.

I made a patch [1] with version ranges that adds support for all kernels from 
the kinetic "linux" repo [2] including 'generic' and other flavours and 
'lowlatency'. It however doesn't cover 'linux-aws' or other repos which share 
ABI numbers with 'lowlatency'.

You can of course use the configure script approach in the Ubuntu package 
'lttng-modules-dkms' by adding your patch to the packaging code.

Regards,

Michael

[1] https://review.lttng.org/c/lttng-modules/+/10485
[2] https://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/kinetic

_______________________________________________
lttng-dev mailing list
lttng-dev@lists.lttng.org
https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

      reply	other threads:[~2023-07-07 18:37 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-03 18:28 Roxana Nicolescu via lttng-dev
2023-07-03 18:28 ` [lttng-dev] [PATCH lttng-modules 1/1] " Roxana Nicolescu via lttng-dev
2023-07-04 15:35 ` [lttng-dev] [PATCH lttng-modules 0/1] " Michael Jeanson via lttng-dev
2023-07-04 18:07   ` Mathieu Desnoyers via lttng-dev
2023-07-04 18:39     ` Roxana Nicolescu via lttng-dev
2023-07-04 19:16       ` Mathieu Desnoyers via lttng-dev
2023-07-06 21:10         ` Roxana Nicolescu via lttng-dev
2023-07-07 18:37           ` Michael Jeanson via lttng-dev [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=9b8f3bca-61f9-998d-3215-48b974a04bd7@efficios.com \
    --to=lttng-dev@lists.lttng.org \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=mjeanson@efficios.com \
    --cc=roxana.nicolescu@canonical.com \
    --cc=stgraber@stgraber.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox