[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200226100256.GK14946@hirez.programming.kicks-ass.net>
Date: Wed, 26 Feb 2020 11:02:56 +0100
From: Peter Zijlstra <peterz@...radead.org>
To: Thomas Gleixner <tglx@...utronix.de>
Cc: LKML <linux-kernel@...r.kernel.org>, x86@...nel.org,
Steven Rostedt <rostedt@...dmis.org>,
Brian Gerst <brgerst@...il.com>,
Juergen Gross <jgross@...e.com>,
Paolo Bonzini <pbonzini@...hat.com>,
Arnd Bergmann <arnd@...db.de>
Subject: Re: [patch 00/15] x86/entry: Consolidation - Part V
On Wed, Feb 26, 2020 at 10:53:19AM +0100, Peter Zijlstra wrote:
> +SYM_CODE_START_LOCAL_NOALIGN(common_idtentry)
> + /* the return address is in the %gs stack slot */
> SAVE_ALL switch_stacks=1 skip_gs=1 unwind_espfix=1
> ENCODE_FRAME_POINTER
>
> /* fixup %gs */
> GS_TO_REG %ecx
> - movl PT_GS(%esp), %edi # get the function address
> - REG_TO_PTGS %ecx
> + pushl PT_GS(%esp) # push return address
> + REG_TO_OTGS %ecx
Aside from the obvious typo, it is also completely broken because
REG_TO_PGTS relies on the stack layout, which we just wrecked.
movl PT_GS(%esp), %edi # get the return address
REG_TO_PTGS %ecx
> SET_KERNEL_GS %ecx
>
> /* fixup orig %eax */
> @@ -1348,9 +1348,8 @@ SYM_CODE_START_LOCAL_NOALIGN(common_exception)
> movl $-1, PT_ORIG_EAX(%esp) # no syscall to restart
>
> movl %esp, %eax # pt_regs pointer
pushl %edi
> + ret
> +SYM_CODE_END(common_idtentry)
Should work, although that push+ret combo is a bit awkward.
Powered by blists - more mailing lists