linux-user should always enable AFP, otherwise our emulated binary
might crash once it tries to make use of additional floating-point
registers or instructions.
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: Alex Bennée <alex.bennee@linaro.org>
Fixes: db0504154e ("s390x/tcg: check for AFP-register, BFP and DFP data exceptions")
Reported-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: David Hildenbrand <david@redhat.com>
---
target/s390x/cpu.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
index 8ed4823d6e..18ba7f85a5 100644
--- a/target/s390x/cpu.c
+++ b/target/s390x/cpu.c
@@ -145,6 +145,11 @@ static void s390_cpu_full_reset(CPUState *s)
env->cregs[0] = CR0_RESET;
env->cregs[14] = CR14_RESET;
+#if defined(CONFIG_USER_ONLY)
+ /* user mode should always be allowed to use the full FPU */
+ env->cregs[0] |= CR0_AFP;
+#endif
+
/* architectured initial value for Breaking-Event-Address register */
env->gbea = 1;
--
2.17.1
On 2018-10-08 17:32, David Hildenbrand wrote:
> linux-user should always enable AFP, otherwise our emulated binary
> might crash once it tries to make use of additional floating-point
> registers or instructions.
>
> Cc: Peter Maydell <peter.maydell@linaro.org>
> Cc: Alex Bennée <alex.bennee@linaro.org>
> Fixes: db0504154e ("s390x/tcg: check for AFP-register, BFP and DFP data exceptions")
> Reported-by: Alex Bennée <alex.bennee@linaro.org>
> Signed-off-by: David Hildenbrand <david@redhat.com>
> ---
> target/s390x/cpu.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
> index 8ed4823d6e..18ba7f85a5 100644
> --- a/target/s390x/cpu.c
> +++ b/target/s390x/cpu.c
> @@ -145,6 +145,11 @@ static void s390_cpu_full_reset(CPUState *s)
> env->cregs[0] = CR0_RESET;
> env->cregs[14] = CR14_RESET;
>
> +#if defined(CONFIG_USER_ONLY)
> + /* user mode should always be allowed to use the full FPU */
> + env->cregs[0] |= CR0_AFP;
> +#endif
> +
> /* architectured initial value for Breaking-Event-Address register */
> env->gbea = 1;
>
>
Reviewed-by: Thomas Huth <thuth@redhat.com>
On Mon, 8 Oct 2018 17:32:04 +0200
David Hildenbrand <david@redhat.com> wrote:
> linux-user should always enable AFP, otherwise our emulated binary
> might crash once it tries to make use of additional floating-point
> registers or instructions.
>
> Cc: Peter Maydell <peter.maydell@linaro.org>
> Cc: Alex Bennée <alex.bennee@linaro.org>
> Fixes: db0504154e ("s390x/tcg: check for AFP-register, BFP and DFP data exceptions")
> Reported-by: Alex Bennée <alex.bennee@linaro.org>
> Signed-off-by: David Hildenbrand <david@redhat.com>
> ---
> target/s390x/cpu.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
> index 8ed4823d6e..18ba7f85a5 100644
> --- a/target/s390x/cpu.c
> +++ b/target/s390x/cpu.c
> @@ -145,6 +145,11 @@ static void s390_cpu_full_reset(CPUState *s)
> env->cregs[0] = CR0_RESET;
> env->cregs[14] = CR14_RESET;
>
> +#if defined(CONFIG_USER_ONLY)
> + /* user mode should always be allowed to use the full FPU */
> + env->cregs[0] |= CR0_AFP;
> +#endif
> +
> /* architectured initial value for Breaking-Event-Address register */
> env->gbea = 1;
>
I've reproduced the crash via make check-tcg, and this patch fixes it.
Thanks, applied.
On 10/8/18 8:32 AM, David Hildenbrand wrote:
> linux-user should always enable AFP, otherwise our emulated binary
> might crash once it tries to make use of additional floating-point
> registers or instructions.
>
> Cc: Peter Maydell <peter.maydell@linaro.org>
> Cc: Alex Bennée <alex.bennee@linaro.org>
> Fixes: db0504154e ("s390x/tcg: check for AFP-register, BFP and DFP data exceptions")
> Reported-by: Alex Bennée <alex.bennee@linaro.org>
> Signed-off-by: David Hildenbrand <david@redhat.com>
> ---
> target/s390x/cpu.c | 5 +++++
> 1 file changed, 5 insertions(+)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
David Hildenbrand <david@redhat.com> writes:
> linux-user should always enable AFP, otherwise our emulated binary
> might crash once it tries to make use of additional floating-point
> registers or instructions.
