linux-kernel - Re: [PATCH 2/2] x86: vdso: Remove redundant ifdeffery around in_ia32

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <5f32ff04-5e3a-49b2-8d97-c95c149d001f@t-8ch.de>
Date: Tue, 10 Sep 2024 20:10:33 +0200
From: Thomas Weißschuh <linux@...ssschuh.net>
To: "Eric W. Biederman" <ebiederm@...ssion.com>
Cc: Thomas Weißschuh <thomas.weissschuh@...utronix.de>, 
	Andy Lutomirski <luto@...nel.org>, Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>, 
	Dave Hansen <dave.hansen@...ux.intel.com>, x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>, 
	Kees Cook <kees@...nel.org>, Thomas Gleixner <tglx@...utronix.de>, 
	linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: [PATCH 2/2] x86: vdso: Remove redundant ifdeffery around
 in_ia32_syscall()

Hi Eric,

On 2024-09-10 09:34:46+0000, Eric W. Biederman wrote:
> Thomas Weißschuh <thomas.weissschuh@...utronix.de> writes:
> 
> > The ifdefs only guard code that is also guarded by in_ia32_syscall(),
> > which already contains the same ifdefs itself.
> >
> > Signed-off-by: Thomas Weißschuh <thomas.weissschuh@...utronix.de>
> > ---
> >  arch/x86/entry/vdso/vma.c | 4 ----
> >  1 file changed, 4 deletions(-)
> >
> > diff --git a/arch/x86/entry/vdso/vma.c b/arch/x86/entry/vdso/vma.c
> > index 9059b9d96393..ab2b011471e0 100644
> > --- a/arch/x86/entry/vdso/vma.c
> > +++ b/arch/x86/entry/vdso/vma.c
> > @@ -75,7 +75,6 @@ static vm_fault_t vdso_fault(const struct vm_special_mapping *sm,
> >  static void vdso_fix_landing(const struct vdso_image *image,
> >  		struct vm_area_struct *new_vma)
> >  {
> > -#if defined CONFIG_X86_32 || defined CONFIG_IA32_EMULATION
> >  	if (in_ia32_syscall() && image == &vdso_image_32) {
> >  		struct pt_regs *regs = current_pt_regs();
> >  		unsigned long vdso_land = image->sym_int80_landing_pad;
> > @@ -86,7 +85,6 @@ static void vdso_fix_landing(const struct vdso_image *image,
> >  		if (regs->ip == old_land_addr)
> >  			regs->ip = new_vma->vm_start + vdso_land;
> >  	}
> > -#endif
> >  }
> >  
> >  static int vdso_mremap(const struct vm_special_mapping *sm,
> > @@ -339,7 +337,6 @@ int compat_arch_setup_additional_pages(struct linux_binprm *bprm,
> >  
> >  bool arch_syscall_is_vdso_sigreturn(struct pt_regs *regs)
> >  {
> > -#if defined(CONFIG_X86_32) || defined(CONFIG_IA32_EMULATION)
> >  	const struct vdso_image *image = current->mm->context.vdso_image;
> >  	unsigned long vdso = (unsigned long) current->mm->context.vdso;
> >  
> > @@ -348,7 +345,6 @@ bool arch_syscall_is_vdso_sigreturn(struct pt_regs *regs)
> >  		    regs->ip == vdso + image->sym_vdso32_rt_sigreturn_landing_pad)
> >  			return true;
> >  	}
> > -#endif
> >  	return false;
> >  }
> 
> Have you tested to verify that after this change
> arch_syscall_is_vdso_signature compiles out the "image" and "vdso"
> variables?

Yes, I did:

$ objdump --disassemble=arch_syscall_is_vdso_sigreturn arch/x86/entry/vdso/vma.o
arch/x86/entry/vdso/vma.o:     file format elf64-x86-64

Disassembly of section .text:

00000000000007f0 <arch_syscall_is_vdso_sigreturn>:
 7f0:	f3 0f 1e fa          	endbr64
 7f4:	e8 00 00 00 00       	call   7f9 <arch_syscall_is_vdso_sigreturn+0x9>
 7f9:	31 c0                	xor    %eax,%eax
 7fb:	e9 00 00 00 00       	jmp    800 <arch_syscall_is_vdso_sigreturn+0x10>


> 
> If the compilers don't it might be worth it rearrange the code as:
> 	if (in_ia32_syscall()) {
> 		const struct vdso_image *image = current->mm->context.vdso_image;
> 		unsigned long vdso = (unsigned long) current->mm->context.vdso;
> 
> 		if (image == &vdso_image_32) {
>                 	....
>                         return true;
>                 }
> 	}
>         return false.
> 
> Making the variables depend upon in_ia32_syscall() so you can be certain
> they are compiles out.

If that structure is preferred I can send a v2.


Thomas