[SLOF] [PATCH v2] broken_sc1: check for H_PRIVILEGE

Michael Ellerman michael at ellerman.id.au
Tue Apr 3 23:21:27 AEST 2018


Nikunj A Dadhania <nikunj at linux.vnet.ibm.com> writes:

> Recently, found that when DAWR was disabled by linux kernel, the hcall started
> returning H_UNSUPPORTED, and VM did not boot up as broken_sc1 patched up SC
> calls falsely.
>
> Instead of checking for various returns, check if its not in privilege mode and
> patch sc1 in that case.
>
> CC: Michael Ellerman <michael at ellerman.id.au>
> CC: Aneesh Kumar K.V <aneesh.kumar at linux.vnet.ibm.com>
> Signed-off-by: Nikunj A Dadhania <nikunj at linux.vnet.ibm.com>

Thanks for doing this.

We may never update the kernel to return H_UNSUPPORTED, but at least we
won't get bitten by this horrible^H^H^Hawesome PR KVM hack again.

cheers

> diff --git a/lib/libhvcall/brokensc1.c b/lib/libhvcall/brokensc1.c
> index e6387e0..d97ae77 100644
> --- a/lib/libhvcall/brokensc1.c
> +++ b/lib/libhvcall/brokensc1.c
> @@ -47,13 +47,13 @@ static int check_broken_sc1(void)
>  	 * supervisor mode.
>  	 */
>  	r = hcall(INS_SC1, H_SET_DABR, 0);
> -	if (r == H_SUCCESS || r == H_HARDWARE) {
> -		/* All is fine */
> -		return 0;
> +	if (r == H_PRIVILEGE) {
> +		/* We found a broken sc1 host! */
> +		return 1;
>  	}
>  
> -	/* We found a broken sc1 host! */
> -	return 1;
> +	/* All is fine */
> +	return 0;
>  }
>  
>  int patch_broken_sc1(void *start, void *end, uint32_t *test_ins)
> diff --git a/lib/libhvcall/libhvcall.h b/lib/libhvcall/libhvcall.h
> index 3fa4398..caa4d6d 100644
> --- a/lib/libhvcall/libhvcall.h
> +++ b/lib/libhvcall/libhvcall.h
> @@ -3,6 +3,7 @@
>  
>  #define H_SUCCESS		0
>  #define H_HARDWARE		-1
> +#define H_PRIVILEGE		-3   /* Caller not privileged */
>  
>  #define H_GET_TCE		0x1C
>  #define H_PUT_TCE		0x20
> -- 
> 2.14.3


More information about the SLOF mailing list