2021-01-20 06:47:21

by Yejune Deng

Subject: [PATCH] x86/fpu/xstate: calculate the number by sizeof and offsetof

In fpstate_sanitize_xstate(), use memset and offsetof instead of '= 0',
and use sizeof instead of a constant.

Signed-off-by: Yejune Deng <[email protected]>
---
arch/x86/kernel/fpu/xstate.c | 10 +++-------
1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
index 5d8047441a0a..2e75630c86e9 100644
--- a/arch/x86/kernel/fpu/xstate.c
+++ b/arch/x86/kernel/fpu/xstate.c
@@ -161,20 +161,16 @@ void fpstate_sanitize_xstate(struct fpu *fpu)
* FP is in init state
*/
if (!(xfeatures & XFEATURE_MASK_FP)) {
+ memset(fx, 0, offsetof(struct fxregs_state, mxcsr));
+ memset(fx->st_space, 0, sizeof(fx->st_space));
fx->cwd = 0x37f;
- fx->swd = 0;
- fx->twd = 0;
- fx->fop = 0;
- fx->rip = 0;
- fx->rdp = 0;
- memset(&fx->st_space[0], 0, 128);
}

/*
* SSE is in init state
*/
if (!(xfeatures & XFEATURE_MASK_SSE))
- memset(&fx->xmm_space[0], 0, 256);
+ memset(fx->xmm_space, 0, sizeof(fx->xmm_space));

/*
* First two features are FPU and SSE, which above we handled
--
2.29.0

2021-01-20 06:54:30

by Dave Hansen

Subject: Re: [PATCH] x86/fpu/xstate: calculate the number by sizeof and offsetof

On 1/19/21 10:44 PM, Yejune Deng wrote:
> In fpstate_sanitize_xstate(), use memset and offsetof instead of '= 0',
> and use sizeof instead of a constant.

What's the benefit to doing this? Saving 4 lines of code?

Your suggestions are not obviously wrong at a glance, but they're also
not obviously right.

2021-01-20 12:06:13

by Borislav Petkov

Subject: Re: [PATCH] x86/fpu/xstate: calculate the number by sizeof and offsetof

On Wed, Jan 20, 2021 at 02:44:15PM +0800, Yejune Deng wrote:
> In fpstate_sanitize_xstate(), use memset and offsetof instead of '= 0',
> and use sizeof instead of a constant.
>
> Signed-off-by: Yejune Deng <[email protected]>
> ---
> arch/x86/kernel/fpu/xstate.c | 10 +++-------
> 1 file changed, 3 insertions(+), 7 deletions(-)
>
> diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
> index 5d8047441a0a..2e75630c86e9 100644
> --- a/arch/x86/kernel/fpu/xstate.c
> +++ b/arch/x86/kernel/fpu/xstate.c
> @@ -161,20 +161,16 @@ void fpstate_sanitize_xstate(struct fpu *fpu)
> * FP is in init state
> */
> if (!(xfeatures & XFEATURE_MASK_FP)) {
> + memset(fx, 0, offsetof(struct fxregs_state, mxcsr));

I'd prefer the explicit zeroing instead of having to look at
fxregs_state and where the offset of mxcsr is.

> + memset(fx->st_space, 0, sizeof(fx->st_space));

This is ok I guess.

> fx->cwd = 0x37f;
> - fx->swd = 0;
> - fx->twd = 0;
> - fx->fop = 0;
> - fx->rip = 0;
> - fx->rdp = 0;
> - memset(&fx->st_space[0], 0, 128);
> }
>
> /*
> * SSE is in init state
> */
> if (!(xfeatures & XFEATURE_MASK_SSE))
> - memset(&fx->xmm_space[0], 0, 256);
> + memset(fx->xmm_space, 0, sizeof(fx->xmm_space));

This too.

--
Regards/Gruss,
Boris.

https://people.kernel.org/tglx/notes-about-netiquette

2021-01-22 01:30:23

by Yejune Deng

Subject: Re: [PATCH] x86/fpu/xstate: calculate the number by sizeof and offsetof

Thank you for your suggestion！

On Wed, Jan 20, 2021 at 6:22 PM Borislav Petkov <[email protected]> wrote:
>
> On Wed, Jan 20, 2021 at 02:44:15PM +0800, Yejune Deng wrote:
> > In fpstate_sanitize_xstate(), use memset and offsetof instead of '= 0',
> > and use sizeof instead of a constant.
> >
> > Signed-off-by: Yejune Deng <[email protected]>
> > ---
> > arch/x86/kernel/fpu/xstate.c | 10 +++-------
> > 1 file changed, 3 insertions(+), 7 deletions(-)
> >
> > diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
> > index 5d8047441a0a..2e75630c86e9 100644
> > --- a/arch/x86/kernel/fpu/xstate.c
> > +++ b/arch/x86/kernel/fpu/xstate.c
> > @@ -161,20 +161,16 @@ void fpstate_sanitize_xstate(struct fpu *fpu)
> > * FP is in init state
> > */
> > if (!(xfeatures & XFEATURE_MASK_FP)) {
> > + memset(fx, 0, offsetof(struct fxregs_state, mxcsr));
>
> I'd prefer the explicit zeroing instead of having to look at
> fxregs_state and where the offset of mxcsr is.
>
> > + memset(fx->st_space, 0, sizeof(fx->st_space));
>
> This is ok I guess.
>
> > fx->cwd = 0x37f;
> > - fx->swd = 0;
> > - fx->twd = 0;
> > - fx->fop = 0;
> > - fx->rip = 0;
> > - fx->rdp = 0;
> > - memset(&fx->st_space[0], 0, 128);
> > }
> >
> > /*
> > * SSE is in init state
> > */
> > if (!(xfeatures & XFEATURE_MASK_SSE))
> > - memset(&fx->xmm_space[0], 0, 256);
> > + memset(fx->xmm_space, 0, sizeof(fx->xmm_space));
>
> This too.
>
> --
> Regards/Gruss,
> Boris.
>
> https://people.kernel.org/tglx/notes-about-netiquette