-stable review patch. If anyone has any objections, please let us know.
---------------------
From: David S. Miller <[email protected]>
Upstream commit: a442585952f137bd4cdb1f2f3166e4157d383b82
Signed-off-by: David S. Miller <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Signed-off-by: Chris Wright <[email protected]>
---
drivers/net/niu.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--- a/drivers/net/niu.c
+++ b/drivers/net/niu.c
@@ -33,8 +33,8 @@
#define DRV_MODULE_NAME "niu"
#define PFX DRV_MODULE_NAME ": "
-#define DRV_MODULE_VERSION "0.6"
-#define DRV_MODULE_RELDATE "January 5, 2008"
+#define DRV_MODULE_VERSION "0.7"
+#define DRV_MODULE_RELDATE "February 18, 2008"
static char version[] __devinitdata =
DRV_MODULE_NAME ".c:v" DRV_MODULE_VERSION " (" DRV_MODULE_RELDATE ")\n";
--
On 21/03/2008, Chris Wright <[email protected]> wrote:
> -stable review patch. If anyone has any objections, please let us know.
> ---------------------
>
I don't see how this one fits this criteria from
Documentation/stable_kernel_rules.txt :
...
- It must fix a real bug that bothers people (not a, "This could be a
problem..." type thing).
...
As I see it, this patch is quite trivial and does not fix any "real problem".
How come this is suitable for -stable?
I have no objections to the patch "as such", I just don't think it
makes sense for -stable.
> From: David S. Miller <[email protected]>
>
> Upstream commit: a442585952f137bd4cdb1f2f3166e4157d383b82
>
> Signed-off-by: David S. Miller <[email protected]>
> Signed-off-by: Greg Kroah-Hartman <[email protected]>
> Signed-off-by: Chris Wright <[email protected]>
>
> ---
> drivers/net/niu.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> --- a/drivers/net/niu.c
> +++ b/drivers/net/niu.c
> @@ -33,8 +33,8 @@
>
> #define DRV_MODULE_NAME "niu"
> #define PFX DRV_MODULE_NAME ": "
> -#define DRV_MODULE_VERSION "0.6"
> -#define DRV_MODULE_RELDATE "January 5, 2008"
> +#define DRV_MODULE_VERSION "0.7"
> +#define DRV_MODULE_RELDATE "February 18, 2008"
>
> static char version[] __devinitdata =
> DRV_MODULE_NAME ".c:v" DRV_MODULE_VERSION " (" DRV_MODULE_RELDATE ")\n";
>
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
From: "Jesper Juhl" <[email protected]>
Date: Sat, 22 Mar 2008 01:10:40 +0100
> As I see it, this patch is quite trivial and does not fix any "real problem".
> How come this is suitable for -stable?
>
> I have no objections to the patch "as such", I just don't think it
> makes sense for -stable.
It makes a difference for driver maintainers when users
report bugs and we ask them for the version printed
by the driver so that we know which fixes have been
applied.
You know, if people are going to be jerky about this,
I'll just include the version bump in the actual bug
fixes which I sometimes do anyways.
Thanks for bringing this up, it's a useful use of
everyone's time :-/
On 22/03/2008, David Miller <[email protected]> wrote:
> From: "Jesper Juhl" <[email protected]>
> Date: Sat, 22 Mar 2008 01:10:40 +0100
>
>
> > As I see it, this patch is quite trivial and does not fix any "real problem".
> > How come this is suitable for -stable?
> >
> > I have no objections to the patch "as such", I just don't think it
> > makes sense for -stable.
>
>
> It makes a difference for driver maintainers when users
> report bugs and we ask them for the version printed
> by the driver so that we know which fixes have been
> applied.
>
Ok, fair enough.
> You know, if people are going to be jerky about this,
> I'll just include the version bump in the actual bug
> fixes which I sometimes do anyways.
>
> Thanks for bringing this up, it's a useful use of
> everyone's time :-/
>
Excuse me for reading patches, trying to spot problems and thinking
about where they are applied and what rules apply etc.
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
> It makes a difference for driver maintainers when users
> report bugs and we ask them for the version printed
> by the driver so that we know which fixes have been
> applied.
I'm not sure if this series of patches really accomplishes that...
you're leaving out some patches such as "[NIU]: Fix 1G PHY link state
handling." that are in the real 0.7. So the 2.6.24.stable version 0.7
has a different set of fixes than the 2.6.25 version 0.7.
- R.
From: Roland Dreier <[email protected]>
Date: Fri, 21 Mar 2008 19:35:40 -0700
> > It makes a difference for driver maintainers when users
> > report bugs and we ask them for the version printed
> > by the driver so that we know which fixes have been
> > applied.
>
> I'm not sure if this series of patches really accomplishes that...
> you're leaving out some patches such as "[NIU]: Fix 1G PHY link state
> handling." that are in the real 0.7. So the 2.6.24.stable version 0.7
> has a different set of fixes than the 2.6.25 version 0.7.
That was an oversight and not intentional, thanks for
for catching that.
