Skip to contentSkip to conversation

rename generic_defconfig to defconfig?

DISCUSSION Jesse Barnes 5 participants 14 messages Sep 22, 2004 – Nov 2, 2004

lore
Share & tools
lore
#0 lore raw

Just thought I'd float the idea of making the ia64 defconfig into a generic ia64 kernel config. It's probably easiest just to rename generic_defconfig to defconfig since that already works everywhere (well, except for akpm's box apparently). That would make it extra easy to build a working kernel for the various ia64 boxes out there...

Jesse

#1 lore raw
Quoted 6 lines

Just thought I'd float the idea of making the ia64 defconfig into a generic ia64 kernel config. It's probably easiest just to rename generic_defconfig to defconfig since that already works everywhere (well, except for akpm's box apparently). That would make it extra easy to build a working kernel for the various ia64 boxes out there...

Sounds plausible. I'd also like to keep some more system specific config files updated as well (to make it easier to see which combinations of NUMA/DISCONTIG/VIRTUAL_MEM_MAP are being used ... and thus try to avoid changes that prevent them from even building).

Does anyone have a zx1_defconfig that works for 2.6.9-rc2? I just tried to use the current one to rebuild for my zx2000 and the kernel made the system sing like a shortwave radio without printing a single character to the screen :-(

-Tony

#2 lore raw
In reply to Jesse Barnes
Quoted 4 lines · Tony

Does anyone have a zx1_defconfig that works for 2.6.9-rc2? I just tried to use the current one to rebuild for my zx2000 and the kernel made the system sing like a shortwave radio without printing a single character to the screen :-(

zx1_defconfig works for me on rx8620 and rx4640 ...

Signature · 6 lines
"Next the statesmen will invent cheap lies, putting the blame upon 
the nation that is attacked, and every man will be glad of those
conscience-soothing falsities, and will diligently study them, and refuse
to examine any refutations of them; and thus he will by and by convince 
himself that the war is just, and will thank God for the better sleep 
he enjoys after this process of grotesque self-deception." -- Mark Twain
#3 lore raw
In reply to Jesse Barnes
Quoted 7 lines · Tony
Quoted 6 lines

Just thought I'd float the idea of making the ia64 defconfig into a generic ia64 kernel config. It's probably easiest just to rename generic_defconfig to defconfig since that already works everywhere (well, except for akpm's box apparently). That would make it extra easy to build a working kernel for the various ia64 boxes out there...

Quoted 5 lines

Sounds plausible. I'd also like to keep some more system specific config files updated as well (to make it easier to see which combinations of NUMA/DISCONTIG/VIRTUAL_MEM_MAP are being used ... and thus try to avoid changes that prevent them from even building).

I could come up with a bigsur_defconfig next week (my machine is due for a kernel upgrade).

Jesse

#4 lore raw
In reply to Jesse Barnes
Quoted 5 lines · Jesse Barnes

Just thought I'd float the idea of making the ia64 defconfig into a generic ia64 kernel config. It's probably easiest just to rename generic_defconfig to defconfig since that already works everywhere (well, except for akpm's box apparently). That would make it extra easy to build a working kernel for the various ia64 boxes out there...

Sounds like a good idea to me. If the idea is to make it as easy as possible to build working kernels, maybe we should also make some things static rather than modules. For starters:

CONFIG_TULIP (HP superdome core I/O and 4-port cards) CONFIG_E100 (i2000)
CONFIG_E1000 (Tiger)
CONFIG_USB_EHCI_HCD (HP boxes) CONFIG_USB_OHCI_HCD (HP boxes) CONFIG_USB_HID

Or maybe I'm the only one who's too lazy to mess with modules for every build.

I think generic_defconfig currently builds a McKinley kernel, which may be the right thing, but you do have to remember to switch back to CONFIG_ITANIUM for i2000.

#5 lore raw
In reply to Jesse Barnes
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Jesse" = Jesse Barnes <[email protected]> writes:

Quoted 6 lines · Tony
Quoted 6 lines
Quoted 1 line

Just thought I'd float the idea of making the ia64 defconfig >into

a generic ia64 kernel config. It's probably easiest just >to rename generic_defconfig to defconfig since that already >works everywhere (well, except for akpm's box apparently). >That would make it extra easy to build a working kernel for >the various ia64 boxes out there...

Generic_defconfig still doesn't work on the simulator, yet. So we at least need that config hanging around too.

Signature · 2 lines
Dr Peter Chubb  http://www.gelato.unsw.edu.au  peterc AT gelato.unsw.edu.au
The technical we do immediately,  the political takes *forever*
#6 lore raw
In reply to Jesse Barnes
Quoted 2 lines

I could come up with a bigsur_defconfig next week (my machine is due for a kernel upgrade).

Attached is the config that I'm using on my Tiger at the moment (derived from generic by turning on DIG, turning off NUMA, dropping cpus from 512 to 4, and messing with a few device drivers).

-Tony

#7 lore raw
In reply to Jesse Barnes
Quoted 7 lines · Tony
Quoted 2 lines

I could come up with a bigsur_defconfig next week (my machine is due for a kernel upgrade).

