12:45 p.m.
I then copied YOUR config (per your email) to .config
I then ran
make old config
and it ended up asking me a huge plethora of questions.
I simply answered by pressing the Enter key.
Now:
If the config file you had sent was the RESULT of
STARTING with your 2.6.34.x config file AND
running
make oldconfig
I would not have seen all those questions pop up
when I ran make old config using your .config.
So, I think you did something wrong.
Thanks JD.
What I did initially is copy my .34.x .config file and after that I ran
"make oldconfig", assuming that for any new/missing options that there
are (i.e. the difference between the .34 and the new .35 version of the
kernel) I will be asked for. That is, what I thought, the sole reason of
running the oldconfig target.
Having *just* done diff comparison between the 'stock' .config and my
.config (the one I attached here) - which I thought *should* amalgamate
all the options between the 2 versions available (i.e. the result of
running make oldconfig) - I discovered that quite a few of the options
present in my .config are not there!
The omission of a crucial group, which in my opinion produces the error
I described in my original post, is this:
CONFIG_DRM_VMWGFX
CONFIG_DRM_NOVEAU**
CONFIG_DRM_NOVEAU_BACKLIGHT
CONFIG_DRM_NOVEAU_DEBUG
The above produces, among other things, noveau.ko (marked with '**'
above), which is the driver used for my display here. Having also
scanned the stock-produced initramfs and the initramfs file generated
after I install the kernel, I can confirm that drivers/noveau/noveau.ko
is NOT there (in my 'custom-build/generated' initramfs file), so the
source of my not-so-good experiences with the .35 kernel, I think, has
been established.
The question is - have I did something wrong by 1) copying an 'old'
(.34) .config; 2) running make oldconfig; and 3) make menuconfig,
thinking that the oldconfig target will take care of the options which
are missing in the 'old' version of the .config file and ask me to fill
in the gaps (i.e. asking all those questions)?
If that is what is supposed to happen, than the oldconfig target is not
doing its job as there are plethora of kernel options I have not been
asked, nor were they included in the resultant .config file (the example
with CONFIG_DRM_NOVEAU I have given above).
If, on the other hand, I have done this the wrong way and the oldconfig
target is supposed to be used only between minor revisions of the
kernel, then, as I did not know that, this is something I will bear in
mind in the future.
I have been building kernels for more than 35 years,
so I know what I am saying is correct.
I do not question your ability in any way, I
am grateful for the help
you provided me with. That, unfortunately, cannot be said for the other
two window-licking jamooks who 'participated' in this thread!
1:03 p.m.
New subject: 2.6.35.10-74 compilation (and build) problems
On Wed, 2011-01-05 at 18:45 +0000, Mr Dash Four wrote:
That, unfortunately, cannot be said for the other
two [whatever] who 'participated' in this thread!
Is this list monitored by the Hall Monitors (see:
http://fedoraproject.org/wiki/Hall_Monitor_Policy )? If not, as it seems
to be, should it be? Or should it be regulated on a case by case basis
(and who is in a position to do so)?
Paul Bolle
1:15 p.m.
New subject: 2.6.35.10-74 compilation (and build) problems
On Wed, Jan 5, 2011 at 2:03 PM, Paul Bolle <pebolle(a)tiscali.nl> wrote:
On Wed, 2011-01-05 at 18:45 +0000, Mr Dash Four wrote:
> That, unfortunately, cannot be said for the other
> two [whatever] who 'participated' in this thread!
Is this list monitored by the Hall Monitors (see:
http://fedoraproject.org/wiki/Hall_Monitor_Policy )? If not, as it seems
to be, should it be? Or should it be regulated on a case by case basis
(and who is in a position to do so)?
Hall Monitors is dead.
josh
1:30 p.m.
New subject: 2.6.35.10-74 compilation (and build) problems
On Jan 5, 2011, at 2:24 PM, Paul Bolle wrote:
On Wed, 2011-01-05 at 14:15 -0500, Josh Boyer wrote:
> Hall Monitors is dead.
I see. So who's going to deal with this situation?
Well, if they see fit, the list maintainers have the power to kick
people from the list. I'm still more amused than anything myself.
--
Jarod Wilson
jarod(a)wilsonet.com
1:38 p.m.
New subject: 2.6.35.10-74 compilation (and build) problems
On Wed, Jan 05, 2011 at 02:30:51PM -0500, Jarod Wilson wrote:
On Jan 5, 2011, at 2:24 PM, Paul Bolle wrote:
> On Wed, 2011-01-05 at 14:15 -0500, Josh Boyer wrote:
>> Hall Monitors is dead.
>
> I see. So who's going to deal with this situation?
Well, if they see fit, the list maintainers have the power to kick
people from the list. I'm still more amused than anything myself.
(list maintainer hat on)
how about we all just end this thread here and move on ?
Dave
1:15 p.m.
New subject: 2.6.35.10-74 compilation (and build) problems
On Jan 5, 2011, at 1:45 PM, Mr Dash Four wrote:
What I did initially is copy my .34.x .config file and after that I
ran
"make oldconfig", assuming that for any new/missing options that there
are (i.e. the difference between the .34 and the new .35 version of the
kernel) I will be asked for. That is, what I thought, the sole reason of
running the oldconfig target.
Having *just* done diff comparison between the 'stock' .config and my
.config (the one I attached here) - which I thought *should* amalgamate
all the options between the 2 versions available (i.e. the result of
running make oldconfig) - I discovered that quite a few of the options
present in my .config are not there!
