Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions

[PATCH v2 00/12] x86/msr: Inline rdmsr/wrmsr instructions · Juergen Gross <jgross@suse.com> · 2025-09-30
[PATCH v2 05/12] x86/msr: Move MSR trace calls one function level up · Juergen Gross <jgross@suse.com> · 2025-09-30
[PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · Juergen Gross <jgross@suse.com> · 2025-09-30
Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · Peter Zijlstra <peterz@infradead.org> · 2025-09-30
Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · Jürgen Groß <jgross@suse.com> · 2025-09-30
Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · Peter Zijlstra <peterz@infradead.org> · 2025-09-30
Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · Jürgen Groß <jgross@suse.com> · 2025-09-30
Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · "H. Peter Anvin" <hpa@zytor.com> · 2025-09-30
Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · "H. Peter Anvin" <hpa@zytor.com> · 2025-09-30
Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · Peter Zijlstra <peterz@infradead.org> · 2025-10-01
Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · kernel test robot <hidden> · 2025-09-30
Re: [PATCH v2 11/12] x86/paravirt: Don't use pv_ops vector for MSR access functions · Jürgen Groß <jgross@suse.com> · 2025-10-01
Re: [PATCH v2 00/12] x86/msr: Inline rdmsr/wrmsr instructions · "H. Peter Anvin" <hpa@zytor.com> · 2025-09-30

From: Peter Zijlstra <peterz@infradead.org>
Date: 2025-10-01 06:45:53
Also in: lkml, xen-devel

On Tue, Sep 30, 2025 at 12:49:21PM -0700, H. Peter Anvin wrote:

/* Xen code, stub sets CF = 1 on failure */

   0:   e8 xx xx xx xx          call   asm_xen_pv_wrmsr
   5:   73 03                   jnc    0xa
   7:   0f 0b                   ud2
   9:   90                      nop
   a:

The trap point even ends up in the same place! UD2 can be any 1-, 2-, or
3-byte trapping instruction.

Please don't rely on flags to be retained by RET. The various
mitigations have trouble with that.

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help