Kernel version of Arnaldos Net Git tree

Kernel version of Arnaldos Net Git tree

[@ 2007-08-20 2:44 UTC (permalink / raw)]

Hi Arnaldo,
     today I clone your git kernel and although it is shown 2.6.24,
when I run make menuconfig it is shown 2.6.23-rc2. Is there something
missing in what I did when I clone the tree?

Leandro.

-- 
Leandro Melo de Sales.
Pervasive and Embedded Computing Laboratory
BRisa and E-Phone Projects Manager & Network Administrator
http://embedded.ufcg.edu.br/
+55 083 3310-1404 (extension 208)

"There are 10 types of people in this world: those who understand
binary, those who don't"

Re: Kernel version of Arnaldos Net Git tree

[Ian McDonald]

On 8/20/07, ツ Leandro Sales <leandroal@gmail.com> wrote:
> Hi Arnaldo,
>      today I clone your git kernel and although it is shown 2.6.24,
> when I run make menuconfig it is shown 2.6.23-rc2. Is there something
> missing in what I did when I clone the tree?
>
> Leandro.
>
Leandro,

This is aimed at the 2.6.24 release, rather than being based on 2.6.24.

It is based on 2.6.23-rc2 as that was the latest at the time when
Arnaldo was working on his tree. 2.6.23 has not yet been released.

Ian
-- 
Web1: http://wand.net.nz/~iam4/
Web2: http://www.jandi.co.nz
Blog: http://iansblog.jandi.co.nz

Re: Kernel version of Arnaldos Net Git tree

[@ 2007-08-20 2:54 UTC (permalink / raw)]

In both tree, Gerrit tree and Arnaldo tree I'm in trouble. I compiled
the kernel and on the boot I'm getting the following messages and
after this the pc stop to work:

Testing NMI watchdog ... <7>APIC error on CPU0: 40(40)
CPU#0: NMI appears to be stuck (0->0)!
CPU#1: NMI appears to be stuck (0->0)!
Using IPI No-Shortcut mode
drivers/rtc/hctosys.c: unable to open rtc device (rtc0)

Any clue to solve this?

Thanks,
Leandro.

2007/8/19, Ian McDonald <ian.mcdonald@jandi.co.nz>:
> On 8/20/07, ツ Leandro Sales <leandroal@gmail.com> wrote:
> > Hi Arnaldo,
> >      today I clone your git kernel and although it is shown 2.6.24,
> > when I run make menuconfig it is shown 2.6.23-rc2. Is there something
> > missing in what I did when I clone the tree?
> >
> > Leandro.
> >
> Leandro,
>
> This is aimed at the 2.6.24 release, rather than being based on 2.6.24.
>
> It is based on 2.6.23-rc2 as that was the latest at the time when
> Arnaldo was working on his tree. 2.6.23 has not yet been released.
>
> Ian
> --
> Web1: http://wand.net.nz/~iam4/
> Web2: http://www.jandi.co.nz
> Blog: http://iansblog.jandi.co.nz
>


-- 
Leandro Melo de Sales.
Pervasive and Embedded Computing Laboratory
BRisa and E-Phone Projects Manager & Network Administrator
http://embedded.ufcg.edu.br/
+55 083 3310-1404 (extension 208)

"There are 10 types of people in this world: those who understand
binary, those who don't"

Re: Kernel version of Arnaldos Net Git tree

[Ian McDonald]

On 8/20/07, ツ Leandro Sales <leandroal@gmail.com> wrote:
> In both tree, Gerrit tree and Arnaldo tree I'm in trouble. I compiled
> the kernel and on the boot I'm getting the following messages and
> after this the pc stop to work:
>
> Testing NMI watchdog ... <7>APIC error on CPU0: 40(40)
> CPU#0: NMI appears to be stuck (0->0)!
> CPU#1: NMI appears to be stuck (0->0)!
> Using IPI No-Shortcut mode
> drivers/rtc/hctosys.c: unable to open rtc device (rtc0)
>
> Any clue to solve this?
>
> Thanks,
> Leandro.
>

Try Linus' tree and see if there is the same issue. From memory this
is a clock issue that has been solved very recently.

If it hasn't been solved then report to the lkml list. If it has been
solved you can wait until the trees are rebased or you can copy the
patches yourself over to a copy of Linus' tree.

Ian
-- 
Web1: http://wand.net.nz/~iam4/
Web2: http://www.jandi.co.nz
Blog: http://iansblog.jandi.co.nz

Re: Kernel version of Arnaldos Net Git tree

[@ 2007-08-20 3:13 UTC (permalink / raw)]

OK. I'll try it! I'm cloning linus tree... Thank you.

Leandro.

