On Thu, Sep 15, 2016 at 08:51:34AM +0000, Joakim Tjernlund wrote:
> On Thu, 2016-09-15 at 07:59 +0200, g...@kroah.com wrote:
> > On Wed, Sep 14, 2016 at 07:07:36PM +0000, Joakim Tjernlund wrote:
> > > 
> > > On Wed, 2016-09-14 at 20:10 +0200, g...@kroah.com wrote:
> > > > 
> > > > On Wed, Sep 14, 2016 at 05:57:30PM +0000, Joakim Tjernlund wrote:
> > > > > 
> > > > > 
> > > > > On Wed, 2016-09-14 at 19:08 +0200, g...@kroah.com wrote:
> > > > > > 
> > > > > > 
> > > > > > On Wed, Sep 14, 2016 at 04:13:52PM +0000, Joakim Tjernlund wrote:
> > > > > > > 
> > > > > > > 
> > > > > > > 
> > > > > > > On Wed, 2016-09-14 at 16:35 +0200, g...@kroah.com wrote:
> > > > > > > > 
> > > > > > > > 
> > > > > > > > 
> > > > > > > > On Wed, Sep 14, 2016 at 12:00:04PM +0000, Joakim Tjernlund 
> > > > > > > > wrote:
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > On Wed, 2016-09-14 at 13:54 +0200, Greg KH wrote:
> > > > > > > > > > 
> > > > > > > > > > 
> > > > > > > > > > 
> > > > > > > > > > 
> > > > > > > > > > On Wed, Sep 14, 2016 at 06:44:50AM +0000, Joakim Tjernlund 
> > > > > > > > > > wrote:
> > > > > > > > > > > 
> > > > > > > > > > > 
> > > > > > > > > > > 
> > > > > > > > > > > 
> > > > > > > > > > > 
> > > > > > > > > > > These patches are picked from Linux upstream repo and 
> > > > > > > > > > > applies directly on 4.4.x
> > > > > > > > > > > Please excuse the attachments but git send email does not 
> > > > > > > > > > > work for us ATM.
> > > > > > > > > > 
> > > > > > > > > > What is the git commit ids of these patches in Linus's tree?
> > > > > > > > > 0002-thinkpad_acpi-Add-support-for-keyboard-backlight.patch:
> > > > > > > > > 6cf4a2c575b4797b967a575b6274e0e40c5f88f1
> > > > > > > > 
> > > > > > > > I do not see that git commit id in Linus's tree, do you?
> > > > > > > 
> > > > > > > ehh, no. I used my 4.4 branch instead of linux master, sorry
> > > > > > > 
> > > > > > > > 
> > > > > > > > 
> > > > > > > > 
> > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 0003-thinkpad_acpi-Remove-ambiguous-logging-for-Unsupport.patch:
> > > > > > > > > 85342b97c2ee2a94a31ea5d59d58cdfd1f045e2a
> > > > > > > > 
> > > > > > > > Nor this one.  I stopped here, assuming that your other ids 
> > > > > > > > aren't
> > > > > > > > correct either :(
> > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > > 0004-thinkpad_acpi-Silence-an-uninitialized-variable-warn.patch:
> > > > > > > > > dcf95a1961f9614eee09066f8148813772ad4268
> > > > > > > > > 0005-thinkpad_acpi-save-kbdlight-state-on-suspend-and-
> > > > > > > > > res.patch:1c6c99c6767035761d17aeecaf0224934ccd00fb
> > > > > > > > > 0006-thinkpad_acpi-Add-support-for-HKEY-version-0x200.patch:
> > > > > > > > > 1bc0df534d11c8c9c97e1e8315b7d22bf034c8d2
> > > > > > > > > > 
> > > > > > > > > > 
> > > > > > > > > > 
> > > > > > > > > > 
> > > > > > > > > > 
> > > > > > > > > > And what happened to the other patches in this series?
> > > > > > > > > >  
> > > > > > > > > The first one was inappropriate, that is why 0001 is missing.
> > > > > > > > 
> > > > > > > > That's pretty strange of you to force us to guess what is going 
> > > > > > > > on here,
> > > > > > > > would you like to get a series like this and wonder why 
> > > > > > > > something
> > > > > > > > obvious like this is gone?
> > > > > > > 
> > > > > > > Yes, it was inconsiderate of me, sorry.
> > > > > > > 
> > > > > > > Trying to do better I figured I should generate a new series from 
> > > > > > > Linus tree but
> > > > > > > I cannot figure out how generate a series from commits that are 
> > > > > > > non sequential in Linux tree.
> > > > > > > git format path does not have a syntax for that from what I can 
> > > > > > > tell, any ideas?
> > > > > > 
> > > > > > Why not just send us the git commit ids that you want applied?  
> > > > > > That's
> > > > > > all we need.
> > > > > >  
> > > > > Oh, didn't occur to me. Here we go(ltes se fi I got it right this 
> > > > > time :)
> > > > > 
> > > > > bb28f3d51ff5e1be541d057708011cc1efe6fae9 - thinkpad_acpi: Add support 
> > > > > for keyboard backlight
> > > > 
> > > > How does this commit meet the Documentation/stable_kernel_rules.txt
> > > > requirements?  It's a new feature, and too big, why would we add it to
> > > > an old kernel like this one?  What's wrong with just using a newer
> > > > kernel for a feature like this?
> > > > 
> > > > > 
> > > > > 
> > > > > 15c75626f0999cce8357c8bf8578247134032acb - thinkpad_acpi: Remove 
> > > > > ambiguous logging for "Unsupported
> > > > > brightness interface"
> > > > > a7718360d91eedbedd58978ec8fff4a67a866f86 - thinkpad_acpi: Silence an 
> > > > > uninitialized variable warning
> > > > > afcedebc6a094224973534f43b396bbbf33fe44e - thinkpad_acpi: save 
> > > > > kbdlight state on suspend and restore
> > > > > it on
> > > > > resume
> > > > 
> > > > Are these 3 fixes for the first one?
> > > > 
> > > > > 
> > > > > 
> > > > > 0118c2d3eac0545d4095877e5a015b5dc763b3c2 - thinkpad_acpi: Add support 
> > > > > for HKEY version 0x200
> > > > 
> > > > Again, another feature :(
> > > 
> > > hmm, this was the one I really wanted as then wifi/video switch
> > > buttons work on our Lenovo laptops we recently bought to replace old
> > > HPs. I did not think this was to regard as a new feature, just a fix
> > > to make the same buttons work as on other models, sorry.
> > 
> > It's a new feature for that model of laptop, otherwise we would be
> > backporting everything all the time :(
> 
> Yes, but allowing new HW to work properly in older kernels was an exception 
> or at least so I thought.
> If here is a new minor version of the e1000 ethernet device which needs a few 
> tweaks to
> function on older kernels, do you not accept patches for that?