From: David Miller <[email protected]>
Date: Fri, 21 Mar 2008 22:21:25 -0700 (PDT)
> From: Roland Dreier <[email protected]>
> Date: Fri, 21 Mar 2008 19:35:40 -0700
>
> > > It makes a difference for driver maintainers when users
> > > report bugs and we ask them for the version printed
> > > by the driver so that we know which fixes have been
> > > applied.
> >
> > I'm not sure if this series of patches really accomplishes that...
> > you're leaving out some patches such as "[NIU]: Fix 1G PHY link state
> > handling." that are in the real 0.7. So the 2.6.24.stable version 0.7
> > has a different set of fixes than the 2.6.25 version 0.7.
>
> That was an oversight and not intentional, thanks for
> for catching that.
Actually Roland, that 1G fix you mention was put in during the
2.6.23 development series.
If you ask git for all the NIU driver changes since 2.6.24
you get:
davem@sunset:~/src/GIT/linux-2.6$ git log v2.6.24.. drivers/net/niu.[ch]
commit fa907895b7b776208a1406efe5ba7ffe0f49f507
Author: Matheos Worku <[email protected]>
Date: Wed Feb 20 00:18:09 2008 -0800
[NIU]: More BMAC alt MAC address fixes.
From: Matheos Worku <[email protected]>
1) niu_enable_alt_mac() needs to be adjusted so that the mask
is computed properly for the BMAC case.
2) BMAC has 6 alt MAC addresses available, not 7.
Signed-off-by: David S. Miller <[email protected]>
commit a442585952f137bd4cdb1f2f3166e4157d383b82
Author: David S. Miller <[email protected]>
Date: Mon Feb 18 21:30:48 2008 -0800
[NIU]: Bump driver version and release date.
Signed-off-by: David S. Miller <[email protected]>
commit 3b5bcedeeb755b6e813537fcf4c32f010b490aef
Author: Matheos Worku <[email protected]>
Date: Mon Feb 18 21:30:03 2008 -0800
[NIU]: Fix BMAC alternate MAC address indexing.
BMAC port alternate MAC address index needs to start at 1. Index 0 is
used for the main MAC address.
Signed-off-by: Matheos Worku <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
commit 2caf62f6cae46e36b1c4a1b0f2d9ef82af89cad2
Author: Joe Perches <[email protected]>
Date: Thu Dec 20 04:07:35 2007 -0800
[NIU]: Use print_mac
Signed-off-by: Joe Perches <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
davem@sunset:~/src/GIT/linux-2.6$
Since using print_mac is a one-liner cleanup rather than a bug fix,
it's reasonable to leave it out for 2.6.24-stable
What's left is exactly what I submitted for -stable.
Hi David,
On Sat, Mar 22, 2008 at 2:14 AM, David Miller <[email protected]> wrote:
> It makes a difference for driver maintainers when users
> report bugs and we ask them for the version printed
> by the driver so that we know which fixes have been
> applied.
>
> You know, if people are going to be jerky about this,
> I'll just include the version bump in the actual bug
> fixes which I sometimes do anyways.
>
> Thanks for bringing this up, it's a useful use of
> everyone's time :-/
Uhm, Jesper was asking a relevant question and definitely not being
"jerky". As you probably know, many of us are dropping explicit driver
versions from the code for the exact reason why you're trying to patch
the stable tree now. They get out of sync real easy and the kernel
version already provides the same information so why bother.
Furthermore, I honestly don't see how your patch meets the (rather
strict) requirements of stable series either. So it would be nice if
one of the stable maintainers stepped up to say that these kind of
patches are okay to set a precedent for others as well.
Pekka
From: "Pekka Enberg" <[email protected]>
Date: Sun, 23 Mar 2008 00:57:29 +0200
> Furthermore, I honestly don't see how your patch meets the (rather
> strict) requirements of stable series either.
Show how changing a version string can cause a regression for
a user.
Then I'll stop being angry about all the time that is being wasted on
this topic.
On Sun, Mar 23, 2008 at 12:57:29AM +0200, Pekka Enberg wrote:
> Furthermore, I honestly don't see how your patch meets the (rather
> strict) requirements of stable series either. So it would be nice if
> one of the stable maintainers stepped up to say that these kind of
> patches are okay to set a precedent for others as well.
These kind of patches are perfectly fine. We trust the maintainers of
drivers and subsystems to tell the -stable maintainers what they want
added to the tree, and if that includes a version number update to make
their lives easier, who are we to say no to that.
Becides, it's not the first time patches like this have been accepted.
thanks,
greg k-h
> Actually Roland, that 1G fix you mention was put in during the
> 2.6.23 development series.
>
> If you ask git for all the NIU driver changes since 2.6.24
> you get:
Sorry for the noise... I did
git log cb77df3e..a4425859 drivers/net/niu.c
to get all the niu changes between when you bumped the version to 0.6
and when you bumped it to 0.7. I didn't think to check which post-0.6
changes were already in 2.6.24.
As long as the version helps you out, I guess it's all good.
- R.