2007/8/19, Ian McDonald <ian.mcdonald@jandi.co.nz>:
> On 8/20/07, ツ Leandro Sales <leandroal@gmail.com> wrote:
> > In both tree, Gerrit tree and Arnaldo tree I'm in trouble. I compiled
> > the kernel and on the boot I'm getting the following messages and
> > after this the pc stop to work:
> >
> > Testing NMI watchdog ... <7>APIC error on CPU0: 40(40)
> > CPU#0: NMI appears to be stuck (0->0)!
> > CPU#1: NMI appears to be stuck (0->0)!
> > Using IPI No-Shortcut mode
> > drivers/rtc/hctosys.c: unable to open rtc device (rtc0)
> >
> > Any clue to solve this?
> >
> > Thanks,
> > Leandro.
> >
>
> Try Linus' tree and see if there is the same issue. From memory this
> is a clock issue that has been solved very recently.
>
> If it hasn't been solved then report to the lkml list. If it has been
> solved you can wait until the trees are rebased or you can copy the
> patches yourself over to a copy of Linus' tree.
>
> Ian
> --
> Web1: http://wand.net.nz/~iam4/
> Web2: http://www.jandi.co.nz
> Blog: http://iansblog.jandi.co.nz
>


-- 
Leandro Melo de Sales.
Pervasive and Embedded Computing Laboratory
BRisa and E-Phone Projects Manager & Network Administrator
http://embedded.ufcg.edu.br/
+55 083 3310-1404 (extension 208)

"There are 10 types of people in this world: those who understand
binary, those who don't"

Re: Kernel version of Arnaldos Net Git tree

[@ 2007-08-20 3:14 UTC (permalink / raw)]

2007/8/19, Ian McDonald <ian.mcdonald@jandi.co.nz>:
> On 8/20/07, ツ Leandro Sales <leandroal@gmail.com> wrote:
> > In both tree, Gerrit tree and Arnaldo tree I'm in trouble. I compiled
> > the kernel and on the boot I'm getting the following messages and
> > after this the pc stop to work:
> >
> > Testing NMI watchdog ... <7>APIC error on CPU0: 40(40)
> > CPU#0: NMI appears to be stuck (0->0)!
> > CPU#1: NMI appears to be stuck (0->0)!
> > Using IPI No-Shortcut mode
> > drivers/rtc/hctosys.c: unable to open rtc device (rtc0)
> >
> > Any clue to solve this?
> >
> > Thanks,
> > Leandro.
> >
>
> Try Linus' tree and see if there is the same issue. From memory this
> is a clock issue that has been solved very recently.
>
> If it hasn't been solved then report to the lkml list. If it has been
> solved you can wait until the trees are rebased or you can copy the
> patches yourself over to a copy of Linus' tree.
>
> Ian
> --
> Web1: http://wand.net.nz/~iam4/
> Web2: http://www.jandi.co.nz
> Blog: http://iansblog.jandi.co.nz
>

BTW, sorry for my top reply in the earlier posts.

Leandro.

Re: Kernel version of Arnaldos Net Git tree

[Ian McDonald]

On 8/21/07, ツ Leandro Sales <leandroal@gmail.com> wrote:
>
> Hello list,

You forgot to cc list!

>    I tested using Linus and David Miller and the problem of watchdog
> didn't happen. But using both Arnaldo or Gerrit I got the watchdog
> problem! :( How often gerrit sync his git repository into Arnaldo git
> and how often Arnaldo do this into David Miller netdev git tree?
>
> Leandro.
>
If it's any consolation then I went to test on one of my machines and
it won't boot either. I'm running an older version of Gerrit's tree
modified a lot (for faster restart testing).

So I'll try doing the same sort of thing as you too now, but first
I'll try Gerrit's more recent tree. Basically Gerrit and Arnaldo and
Dave sync their trees when they feel like it. Generally before
submitting upstream people tend to resync to check their patches still
apply. Gerrit and Arnaldo and I will often sync off Dave M's release+1
tree (currrently 2.6.24) as this is where networking changes go if the
merge window is not currently open (Linus only allows changes up to
rc1 of any release).

What you and I might have to do is take Gerrit's tree and pull his
changes out of the tree (they are the most recent patches) and apply
them to Linus' latest tree. I don't know the syntax for this yet but
will tell you if I have to do this (please tell me the syntax if you
do it first!!). We will need Linus' tree rather than Dave
M's/Arnaldo's tree (even if they rebase) as at the moment with
Arnaldo's recent patch set I'm pretty sure Gerrit's patches won't
apply cleanly any more until he reworks his patch set. The joys of
distributed development!

Hope this is of some use.

