Re: [PATCH 01/13] KVM: PPC: Ultravisor: generic ucall handler plpar_ucall()

[Paul Mackerras <paulus@ozlabs.org>]

On Tue, Jan 22, 2019 at 07:59:32AM -0800, Ram Pai wrote:
> plpar_ucall() function can now be used to make ucalls
> with varied number of in and out arguments. It handles
> 4 out arguments and a lot of in arguments.
> 
> copies the implementation of plpar_hcall().
> 
> Signed-off-by: Ram Pai <linuxram@us.ibm.com>
> ---
>  arch/powerpc/platforms/powernv/Makefile |  1 +
>  arch/powerpc/platforms/powernv/uvCall.S | 41 +++++++++++++++++++++++++++++++++

I don't think arch/powerpc/platforms/powernv is the right place for
this, since we will also need to do ucalls from a secure guest and its
kernel may not have CONFIG_PPC_POWERNV=y.  Also, the "plpar_" prefix
gives rise to a cognitive dissonance, because "plpar" is short for
pSeries logical partition, and we definitely are not in a pSeries
logical partition here.

> +_GLOBAL_TOC(plpar_ucall)
> +
> +	mfcr	r0
> +	stw	r0,8(r1)
> +	mr	r0,r3
> +
> +	std     r4,STK_PARAM(R4)(r1)     /* Save ret buffer */
> +
> +	mr	r3,r5
> +	mr	r4,r6
> +	mr	r5,r7
> +	mr	r6,r8
> +	mr	r7,r9
> +	mr	r8,r10
> +
> +	sc 2				/* invoke the ultravisor */
> +
> +	ld	r12,STK_PARAM(R4)(r1)
> +	std	r4,  0(r12)
> +	std	r5,  8(r12)
> +	std	r6, 16(r12)
> +	std	r7, 24(r12)
> +
> +	lwz	r0,8(r1)
> +	mtcrf	0xff,r0
> +
> +	blr				/* return r3 = status */

So we save/restore CR across the ucall, but not LR.  Do we really not
trust the ultravisor to preserve the non-volatile CR fields, or is it
just that you were copying plpar_hcall()?  Either way it deserves a
note in the patch description.

Paul.