question about one of your ubuntu patch

[David Lee]

Hi Kamal,

Thanks for your reply.

 but for the 4.19-stable patch sets, only those that can be reasonably
ported to 4.15 are accepted (a large portion of them aren't suitable).

How do you know if a patch is suitable for 4.15?

Besides, commits listed in
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1889474 are already
filtered by your rules? So the date in web page title--"2020-07-29" is the
date of starting filtering patches or ending filtering patches?
Usually, how many days you spend on filtering patches?

Thanks


On Tue, Sep 8, 2020 at 11:48 AM Kamal Mostafa <kamal at canonical.com> wrote:

> Hi David-
>
> On Tue, Sep 8, 2020 at 5:11 AM David Lee <saynice111 at gmail.com> wrote:
>
>> I think if commits are chosen, porting them should be easy. Can you give
>> me more details on why
>> it's time consuming?
>>
>
> Because each of the upstream stable trees gain patches at a rate > 5000
> per year.  The majority of those patches are indeed relatively easy to
> port, but a significant fraction do not apply easily, and require analysis,
> porting, and testing effort.   We do this porting process continuously for
> multiple supported Ubuntu kernel versions.
>
>
>> > I just found a patch in
>>> > https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1857158, "rtc:
>>> disable
>>> > uie before setting time and enable after"--7e7c005b4b1f1f169bcc. It's
>>> > in v4.19.89. But it's not patched in Ubuntu-4.15.0-76.86,which is
>>> issued at
>>> > Jan 17 2020.
>>> > Ubuntu-4.15.0-76.86 contains the commit in the "Fixes:"
>>> > in 7e7c005b4b1f1f169bcc, which means that ubuntu introduces the bug.
>>> It's
>>> > necessary to be patched. And the patch is also in the list of
>>> > https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1857158. I
>>> thought all
>>> > the patches
>>> > in that website are to be added in ubuntu. Why is not it added in
>>> ubuntu?
>>>
>>
> As Jeurg noted, that patch was indeed ported to Ubuntu Bionic and released
> as indicated in the bug you linked.
>
> It has a different SHA in the bionic repo (which will always be true for
> stable patches) than the upstream SHA.  In general, when you're looking for
> an upstream commit in an Ubuntu kernel repo, your best bet is to search by
> title.  Here's your example, and another one where the patch title doesn't
> exactly match upstream (because we merged it before upstream did):
>
> $ git log --oneline --grep="rtc: disable uie"
> be2063280b82 rtc: disable uie before setting time and enable after
>
> $ git log --oneline --grep="s390/mm: fix page table"
> b77f29531a23 UBUNTU: SAUCE: s390/mm: fix page table upgrade vs 2ndary
> address mode accesses
>
>
> > Is there any other rules?
>>>
>>
> For Bionic's 4.15 kernel (where we port upstream stable patch sets from
> 4.14-stable and 4.19-stable) the general rule is that almost every patch
> from 4.14-stable will get applied or ported (rarely, we determine that a
> 4.14-stable patch is just unsuitable for 4.15) -- but for the 4.19-stable
> patch sets, only those that can be reasonably ported to 4.15 are accepted
> (a large portion of them aren't suitable).
>
> As Thadeu described, the upstream patch sets I'm referring to above are
> the "stable" and "longterm" series, here: https://www.kernel.org/
>
> Hope that helps answer your questions.
>
>  -Kamal Mostafa (Canonical Kernel Team)
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20200910/98d800b0/attachment.html>