* Greg KH ([email protected]) wrote:
> On Sun, Mar 23, 2008 at 12:57:29AM +0200, Pekka Enberg wrote:
> > Furthermore, I honestly don't see how your patch meets the (rather
> > strict) requirements of stable series either. So it would be nice if
> > one of the stable maintainers stepped up to say that these kind of
> > patches are okay to set a precedent for others as well.
>
> These kind of patches are perfectly fine. We trust the maintainers of
> drivers and subsystems to tell the -stable maintainers what they want
> added to the tree, and if that includes a version number update to make
> their lives easier, who are we to say no to that.
I completely agree. The -stable rules are guidelines, primarily to help
us say no. When it's submitted by a maintainer as something that helps
them support their code (esp a no risk change like version bump to keep
things in line), there's little reason to reject. Consider the version
bump part of the bug fixes.
thanks,
-chris
Hi David,
On Sun, Mar 23, 2008 at 1:26 AM, David Miller <[email protected]> wrote:
> Show how changing a version string can cause a regression for
> a user.
It obviously can't but then again, we don't put comment or
documentation fixes in the stable tree either just because. But agreed
this is pointless now as Chris and Greg have explicitly said they're
ok.
Pekka
From: "Pekka Enberg" <[email protected]>
Date: Sun, 23 Mar 2008 10:42:46 +0200
> It obviously can't but then again, we don't put comment or
> documentation fixes in the stable tree either just because. But agreed
> this is pointless now as Chris and Greg have explicitly said they're
> ok.
Thus, thanks for continuing to waste everyone's time.
On Sun, Mar 23, 2008 at 12:57:29AM +0200, Pekka Enberg wrote:
> Hi David,
>
> On Sat, Mar 22, 2008 at 2:14 AM, David Miller <[email protected]> wrote:
> > It makes a difference for driver maintainers when users
> > report bugs and we ask them for the version printed
> > by the driver so that we know which fixes have been
> > applied.
> >
> > You know, if people are going to be jerky about this,
> > I'll just include the version bump in the actual bug
> > fixes which I sometimes do anyways.
> >
> > Thanks for bringing this up, it's a useful use of
> > everyone's time :-/
>
> Uhm, Jesper was asking a relevant question and definitely not being
> "jerky". As you probably know, many of us are dropping explicit driver
> versions from the code for the exact reason why you're trying to patch
> the stable tree now. They get out of sync real easy and the kernel
> version already provides the same information so why bother.
>...
There are two completely different cases one mustn't confuse:
> 90% of all version numbers, cvs tags,... in the kernel haven't been
updated for ages and should be removed.
But there are some drivers and ALSA that have an active maintainer, an
actively maintained version number, and tend to often be backported into
distribution kernels.
> Pekka
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
David,
When I made the reply below I was rather tired and edgy. I would like
to explain why I commented as I did and offer an appology for wasting
your time and/or annoying you - that was never my intention.
I try to participate constructively where and when I can and one of
the ways I try to do so is by reading patches that people submit. Most
of the time I'm not qualified to comment on them and I try not to.
In this specific case I felt I knew the -stable rules well enough and
could understand the patch well enough to offer up a constructive
comment, so I did.
My Intention was simply to point out that the patch *seemed* to fall
outside what the -stable rules allowed, and my understanding was that
the rules were there for a reason - to keep -stable patches to the
very minimum required to fix bugs and only that.
When you replied the way you did I got rather irritated by the fact
that you chose to see it as an attempt to waste peoples time rather
than see it how I intended it; as an attempt to help out by reviewing
patches and point out any oddities/guideline-deviation I observed.
I should have waited to respond and certainly not written my response
while still being annoyed.
I realize now that patches like this are acceptable for -stable, so
I'll refrain from commenting on them in the future.
I hope you can accept an appology and see that I did not intend to
waste anyones time.
Kind regards,
Jesper Juhl
On 22/03/2008, Jesper Juhl <[email protected]> wrote:
> On 22/03/2008, David Miller <[email protected]> wrote:
> > From: "Jesper Juhl" <[email protected]>
> > Date: Sat, 22 Mar 2008 01:10:40 +0100
> >
> >
> > > As I see it, this patch is quite trivial and does not fix any "real problem".
> > > How come this is suitable for -stable?
> > >
> > > I have no objections to the patch "as such", I just don't think it
> > > makes sense for -stable.
> >
> >
> > It makes a difference for driver maintainers when users
> > report bugs and we ask them for the version printed
> > by the driver so that we know which fixes have been
> > applied.
> >
>
> Ok, fair enough.
>
>
> > You know, if people are going to be jerky about this,
> > I'll just include the version bump in the actual bug
> > fixes which I sometimes do anyways.
> >
> > Thanks for bringing this up, it's a useful use of
> > everyone's time :-/
> >
>
> Excuse me for reading patches, trying to spot problems and thinking
> about where they are applied and what rules apply etc.
>
>
> --
>
> Jesper Juhl <[email protected]>
>
> Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
> Plain text mails only, please http://www.expita.com/nomime.html
>
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html