linux-kernel - Re: [PATCH v3 09/29] mm: abstract shadow stack vma behind `vma_is_shadow

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

Date: Thu, 4 Apr 2024 21:02:17 +0200
From: David Hildenbrand <david@...hat.com>
To: Deepak Gupta <debug@...osinc.com>, paul.walmsley@...ive.com,
 rick.p.edgecombe@...el.com, broonie@...nel.org, Szabolcs.Nagy@....com,
 kito.cheng@...ive.com, keescook@...omium.org, ajones@...tanamicro.com,
 conor.dooley@...rochip.com, cleger@...osinc.com, atishp@...shpatra.org,
 alex@...ti.fr, bjorn@...osinc.com, alexghiti@...osinc.com,
 samuel.holland@...ive.com, conor@...nel.org
Cc: linux-doc@...r.kernel.org, linux-riscv@...ts.infradead.org,
 linux-kernel@...r.kernel.org, devicetree@...r.kernel.org,
 linux-mm@...ck.org, linux-arch@...r.kernel.org,
 linux-kselftest@...r.kernel.org, corbet@....net, palmer@...belt.com,
 aou@...s.berkeley.edu, robh+dt@...nel.org,
 krzysztof.kozlowski+dt@...aro.org, oleg@...hat.com,
 akpm@...ux-foundation.org, arnd@...db.de, ebiederm@...ssion.com,
 Liam.Howlett@...cle.com, vbabka@...e.cz, lstoakes@...il.com,
 shuah@...nel.org, brauner@...nel.org, andy.chiu@...ive.com,
 jerry.shih@...ive.com, hankuan.chen@...ive.com, greentime.hu@...ive.com,
 evan@...osinc.com, xiao.w.wang@...el.com, charlie@...osinc.com,
 apatel@...tanamicro.com, mchitale@...tanamicro.com,
 dbarboza@...tanamicro.com, sameo@...osinc.com, shikemeng@...weicloud.com,
 willy@...radead.org, vincent.chen@...ive.com, guoren@...nel.org,
 samitolvanen@...gle.com, songshuaishuai@...ylab.org, gerg@...nel.org,
 heiko@...ech.de, bhe@...hat.com, jeeheng.sia@...rfivetech.com,
 cyy@...self.name, maskray@...gle.com, ancientmodern4@...il.com,
 mathis.salmen@...sal.de, cuiyunhui@...edance.com, bgray@...ux.ibm.com,
 mpe@...erman.id.au, baruch@...s.co.il, alx@...nel.org,
 catalin.marinas@....com, revest@...omium.org, josh@...htriplett.org,
 shr@...kernel.io, deller@....de, omosnace@...hat.com, ojeda@...nel.org,
 jhubbard@...dia.com, Mike Rapoport <rppt@...nel.org>
Subject: Re: [PATCH v3 09/29] mm: abstract shadow stack vma behind
 `vma_is_shadow_stack`

On 04.04.24 01:34, Deepak Gupta wrote:
> VM_SHADOW_STACK (alias to VM_HIGH_ARCH_5) to encode shadow stack VMA.
> 
> This patch changes checks of VM_SHADOW_STACK flag in generic code to call
> to a function `vma_is_shadow_stack` which will return true if its a
> shadow stack vma and default stub (when support doesnt exist) returns false.
> 
> Signed-off-by: Deepak Gupta <debug@...osinc.com>
> Suggested-by: Mike Rapoport <rppt@...nel.org>
> ---
>   include/linux/mm.h | 13 ++++++++++++-
>   mm/gup.c           |  5 +++--
>   mm/internal.h      |  2 +-
>   3 files changed, 16 insertions(+), 4 deletions(-)
> 
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 64109f6c70f5..9952937be659 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -363,8 +363,19 @@ extern unsigned int kobjsize(const void *objp);
>   
>   #ifndef VM_SHADOW_STACK
>   # define VM_SHADOW_STACK	VM_NONE
> +
> +static inline bool vma_is_shadow_stack(vm_flags_t vm_flags)
> +{
> +	return false;
> +}
> +#else
> +static inline bool vma_is_shadow_stack(vm_flags_t vm_flags)
> +{
> +	return (vm_flags & VM_SHADOW_STACK);
> +}
>   #endif

You can simply do outside the ifdef

static inline bool vma_is_shadow_stack(vm_flags_t vm_flags)
{
	return !!(vm_flags & VM_SHADOW_STACK);
}

This will work even when VM_SHADOW_STACK is defined to be VM_NONE.

>   
> +

unrelated code change

>   #if defined(CONFIG_X86)
>   # define VM_PAT		VM_ARCH_1	/* PAT reserves whole VMA at once (x86) */
>   #elif defined(CONFIG_PPC)
> @@ -3473,7 +3484,7 @@ static inline unsigned long stack_guard_start_gap(struct vm_area_struct *vma)
>   		return stack_guard_gap;
>   
>   	/* See reasoning around the VM_SHADOW_STACK definition */
> -	if (vma->vm_flags & VM_SHADOW_STACK)
> +	if (vma->vm_flags && vma_is_shadow_stack(vma->vm_flags))

Pretty sure:

if (vma_is_shadow_stack(vma->vm_flags))

>   		return PAGE_SIZE;
>   
>   	return 0;
> diff --git a/mm/gup.c b/mm/gup.c
> index df83182ec72d..a7a02eb0a6b3 100644
> --- a/mm/gup.c
> +++ b/mm/gup.c
> @@ -1053,7 +1053,7 @@ static int check_vma_flags(struct vm_area_struct *vma, unsigned long gup_flags)
>   		    !writable_file_mapping_allowed(vma, gup_flags))
>   			return -EFAULT;
>   
> -		if (!(vm_flags & VM_WRITE) || (vm_flags & VM_SHADOW_STACK)) {
> +		if (!(vm_flags & VM_WRITE) || vma_is_shadow_stack(vm_flags)) {
>   			if (!(gup_flags & FOLL_FORCE))
>   				return -EFAULT;
>   			/* hugetlb does not support FOLL_FORCE|FOLL_WRITE. */
> @@ -1071,7 +1071,8 @@ static int check_vma_flags(struct vm_area_struct *vma, unsigned long gup_flags)
>   			if (!is_cow_mapping(vm_flags))
>   				return -EFAULT;
>   		}
> -	} else if (!(vm_flags & VM_READ)) {
> +	} else if (!(vm_flags & VM_READ) && !vma_is_shadow_stack(vm_flags)) {
> +	/* reads allowed if its shadow stack vma */
>   		if (!(gup_flags & FOLL_FORCE))
>   			return -EFAULT;
>   		/*

Unless I am missing something, this is not a simple cleanup. It should 
go into a separate patch with a clearly documented reason for that change.

-- 
Cheers,

David / dhildenb