Jeff Dike <[EMAIL PROTECTED]> ha scritto:
> On Sun, Aug 06, 2006 at 05:47:00PM +0200, Paolo 'Blaisorblade'
> Giarrusso wrote:
> > +_extra_flags_ = -fno-builtin -m64 -mcmodel=kernel
> What exactly does this do
go to "man gcc" and search mcmodel for the answer to this one.
And x86_64 uses it too,
Jeff Dike <[EMAIL PROTECTED]> ha scritto:
> On Sun, Aug 06, 2006 at 05:47:03PM +0200, Paolo 'Blaisorblade'
> Giarrusso wrote:
> > From: Paolo 'Blaisorblade' Giarrusso <[EMAIL PROTECTED]>
> >
> > This spinlock can be taken on interrupt too, so
> spin_lock_irq[save] must be used.
> >
> > However,
Paolo 'Blaisorblade' Giarrusso <[EMAIL PROTECTED]> writes:
> From: Paolo 'Blaisorblade' Giarrusso <[EMAIL PROTECTED]>
>
> We have never used this flag and recently one user experienced a complaining
> warning about this (there was a symbol in the positive half of the address
> space
> IIRC). So
Andi Kleen <[EMAIL PROTECTED]> ha scritto:
> Paolo 'Blaisorblade' Giarrusso <[EMAIL PROTECTED]> writes:
>
> > From: Paolo 'Blaisorblade' Giarrusso <[EMAIL PROTECTED]>
> >
> > We have never used this flag and recently one user experienced a
> complaining
> > warning about this (there was a symbo
On Tuesday 08 August 2006 16:03, Paolo Giarrusso wrote:
> Andi Kleen <[EMAIL PROTECTED]> ha scritto:
>
> > Paolo 'Blaisorblade' Giarrusso <[EMAIL PROTECTED]> writes:
> >
> > > From: Paolo 'Blaisorblade' Giarrusso <[EMAIL PROTECTED]>
> > >
> > > We have never used this flag and recently one user
On Tue, 2006-08-08 at 16:03 +0200, Paolo Giarrusso wrote:
> Moreover, who has recently tested module loading in x86-64 uml
> kernels? I don't remember doing such testing recently...
>
Well, modules seem to work fine here on x86_64. I'm running 2.6.16-bs2.
I only tested the loop module, but othe
On Tuesday 08 August 2006 06:52, Alejandro Liu wrote:
> Hi,
>
> What is a "current" version of UML to use as a guest? I am currently using
> the one that comes with the vanilla 2.6.17.6 kernel and trying to use
> hostfs as root and found a number of bugs.
> Of course I went ahead and fixed those
On Tue, Aug 08, 2006 at 12:59:05PM +0200, Paolo Giarrusso wrote:
> I could be wrong, but I trust that thanks to deep and good work by
> who designed locking in the network layer, this patch is correct. And
> indeed I addressed your issues below.
OK, but there will need to be comments explaining wh
The use of SEGMENT_RPL_MASK in the i386 ptrace.h introduced by
x86-allow-a-kernel-to-not-be-in-ring-0.patch broke the UML build, as
UML includes the underlying architecture's ptrace.h, but has no easy
access to the x86 segment definitions.
Rather than kludging around this, as in the past, this pat
Include the host architecture's ptrace-abi.h instead of ptrace.h.
There was some cpp mangling of names around the ptrace.h include to
avoid symbol clashes between UML and the host architecture. Most of
these can go away. The exception is struct pt_regs, which is
convenient to have in userspace,
Jeff Dike <[EMAIL PROTECTED]> writes:
>
> x86_64 is also treated in this way for consistency.
>
> This patch was run by linux-arch yesterday with no comment.
Well you should have sent it to me.
I think I would prefer a well placed ifdef __KERNEL__ or two for this - i don't
like it when it beco
On Wed, Aug 09, 2006 at 01:33:43AM +0200, Andi Kleen wrote:
> I think I would prefer a well placed ifdef __KERNEL__ or two for this - i
> don't
> like it when it becomes harder to grep include files like this
> (like the errno->errno-base split was quite bad in this regard)
__KERNEL__ doesn't hel
12 matches
Mail list logo