Ian
-- 
Web1: http://wand.net.nz/~iam4/
Web2: http://www.jandi.co.nz
Blog: http://iansblog.jandi.co.nz

Re: Kernel version of Arnaldos Net Git tree

[Ian McDonald]

On 8/21/07, Ian McDonald <ian.mcdonald@jandi.co.nz> wrote:
> On 8/21/07, ツ Leandro Sales <leandroal@gmail.com> wrote:
> If it's any consolation then I went to test on one of my machines and
> it won't boot either. I'm running an older version of Gerrit's tree
> modified a lot (for faster restart testing).
>
> So I'll try doing the same sort of thing as you too now, but first
> I'll try Gerrit's more recent tree. Basically Gerrit and Arnaldo and

Well I've just pulled Gerrit's tree and it's based off 2.6.23rc3 or
later now and recompiled it and my problems have disappeared. You
might find the same too.

Ian
-- 
Web1: http://wand.net.nz/~iam4/
Web2: http://www.jandi.co.nz
Blog: http://iansblog.jandi.co.nz

Re: Kernel version of Arnaldos Net Git tree

[Gerrit Renker]

Ian, Leandro

I fully sympathise with your experiences, having just been through much hell of
rebasing the test tree to track Arnaldo's recent changes (dont' worry, it works
and I will post the changes) and then finding out that 2 out of 3 computers no longer
even boot due to changes somewhere else in the tree. 

I am open to suggestions as to when to resynchronise the tree. 

What I will do, in any case, is put up a diff from a working version online.

The test tree was synchronised yesterday against davem-2.6

Gerrit

Quoting Ian McDonald:
|  >    I tested using Linus and David Miller and the problem of watchdog
|  > didn't happen. But using both Arnaldo or Gerrit I got the watchdog
|  > problem! :( How often gerrit sync his git repository into Arnaldo git
|  > and how often Arnaldo do this into David Miller netdev git tree?
|  >
|  > Leandro.
|  >
|  If it's any consolation then I went to test on one of my machines and
|  it won't boot either. I'm running an older version of Gerrit's tree
|  modified a lot (for faster restart testing).
|  
|  So I'll try doing the same sort of thing as you too now, but first
|  I'll try Gerrit's more recent tree. Basically Gerrit and Arnaldo and
|  Dave sync their trees when they feel like it. Generally before
|  submitting upstream people tend to resync to check their patches still
|  apply. Gerrit and Arnaldo and I will often sync off Dave M's release+1
|  tree (currrently 2.6.24) as this is where networking changes go if the
|  merge window is not currently open (Linus only allows changes up to
|  rc1 of any release).
|  
|  What you and I might have to do is take Gerrit's tree and pull his
|  changes out of the tree (they are the most recent patches) and apply
|  them to Linus' latest tree. I don't know the syntax for this yet but
|  will tell you if I have to do this (please tell me the syntax if you
|  do it first!!). We will need Linus' tree rather than Dave
|  M's/Arnaldo's tree (even if they rebase) as at the moment with
|  Arnaldo's recent patch set I'm pretty sure Gerrit's patches won't
|  apply cleanly any more until he reworks his patch set. The joys of
|  distributed development!

Re: Kernel version of Arnaldos Net Git tree

[Ian McDonald]

On 8/21/07, Gerrit Renker <gerrit@erg.abdn.ac.uk> wrote:
> Ian, Leandro
>
> I fully sympathise with your experiences, having just been through much hell of
> rebasing the test tree to track Arnaldo's recent changes (dont' worry, it works
> and I will post the changes) and then finding out that 2 out of 3 computers no longer
> even boot due to changes somewhere else in the tree.
>
At least mine are working again, but versions are very unstable at the
moment. For example one machine won't run on a 2.6.22 Debian kernel at
the moment...

> I am open to suggestions as to when to resynchronise the tree.
>
I don't think there is any easy answers :-(

> What I will do, in any case, is put up a diff from a working version online.
>
> The test tree was synchronised yesterday against davem-2.6

2.6 or 2.6.24?? I thought Arnaldo's changes went into 24...

Ian

-- 
Web1: http://wand.net.nz/~iam4/
Web2: http://www.jandi.co.nz
Blog: http://iansblog.jandi.co.nz

Re: Kernel version of Arnaldos Net Git tree

[Gerrit Renker]

|  At least mine are working again, but versions are very unstable at the
|  moment. For example one machine won't run on a 2.6.22 Debian kernel at
|  the moment...
I was using the 2.6.24 netdev tree and it produced an oops for the e100 card;
hung when it tried to re-initialise the serial consoles; on a double Xeon PIII
will not boot; on the laptop there is apparently the ACPI problem again.

|  > The test tree was synchronised yesterday against davem-2.6
|  
|  2.6 or 2.6.24?? I thought Arnaldo's changes went into 24...
Oh no, yesterday it was 2.6.23-rc2/3 without Arnaldo's changes. 

In any case, I shall from now on put known-to-work test-tree patches per-kernel-version on
http://www.erg.abdn.ac.uk/users/gerrit/dccp/testing_dccp/test-tree/

The test tree is now resynched to match Arnaldo's set, but I need a separate email to
list the changes.

