Hi,
in setup.c the kernel tries to reserve ram resources for system ram etc
etc. However it seems it's done with the parameters to request_resource in
the wrong order (it certainly is opposite order from other neighboring
code). Can someone confirm I'm not overlooking something?
Greetings,
Arjan van de Ven
--- linux-2.6.0/arch/i386/kernel/setup.c~ 2004-01-04 16:37:34.622450000 +0100
+++ linux-2.6.0/arch/i386/kernel/setup.c 2004-01-04 16:37:34.622450000 +0100
@@ -834,8 +834,8 @@
* so we try it repeatedly and let the resource manager
* test it.
*/
- request_resource(res, &code_resource);
- request_resource(res, &data_resource);
+ request_resource(&code_resource, res);
+ request_resource(&data_resource, res);
}
}
On Sun, Jan 04, 2004 at 01:08:52PM -0800, Linus Torvalds wrote:
>
>
> On Sun, 4 Jan 2004, Arjan van de Ven wrote:
> >
> > in setup.c the kernel tries to reserve ram resources for system ram etc
> > etc. However it seems it's done with the parameters to request_resource in
> > the wrong order (it certainly is opposite order from other neighboring
> > code). Can someone confirm I'm not overlooking something?
>
> You've overlooked something.
>
> The core uses the rigth order: it's literally trying to find _which_ of
> the e820 resources contains the "code" and "data" resource.
>
> In other words: the code and data resources don't contain anything. They
> are contained _in_ something, but we don't know which one off-hand, so we
> try to register them in all the memory resources we find.
> and not used for anything else.
ok fair enough; maybe deserves more comment but it makes sense.
On Sun, Jan 04, 2004 at 01:08:52PM -0800, Linus Torvalds wrote:
>
>
> On Sun, 4 Jan 2004, Arjan van de Ven wrote:
> >
> > in setup.c the kernel tries to reserve ram resources for system ram etc
> > etc. However it seems it's done with the parameters to request_resource in
> > the wrong order (it certainly is opposite order from other neighboring
> > code). Can someone confirm I'm not overlooking something?
>
> You've overlooked something.
>
> The core uses the rigth order: it's literally trying to find _which_ of
> the e820 resources contains the "code" and "data" resource.
>
> In other words: the code and data resources don't contain anything. They
> are contained _in_ something, but we don't know which one off-hand, so we
> try to register them in all the memory resources we find.
> and not used for anything else.
ok fair enough; maybe deserves more comment but it makes sense.