From: KONRAD Frederic <frederic.konrad@adacore.com>
Since the commit af7a06bac7d3abb2da48ef3277d2a415772d2ae8:
`casa [..](10), .., ..` (and probably others alternate space instructions)
triggers a data access exception when the MMU is disabled.
When we enter get_asi(...) dc->mem_idx is set to MMU_PHYS_IDX when the MMU
is disabled. Just keep mem_idx unchanged in this case so we passthrough the
MMU when it is disabled.
Signed-off-by: KONRAD Frederic <frederic.konrad@adacore.com>
---
Notes:
Changes RFC -> V1:
* emit the instruction with MMU_PHYS_IDX instead of checking that the MMU
is enabled in get_physical_address(..)
target/sparc/translate.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/target/sparc/translate.c b/target/sparc/translate.c
index 71e0853..5aa367a 100644
--- a/target/sparc/translate.c
+++ b/target/sparc/translate.c
@@ -2093,6 +2093,11 @@ static DisasASI get_asi(DisasContext *dc, int insn, TCGMemOp memop)
type = GET_ASI_BFILL;
break;
}
+
+ /* MMU_PHYS_IDX is used when the MMU is disabled to passthrough the
+ * permissions check in get_physical_address(..).
+ */
+ mem_idx = (dc->mem_idx == MMU_PHYS_IDX) ? MMU_PHYS_IDX : mem_idx;
} else {
gen_exception(dc, TT_PRIV_INSN);
type = GET_ASI_EXCP;
--
1.8.3.1
On 03/02/2018 01:59 AM, KONRAD Frederic wrote: > From: KONRAD Frederic <frederic.konrad@adacore.com> > > Since the commit af7a06bac7d3abb2da48ef3277d2a415772d2ae8: > `casa [..](10), .., ..` (and probably others alternate space instructions) > triggers a data access exception when the MMU is disabled. > > When we enter get_asi(...) dc->mem_idx is set to MMU_PHYS_IDX when the MMU > is disabled. Just keep mem_idx unchanged in this case so we passthrough the > MMU when it is disabled. > > Signed-off-by: KONRAD Frederic <frederic.konrad@adacore.com> > --- > > Notes: > Changes RFC -> V1: > * emit the instruction with MMU_PHYS_IDX instead of checking that the MMU > is enabled in get_physical_address(..) > > target/sparc/translate.c | 5 +++++ > 1 file changed, 5 insertions(+) Reviewed-by: Richard Henderson <richard.henderson@linaro.org> r~
On 03/02/2018 08:04 PM, Richard Henderson wrote: > On 03/02/2018 01:59 AM, KONRAD Frederic wrote: >> From: KONRAD Frederic <frederic.konrad@adacore.com> >> >> Since the commit af7a06bac7d3abb2da48ef3277d2a415772d2ae8: >> `casa [..](10), .., ..` (and probably others alternate space instructions) >> triggers a data access exception when the MMU is disabled. >> >> When we enter get_asi(...) dc->mem_idx is set to MMU_PHYS_IDX when the MMU >> is disabled. Just keep mem_idx unchanged in this case so we passthrough the >> MMU when it is disabled. >> >> Signed-off-by: KONRAD Frederic <frederic.konrad@adacore.com> >> --- >> >> Notes: >> Changes RFC -> V1: >> * emit the instruction with MMU_PHYS_IDX instead of checking that the MMU >> is enabled in get_physical_address(..) >> >> target/sparc/translate.c | 5 +++++ >> 1 file changed, 5 insertions(+) > > Reviewed-by: Richard Henderson <richard.henderson@linaro.org> > > r~ > Thanks for the review Richard. BTW I think it makes sense to have it in stable release.. So CC'ing qemu-stable. Regards, Fred
On 02/03/18 19:04, Richard Henderson wrote: > On 03/02/2018 01:59 AM, KONRAD Frederic wrote: >> From: KONRAD Frederic <frederic.konrad@adacore.com> >> >> Since the commit af7a06bac7d3abb2da48ef3277d2a415772d2ae8: >> `casa [..](10), .., ..` (and probably others alternate space instructions) >> triggers a data access exception when the MMU is disabled. >> >> When we enter get_asi(...) dc->mem_idx is set to MMU_PHYS_IDX when the MMU >> is disabled. Just keep mem_idx unchanged in this case so we passthrough the >> MMU when it is disabled. >> >> Signed-off-by: KONRAD Frederic <frederic.konrad@adacore.com> >> --- >> >> Notes: >> Changes RFC -> V1: >> * emit the instruction with MMU_PHYS_IDX instead of checking that the MMU >> is enabled in get_physical_address(..) >> >> target/sparc/translate.c | 5 +++++ >> 1 file changed, 5 insertions(+) > > Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Thanks for reviewing the patch Richard. Do you want me to take this one via my qemu-sparc branch? ATB, Mark.
On 03/05/2018 09:56 PM, Mark Cave-Ayland wrote: > Thanks for reviewing the patch Richard. Do you want me to take this one via my > qemu-sparc branch? Yes, please. r~
On 02/03/18 09:59, KONRAD Frederic wrote:
> From: KONRAD Frederic <frederic.konrad@adacore.com>
>
> Since the commit af7a06bac7d3abb2da48ef3277d2a415772d2ae8:
> `casa [..](10), .., ..` (and probably others alternate space instructions)
> triggers a data access exception when the MMU is disabled.
>
> When we enter get_asi(...) dc->mem_idx is set to MMU_PHYS_IDX when the MMU
> is disabled. Just keep mem_idx unchanged in this case so we passthrough the
> MMU when it is disabled.
>
> Signed-off-by: KONRAD Frederic <frederic.konrad@adacore.com>
> ---
>
> Notes:
> Changes RFC -> V1:
> * emit the instruction with MMU_PHYS_IDX instead of checking that the MMU
> is enabled in get_physical_address(..)
>
> target/sparc/translate.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/target/sparc/translate.c b/target/sparc/translate.c
> index 71e0853..5aa367a 100644
> --- a/target/sparc/translate.c
> +++ b/target/sparc/translate.c
> @@ -2093,6 +2093,11 @@ static DisasASI get_asi(DisasContext *dc, int insn, TCGMemOp memop)
> type = GET_ASI_BFILL;
> break;
> }
> +
> + /* MMU_PHYS_IDX is used when the MMU is disabled to passthrough the
> + * permissions check in get_physical_address(..).
> + */
> + mem_idx = (dc->mem_idx == MMU_PHYS_IDX) ? MMU_PHYS_IDX : mem_idx;
> } else {
> gen_exception(dc, TT_PRIV_INSN);
> type = GET_ASI_EXCP;
>
Thanks Frederic, I've applied this to my qemu-sparc branch.
ATB,
Mark.
© 2016 - 2026 Red Hat, Inc.