linux-kernel - Re: [PATCH 1/2] x86/fpu: Extend kernel_fpu_begin

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 for Android: free password hash cracker in your pocket

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

Date: Tue, 30 Apr 2024 16:42:25 -0700
From: Ashok Raj <ashok.raj@...el.com>
To: "Chang S. Bae" <chang.seok.bae@...el.com>
CC: <linux-kernel@...r.kernel.org>, <x86@...nel.org>,
	<platform-driver-x86@...r.kernel.org>, <tglx@...utronix.de>,
	<mingo@...hat.com>, <bp@...en8.de>, <dave.hansen@...ux.intel.com>,
	<hdegoede@...hat.com>, <ilpo.jarvinen@...ux.intel.com>,
	<tony.luck@...el.com>, <jithu.joseph@...el.com>, Ashok Raj
	<ashok.raj@...el.com>
Subject: Re: [PATCH 1/2] x86/fpu: Extend kernel_fpu_begin_mask() to
 initialize AMX state

On Tue, Apr 30, 2024 at 02:25:07PM -0700, Bae, Chang Seok wrote:
> The In-Field Scan [1] test does not start its operation when the CPU is
> unprepared. If the AMX state is uninitialized, the scan test will
> immediately terminate. Therefore, a proper initialization of the AMX
> state is necessary to run the test.

Instead of using new terminology like "unprepared", maybe use the
term "non-initialized state" as in the SDM for clarity? 

	Maintaining AMX state in a non-initialized state will prevent In-field
	Scan test [1] to abort without making progress.

> 
> Although fpu_idle_fpregs() initializes the state, its usage should be
> limited to specialized cases, primarily before entering the sleep state.
> The restore_fpregs_from_fpstate() function offers a suitable mechanism
> for initializing fpstate in general, which remains within the core code.
> 
> Extend kernel_fpu_begin_mask() to include AMX state initialization,
> providing the in-field scan driver code access to the appropriate
> initialization method while maintaining compliance with established FPU
> API semantics.

Maybe simplify the above paragraph with

Extend kernel_fpu_begin_mask() to allow In-field scan driver to initialize
AMX state before running tests.

> 
> [1] https://docs.kernel.org/arch/x86/ifs.html
> Signed-off-by: Chang S. Bae <chang.seok.bae@...el.com>
> ---
> The necessity for AMX initialization is clarified in the Intel Software
> Development Manual as of March 2024, particularly in Section 18.2
> RECOMMENDATIONS FOR SYSTEM SOFTWARE of Vol. 1.
> 
> Side note: restore_fpregs_from_fpstate() also sets the x87 state to a
> fixed value. However, this only applies to AMD CPUs with the FXSAVE_LEAK
> quirk.
> ---
>  arch/x86/include/asm/fpu/api.h | 1 +
>  arch/x86/kernel/fpu/core.c     | 3 +++
>  2 files changed, 4 insertions(+)
> 
> diff --git a/arch/x86/include/asm/fpu/api.h b/arch/x86/include/asm/fpu/api.h
> index a2be3aefff9f..67887fc45c24 100644
> --- a/arch/x86/include/asm/fpu/api.h
> +++ b/arch/x86/include/asm/fpu/api.h
> @@ -25,6 +25,7 @@
>  /* Kernel FPU states to initialize in kernel_fpu_begin_mask() */
>  #define KFPU_387	_BITUL(0)	/* 387 state will be initialized */
>  #define KFPU_MXCSR	_BITUL(1)	/* MXCSR will be initialized */
> +#define KFPU_AMX	_BITUL(2)	/* AMX will be initialized */
>  
>  extern void kernel_fpu_begin_mask(unsigned int kfpu_mask);
>  extern void kernel_fpu_end(void);
> diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c
> index 520deb411a70..0c0235b4a901 100644
> --- a/arch/x86/kernel/fpu/core.c
> +++ b/arch/x86/kernel/fpu/core.c
> @@ -440,6 +440,9 @@ void kernel_fpu_begin_mask(unsigned int kfpu_mask)
>  
>  	if (unlikely(kfpu_mask & KFPU_387) && boot_cpu_has(X86_FEATURE_FPU))
>  		asm volatile ("fninit");
> +
> +	if (unlikely(kfpu_mask & KFPU_AMX) && boot_cpu_has(X86_FEATURE_AMX_TILE))
> +		restore_fpregs_from_fpstate(&init_fpstate, XFEATURE_MASK_XTILE);
>  }
>  EXPORT_SYMBOL_GPL(kernel_fpu_begin_mask);
>  
> -- 
> 2.34.1
> 

-- 
Cheers,
Ashok