Attached is the config that I'm using on my Tiger at the moment (derived from generic by turning on DIG, turning off NUMA, dropping cpus from 512 to 4, and messing with a few device drivers).

I'm pretty sure CONFIG_VIRTUAL_MEM_MAP isn't needed either. Also, which device drivers did you have to tweak? Should those changes also go into generic_defconfig?

Thanks,
Jesse

#8 lore raw
In reply to Jesse Barnes
Quoted 18 lines · Bjorn Helgaas
Quoted 5 lines · Jesse Barnes

Just thought I'd float the idea of making the ia64 defconfig into a generic ia64 kernel config. It's probably easiest just to rename generic_defconfig to defconfig since that already works everywhere (well, except for akpm's box apparently). That would make it extra easy to build a working kernel for the various ia64 boxes out there...

Sounds like a good idea to me. If the idea is to make it as easy as possible to build working kernels, maybe we should also make some things static rather than modules. For starters:

CONFIG_TULIP (HP superdome core I/O and 4-port cards) CONFIG_E100 (i2000)
CONFIG_E1000 (Tiger)
CONFIG_USB_EHCI_HCD (HP boxes) CONFIG_USB_OHCI_HCD (HP boxes) CONFIG_USB_HID

Quoted 2 lines

Or maybe I'm the only one who's too lazy to mess with modules for every build.

No, I think you have a point there :) I *really* don't like having to copy modules over. If tg3 isn't builtin yet, it would be nice too.

Quoted 3 lines

I think generic_defconfig currently builds a McKinley kernel, which may be the right thing, but you do have to remember to switch back to CONFIG_ITANIUM for i2000.

Right, but those boxes are mainly for the masochists among us, so I think a McKinley setting is the right choice.

Jesse

#9 lore raw
In reply to Jesse Barnes
Quoted 6 lines

I'm pretty sure CONFIG_VIRTUAL_MEM_MAP isn't needed either. Also, which
device drivers did you have to tweak? Should those changes also go into
generic_defconfig?

Tigers with more than 2G of memory have a hole in the physical address space (from 2G to 4G) reserved for memory mapped I/O. So VIRTUAL_MEM_MAP (or I suppose DISCONTIG) is needed to avoid wasting memory on page structs for non-existant memory.

The device driver bits are changing 'M' to 'Y' on devices. Like Bjorn I'm too lazy to mess with modules on every build.

-Tony

#10 lore raw
In reply to Jesse Barnes
Quoted 6 lines · Jesse Barnes
Quoted 2 lines

Or maybe I'm the only one who's too lazy to mess with modules for every build.

No, I think you have a point there :) I *really* don't like having to copy modules over. If tg3 isn't builtin yet, it would be nice too.

Quoted 6 lines
Quoted 3 lines

I think generic_defconfig currently builds a McKinley kernel, which may be the right thing, but you do have to remember to switch back to CONFIG_ITANIUM for i2000.

Right, but those boxes are mainly for the masochists among us, so I think a McKinley setting is the right choice.

Given the above, generic_defconfig seems to be in good shape to takeover as defconfig. Tony?

Jesse

#11 lore raw
In reply to Jesse Barnes
Quoted 2 lines

Given the above, generic_defconfig seems to be in good shape to takeover as defconfig. Tony?

That sounds ok to me ... "arch/ia64/defconfig" doesn't really fit anyone's machine right now (does it ... speak up now if you are somehow depending on it).

So I assume the proposal to to update defconfig with the current contents of arch/ia64/configs/gen_defconfig, and then delete gen_defconfig? That way defconfig is something that actually builds and boots on a fairly wide variety of machines.

-Tony

#12 lore raw
In reply to Jesse Barnes
Quoted 6 lines · Tony
Quoted 2 lines

Given the above, generic_defconfig seems to be in good shape to takeover as defconfig. Tony?

That sounds ok to me ... "arch/ia64/defconfig" doesn't really fit anyone's machine right now (does it ... speak up now if you are somehow depending on it).

It looks like some sort of zx1 config.

Quoted 4 lines

So I assume the proposal to to update defconfig with the current contents of arch/ia64/configs/gen_defconfig, and then delete gen_defconfig? That way defconfig is something that actually builds and boots on a fairly wide variety of machines.

Yep, that sounds great.

Jesse

#13 lore raw
In reply to Jesse Barnes
Quoted 6 lines
Quoted 4 lines

So I assume the proposal to to update defconfig with the current contents of arch/ia64/configs/gen_defconfig, and then delete gen_defconfig? That way defconfig is something that actually builds and boots on a fairly wide variety of machines.

Yep, that sounds great.

Ok. I just pushed a change up to bkbits.net (whole defconfig file attached to this e-mail for non-BK users). It boots on Tiger, and meets my module-pain threshhold (doesn't require initrd to boot), but it does pull in a bunch of USB and ACPI modules after it boots. I'm happy to take more patches to make it run standalone.

-Tony