The omission of a crucial group, which in my opinion produces the error
I described in my original post, is this:
CONFIG_DRM_VMWGFX
CONFIG_DRM_NOVEAU**
CONFIG_DRM_NOVEAU_BACKLIGHT
CONFIG_DRM_NOVEAU_DEBUG
The above produces, among other things, noveau.ko (marked with '**'
above), which is the driver used for my display here. Having also
scanned the stock-produced initramfs and the initramfs file generated
after I install the kernel, I can confirm that drivers/noveau/noveau.ko
is NOT there (in my 'custom-build/generated' initramfs file), so the
source of my not-so-good experiences with the .35 kernel, I think, has
been established.
The question is - have I did something wrong by 1) copying an 'old'
(.34) .config; 2) running make oldconfig; and 3) make menuconfig,
thinking that the oldconfig target will take care of the options which
are missing in the 'old' version of the .config file and ask me to fill
in the gaps (i.e. asking all those questions)?
Running make oldconfig is fine, provided you actually know the right
answers to get a config that matches what Fedora ships. Starting from
the stock config for the kernel you're rebuilding is a far saner route
than using one from an old kernel though. Its even the method used in
the "Building a custom kernel" wiki page I provided a link to. One more
time, just in case:
https://fedoraproject.org/wiki/Building_a_custom_kernel
Had you followed that, your kernel would have probably worked just
fine on the first try.
If that is what is supposed to happen, than the oldconfig target is
not
doing its job as there are plethora of kernel options I have not been
asked, nor were they included in the resultant .config file (the example
with CONFIG_DRM_NOVEAU I have given above).
If, on the other hand, I have done this the wrong way and the oldconfig
target is supposed to be used only between minor revisions of the
kernel, then, as I did not know that, this is something I will bear in
mind in the future.
See above.
> I have been building kernels for more than 35 years,
> so I know what I am saying is correct.
I do not question your ability in any way, I am grateful for the help
you provided me with. That, unfortunately, cannot be said for the other
two window-licking jamooks who 'participated' in this thread!
Just fyi, its considered bad etiquette to reply to private responses
with the list cc'd. Not a single one of JD's mails went to the list.
And eloquent though yours may be, slinging insults around is also
generally considered bad etiquette, and goes against Fedora project
policies, per the link Paul just provided.
--
Jarod Wilson
jarod(a)wilsonet.com
1:44 p.m.
New subject: 2.6.35.10-74 compilation (and build) problems
Running make oldconfig is fine, provided you actually know the right
answers to get a config that matches what Fedora ships.
Read my answer more
carefully. I was not asked whether or not to include
the noveau (and various other options, which I did not specify in my
previous post) and as a result of this, the video driver was not
included. Have I been asked or, at least, the appropriate options
'masked' with '#' and 'not set' I would have been alerted to this
and
would have fixed this error and not gone through with posting on here
(may be!).
As it turns out, I was not asked for these missing options and they were
not included in the resultant .config file.
Starting from the stock config for the kernel you're rebuilding
is a far saner route
than using one from an old kernel though.
Whether it is 'saner' is a matter
of personal opinion!
I have been building my kernel from source and depending on the target
set of machines am using a specific set of options designed for these.
If I am supposed to just start from scratch with every (major?) revision
of the kernel and not run the oldconfig target, then that's fine,
although I think this is wrong!
For me it makes more sense to use the oldconfig target to fill in the
gaps left between the *old* config and the new one as that target (is
supposed to) ask the appropriate questions for all options not specified
in that .config file.
If you think that it makes more sense to input a few hundred options
every now and again when a new version of the kernel gets released, then
go right ahead - for me it makes more sense oldconfig to be used to fill
in the gaps.
Its even the method used in the "Building a custom kernel"
wiki page I provided a link to. One more
time, just in case:
https://fedoraproject.org/wiki/Building_a_custom_kernel
Frankly, I do not give two hoots what some random wiki page
designed/created by Fedora says. Do you follow every Fedora wiki page
with instructions to the letter? If so, then I suppose you are very
happy man using Anaconda in the way it is described in the relevant wiki
page then, eh?
Had you followed that, your kernel would have probably worked just
fine on the first try.
Yeah, after manually typing/transferring a couple of hundred options
from my old config, sure...
> I do not question your ability in any way, I am grateful for the
help
> you provided me with. That, unfortunately, cannot be said for the other
> two window-licking jamooks who 'participated' in this thread!
>
Just fyi, its considered bad etiquette to reply to private responses
with the list cc'd. Not a single one of JD's mails went to the list.
So, how come when I send *you* the same email with the list cc'd it gets
there, eh?
And eloquent though yours may be, slinging insults around is also
generally considered bad etiquette, and goes against Fedora project
policies, per the link Paul just provided.
I am not the one who started with the insults - have a look at the
second post of this thread and tell me whether this constitutes "Be
excellent to each other" - a policy you profess to know, love and follow
to the letter!
I am not even going to comment on the post from that Michal individual
as 1) it is not in English (I though the language used in this list is
USA English, right?) and 2) it is as far as you can possibly get from
the "Be excellent to each other" policy you profess to know and follow.
We have a say here in my part of the land where I live - you reap what
you saw, so there...
4867
days inactive
4867
days old
kernel@lists.fedoraproject.org
7 comments
5 participants
participants (5)
-
Dave Jones -
Jarod Wilson -
Josh Boyer -
Mr Dash Four -
Paul Bolle