[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHk-=whfQoTDCd=8DYfCiX0a2ndqM-mmoxDm1xA7Kud+WQ9T8w@mail.gmail.com>
Date: Tue, 16 Jul 2024 16:18:49 -0700
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Nathan Chancellor <nathan@...nel.org>
Cc: linux-um@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: Crash when booting UML after e3c92e81711d14b46c3121d36bc8e152cb843923
On Tue, 16 Jul 2024 at 07:36, Nathan Chancellor <nathan@...nel.org> wrote:
>
> Our continuous integration notices a crash when booting User Mode Linux
> after commit e3c92e81711d ("runtime constants: add x86 architecture
> support")
Ahh.
Yes, UML should *not* be using the x86-specific (or any "native")
runtime constants, and that was never the intent.
But it turns out UML ends up just blindly getting it, because it ends
up using the native architecture header files here when it exists.
How very annoying.
I'm not entirely sure how to tell UML to use the generic header and
not the native one.
Hmm... Does this simple (whitespace-damaged) patch fix it for you?
--- a/arch/um/include/asm/Kbuild
+++ b/arch/um/include/asm/Kbuild
@@ -19,6 +19,7 @@ generic-y += param.h
generic-y += parport.h
generic-y += percpu.h
generic-y += preempt.h
+generic-y += runtime-const.h
generic-y += softirq_stack.h
generic-y += switch_to.h
generic-y += topology.h
basically telling UML to not use the runtime-const.h file from the
native architecture..
Somebody who knows UML better can probably confirm whether this is the
right way to say "don't use the low-level architecture file".
Linus
Powered by blists - more mailing lists