Re: [Qemu-devel] [for-2.6 PATCH 1/3] target-i386: Define structs for layout of xsave area
On 11/30/2015 03:18 AM, Paolo Bonzini wrote: Because this is always little endian, I would write it as uint8_t[16][16]. Maybe. That isn't altogether handy for TCG, since we'll be wanting to bswap these buffers (probably in uint64_t chunks). r~ -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [Qemu-devel] [for-2.6 PATCH 1/3] target-i386: Define structs for layout of xsave area
On Tue, Dec 01, 2015 at 09:09:47AM -0800, Richard Henderson wrote: > On 11/30/2015 03:18 AM, Paolo Bonzini wrote: > >Because this is always little endian, I would write it as uint8_t[16][16]. > > Maybe. That isn't altogether handy for TCG, since we'll be wanting to bswap > these buffers (probably in uint64_t chunks). X86XSaveArea will be used only when loading/saving state using xsave, not for executing regular instructions. In X86CPU, the data is already stored as XMMReg unions (the one with the XMM_[BWDQ] helpers). -- Eduardo -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [Qemu-devel] [for-2.6 PATCH 1/3] target-i386: Define structs for layout of xsave area
On 12/01/2015 09:15 AM, Eduardo Habkost wrote: On Tue, Dec 01, 2015 at 09:09:47AM -0800, Richard Henderson wrote: On 11/30/2015 03:18 AM, Paolo Bonzini wrote: Because this is always little endian, I would write it as uint8_t[16][16]. Maybe. That isn't altogether handy for TCG, since we'll be wanting to bswap these buffers (probably in uint64_t chunks). X86XSaveArea will be used only when loading/saving state using xsave, not for executing regular instructions. ... like the regular instruction xsave? https://patchwork.ozlabs.org/patch/493318/ In X86CPU, the data is already stored as XMMReg unions (the one with the XMM_[BWDQ] helpers). Of course. But those unions are arranged to be in big-endian format on big-endian hosts. So we need to swap the data back to little-endian format for storage into guest memory. r~ -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [Qemu-devel] [for-2.6 PATCH 1/3] target-i386: Define structs for layout of xsave area
On 01/12/2015 18:20, Richard Henderson wrote: >> >> X86XSaveArea will be used only when loading/saving state using >> xsave, not for executing regular instructions. > > ... like the regular instruction xsave? > > https://patchwork.ozlabs.org/patch/493318/ Right, but that's a helper anyway. >> In X86CPU, the >> data is already stored as XMMReg unions (the one with the >> XMM_[BWDQ] helpers). > > Of course. But those unions are arranged to be in big-endian format on > big-endian hosts. So we need to swap the data back to little-endian > format for storage into guest memory. Yes, you can use byte moves with XMM_B (more obvious), or stq_le_p with XMM_Q (faster I guess---though the compiler might optimize the former on little-endian hosts). Either works with an uint8_t[] destination. Paolo -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [Qemu-devel] [for-2.6 PATCH 1/3] target-i386: Define structs for layout of xsave area
On Tue, Dec 01, 2015 at 06:27:17PM +0100, Paolo Bonzini wrote:
> On 01/12/2015 18:20, Richard Henderson wrote:
> >>
> >> X86XSaveArea will be used only when loading/saving state using
> >> xsave, not for executing regular instructions.
> >
> > ... like the regular instruction xsave?
> >
> > https://patchwork.ozlabs.org/patch/493318/
>
> Right, but that's a helper anyway.
>
> >> In X86CPU, the
> >> data is already stored as XMMReg unions (the one with the
> >> XMM_[BWDQ] helpers).
> >
> > Of course. But those unions are arranged to be in big-endian format on
> > big-endian hosts. So we need to swap the data back to little-endian
> > format for storage into guest memory.
>
> Yes, you can use byte moves with XMM_B (more obvious), or stq_le_p with
> XMM_Q (faster I guess---though the compiler might optimize the former on
> little-endian hosts). Either works with an uint8_t[] destination.
stq_le_p() (more exactly, stq_p()) is exactly what is already
done by kvm_{get,put}_xsave(), using uint8_t pointers.
BTW, if we are going to implement xsave in TCG, the
X86CPU<->xsave translation logic in kvm_{get,put}_xsave() could
be moved to generic code and reused by TCG instead of being
reimplemented.
--
Eduardo
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [Qemu-devel] [for-2.6 PATCH 1/3] target-i386: Define structs for layout of xsave area
On 12/01/2015 10:34 AM, Eduardo Habkost wrote:
BTW, if we are going to implement xsave in TCG, the
X86CPU<->xsave translation logic in kvm_{get,put}_xsave() could
be moved to generic code and reused by TCG instead of being
reimplemented.
That's not trivial.
In particular, stq_p isn't what the tcg helper needs to use, but rather
cpu_stq_data_ra. Given the differing parameters, we'd have to resort to some
sort of macro-ization. It's probably easiest to simply keep the two
implementations separate.
r~
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