Re: Kernel version of Arnaldos Net Git tree

[@ 2007-08-21 18:41 UTC (permalink / raw)]

2007/8/21, Gerrit Renker <gerrit@erg.abdn.ac.uk>:
> Ian, Leandro

Hi Gerrit and Ian,

>
> I fully sympathise with your experiences, having just been through much hell of
> rebasing the test tree to track Arnaldo's recent changes (dont' worry, it works
> and I will post the changes) and then finding out that 2 out of 3 computers no longer
> even boot due to changes somewhere else in the tree.
>

Yes, this is my case and yesterday I used linus-git with your dccp
code and everything worked fine. Today I pulled from your git tree and
also worked, but my laptop froze to times and I need to reboot it, I
don't now why!

> I am open to suggestions as to when to resynchronise the tree.
>

I don't have any particular opinion about that... I think that if we
report some malfunction (like this with watchdog, APIC, and so on) in
your tree and you merge your tree with the latest linus-git (or davem)
it is enough.

> What I will do, in any case, is put up a diff from a working version online.
>
> The test tree was synchronised yesterday against davem-2.6
>

Good from now!

> Gerrit
>
> Quoting Ian McDonald:
> |  > I tested using Linus and David Miller and the problem of watchdog
> |  > didn't happen. But using both Arnaldo or Gerrit I got the watchdog
> |  > problem! :( How often gerrit sync his git repository into Arnaldo git
> |  > and how often Arnaldo do this into David Miller netdev git tree?
> |  >
> |  > Leandro.
> |  >
> |  If it's any consolation then I went to test on one of my machines and
> |  it won't boot either. I'm running an older version of Gerrit's tree
> |  modified a lot (for faster restart testing).
> |
> |  So I'll try doing the same sort of thing as you too now, but first
> |  I'll try Gerrit's more recent tree. Basically Gerrit and Arnaldo and
> |  Dave sync their trees when they feel like it. Generally before
> |  submitting upstream people tend to resync to check their patches still
> |  apply. Gerrit and Arnaldo and I will often sync off Dave M's release+1
> |  tree (currrently 2.6.24) as this is where networking changes go if the
> |  merge window is not currently open (Linus only allows changes up to
> |  rc1 of any release).
> |
> |  What you and I might have to do is take Gerrit's tree and pull his
> |  changes out of the tree (they are the most recent patches) and apply
> |  them to Linus' latest tree. I don't know the syntax for this yet but
> |  will tell you if I have to do this (please tell me the syntax if you
> |  do it first!!). We will need Linus' tree rather than Dave
> |  M's/Arnaldo's tree (even if they rebase) as at the moment with
> |  Arnaldo's recent patch set I'm pretty sure Gerrit's patches won't
> |  apply cleanly any more until he reworks his patch set. The joys of
> |  distributed development!
>

Re: Kernel version of Arnaldos Net Git tree

[Gerrit Renker]

|  Yes, this is my case and yesterday I used linus-git with your dccp
|  code and everything worked fine. Today I pulled from your git tree and
|  also worked, but my laptop froze to times and I need to reboot it, I
|  don't now why!
|
I think I know ... I had the same problems - three computers with e100 cards
just froze. It really seems that this is due to a half-finished merge window.

What I have done for the moment is - created a dummy patch which collects
all patches sent by Arnaldo, and put this as first patch into the tree; and rebased
everything onto davem-2.6 (which works fine, tested on 5 different computers).

I'll leave that until the 2.6.24 tree becomes a bit better, so that the code at least
works. Anything else - just let us know.

Gerrit
  

Re: Kernel version of Arnaldos Net Git tree

[Ian McDonald]

On 8/22/07, Gerrit Renker <gerrit@erg.abdn.ac.uk> wrote:
> I think I know ... I had the same problems - three computers with e100 cards
> just froze. It really seems that this is due to a half-finished merge window.
>
Gerrit - this is well worth reporting to netdev mailing list as the
2.6.24 later becomes the stable tree. I'm not saying you should fix
but it's worth reporting.

-- 
Web1: http://wand.net.nz/~iam4/
Web2: http://www.jandi.co.nz
Blog: http://iansblog.jandi.co.nz

Re: Kernel version of Arnaldos Net Git tree

[Gerrit Renker]

|  > I think I know ... I had the same problems - three computers with e100 cards
|  > just froze. It really seems that this is due to a half-finished merge window.
|  >
|  Gerrit - this is well worth reporting to netdev mailing list as the
|  2.6.24 later becomes the stable tree. I'm not saying you should fix
|  but it's worth reporting.
|  
I have cribbed the oops message from the screen and sent this to netdev. Today it is
still not better (same oops as yesterday).