>
> Cc: Peter Maydell <peter.maydell@linaro.org>
> Cc: Alex Bennée <alex.bennee@linaro.org>
> Fixes: db0504154e ("s390x/tcg: check for AFP-register, BFP and DFP data exceptions")
> Reported-by: Alex Bennée <alex.bennee@linaro.org>
> Signed-off-by: David Hildenbrand <david@redhat.com>
> ---
> target/s390x/cpu.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
> index 8ed4823d6e..18ba7f85a5 100644
> --- a/target/s390x/cpu.c
> +++ b/target/s390x/cpu.c
> @@ -145,6 +145,11 @@ static void s390_cpu_full_reset(CPUState *s)
> env->cregs[0] = CR0_RESET;
> env->cregs[14] = CR14_RESET;
>
> +#if defined(CONFIG_USER_ONLY)
> + /* user mode should always be allowed to use the full FPU */
> + env->cregs[0] |= CR0_AFP;
> +#endif
> +
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
> /* architectured initial value for Breaking-Event-Address register */
> env->gbea = 1;
--
Alex Bennée
On Mon, 8 Oct 2018 17:32:04 +0200
David Hildenbrand <david@redhat.com> wrote:
> linux-user should always enable AFP, otherwise our emulated binary
> might crash once it tries to make use of additional floating-point
> registers or instructions.
>
> Cc: Peter Maydell <peter.maydell@linaro.org>
> Cc: Alex Bennée <alex.bennee@linaro.org>
> Fixes: db0504154e ("s390x/tcg: check for AFP-register, BFP and DFP data exceptions")
> Reported-by: Alex Bennée <alex.bennee@linaro.org>
> Signed-off-by: David Hildenbrand <david@redhat.com>
> ---
> target/s390x/cpu.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
> index 8ed4823d6e..18ba7f85a5 100644
> --- a/target/s390x/cpu.c
> +++ b/target/s390x/cpu.c
> @@ -145,6 +145,11 @@ static void s390_cpu_full_reset(CPUState *s)
> env->cregs[0] = CR0_RESET;
> env->cregs[14] = CR14_RESET;
>
> +#if defined(CONFIG_USER_ONLY)
> + /* user mode should always be allowed to use the full FPU */
> + env->cregs[0] |= CR0_AFP;
> +#endif
> +
> /* architectured initial value for Breaking-Event-Address register */
> env->gbea = 1;
>
In case this is applied directly:
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Tested-by: Cornelia Huck <cohuck@redhat.com>
On 8 October 2018 at 17:16, Cornelia Huck <cohuck@redhat.com> wrote:
> On Mon, 8 Oct 2018 17:32:04 +0200
> David Hildenbrand <david@redhat.com> wrote:
>
>> linux-user should always enable AFP, otherwise our emulated binary
>> might crash once it tries to make use of additional floating-point
>> registers or instructions.
>>
>> Cc: Peter Maydell <peter.maydell@linaro.org>
>> Cc: Alex Bennée <alex.bennee@linaro.org>
>> Fixes: db0504154e ("s390x/tcg: check for AFP-register, BFP and DFP data exceptions")
>> Reported-by: Alex Bennée <alex.bennee@linaro.org>
>> Signed-off-by: David Hildenbrand <david@redhat.com>
>> ---
>> target/s390x/cpu.c | 5 +++++
>> 1 file changed, 5 insertions(+)
>>
>> diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
>> index 8ed4823d6e..18ba7f85a5 100644
>> --- a/target/s390x/cpu.c
>> +++ b/target/s390x/cpu.c
>> @@ -145,6 +145,11 @@ static void s390_cpu_full_reset(CPUState *s)
>> env->cregs[0] = CR0_RESET;
>> env->cregs[14] = CR14_RESET;
>>
>> +#if defined(CONFIG_USER_ONLY)
>> + /* user mode should always be allowed to use the full FPU */
>> + env->cregs[0] |= CR0_AFP;
>> +#endif
>> +
>> /* architectured initial value for Breaking-Event-Address register */
>> env->gbea = 1;
>>
>
> In case this is applied directly:
>
> Reviewed-by: Cornelia Huck <cohuck@redhat.com>
> Tested-by: Cornelia Huck <cohuck@redhat.com>
Thanks; I've applied it to master, and also added 'make check-tcg'
to my set of things I run as part of a merge, which should catch
this kind of bug in future.
-- PMM
© 2016 - 2025 Red Hat, Inc.