As the stable_kernel_rules.txt file says, yes, for simple quirk table
and device id additions, yes we do.  But this involves new code and
functions and really, a new feature implemented for that hardware.

> > > The other patches were just there to make sure I didn't run inte
> > > conflicts I didn't want to resolve.  They all made sense except for
> > > maybe the one adding keyboard backlight. 
> > > 
> > > > 
> > > > And why aren't we cc:ing the maintainers of this subsystem?
> > > 
> > > Right, they are now. If you and/or maintainers do feel this is
> > > inappropriate for stable I will just carry them my self until the next
> > > major stable kernel is released.
> > 
> > What is wrong with 4.7-stable kernels?  Why are you "stuck" at 4.4?
> > 
> 
> Non LTS kernels are fairly short lived. First you have wait for a few
> patch levels(say >=5) before one can consider the kernel fairly
> stable.

Why do you say that?  How do you judge stability?  You are getting
bugfixes that are in Linus's tree, not that were specifically for that
stable kernel release alone.

> After that there isn't that many release before it retired and you
> have to look for the next stable kernel. 

Yes, it's a never ending treadmill, welcome to the real world :)

> Each time you migrate your kernel config between major releases there
> is a risk something is mis migrated so one has to spend some time to
> review/test the kernel.

How much time does this take you?  How do you do this review?  What does
it entail?

> So I stay on LTS were I can and only systems really needing something
> newer gets a non LTS kernel.
> 
> Note that this our company kernel, used on > 100 boxes and it is not
> uncommon that something breaks somewhere after a major kernel upgrade,
> usually because I missed some kernel config.

Are you testing the -rc kernels all along?  That's the best way to do
this incrementally such that you don't miss something, and to let the
developers know that there is something wrong.

I've been running every -rc on my personal machines for years without
problems, people shouldn't be afraid of this.

By waiting a full year, your "jump" is much larger and harder,
especially for any new hardware, as you are finding out.  We can't go
back in time and add support for hardware that wasn't around when that
kernel release was being developed, so I recommend using the latest
releases, or use a distro that does this work for you (Fedora,
Tumbleweed, Arch, Debian unstable, etc.)

good luck!

greg k-h

------------------------------------------------------------------------------
_______________________________________________
ibm-acpi-devel mailing list
ibm-acpi-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ibm-acpi-devel

Reply via email to