[PATCH] target/arm: don't bother with id_aa64pfr0_read for USER_ONLY

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

[PATCH] target/arm: don't bother with id_aa64pfr0_read for USER_ONLY

Alex Bennée-2
For system emulation we need to check the state of the GIC before we
report the value. However this isn't relevant to exporting of the
value to linux-user and indeed breaks the exported value as set by
modify_arm_cp_regs.

[AJB: the other option would be just to set reset value anyway and not
ifdef out the readfn as the register will become const anyway]

Signed-off-by: Alex Bennée <[hidden email]>
---
 target/arm/helper.c | 16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/target/arm/helper.c b/target/arm/helper.c
index f78dd3b5fe3..489c31504a6 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -5867,6 +5867,7 @@ static uint64_t id_pfr1_read(CPUARMState *env, const ARMCPRegInfo *ri)
     return pfr1;
 }
 
+#ifndef CONFIG_USER_ONLY
 static uint64_t id_aa64pfr0_read(CPUARMState *env, const ARMCPRegInfo *ri)
 {
     ARMCPU *cpu = env_archcpu(env);
@@ -5877,6 +5878,7 @@ static uint64_t id_aa64pfr0_read(CPUARMState *env, const ARMCPRegInfo *ri)
     }
     return pfr0;
 }
+#endif
 
 /* Shared logic between LORID and the rest of the LOR* registers.
  * Secure state has already been delt with.
@@ -6297,16 +6299,22 @@ void register_cp_regs_for_features(ARMCPU *cpu)
          * define new registers here.
          */
         ARMCPRegInfo v8_idregs[] = {
-            /* ID_AA64PFR0_EL1 is not a plain ARM_CP_CONST because we don't
-             * know the right value for the GIC field until after we
-             * define these regs.
+            /*
+             * ID_AA64PFR0_EL1 is not a plain ARM_CP_CONST in system
+             * emulation because we don't know the right value for the
+             * GIC field until after we define these regs.
              */
             { .name = "ID_AA64PFR0_EL1", .state = ARM_CP_STATE_AA64,
               .opc0 = 3, .opc1 = 0, .crn = 0, .crm = 4, .opc2 = 0,
               .access = PL1_R, .type = ARM_CP_NO_RAW,
               .accessfn = access_aa64_tid3,
+#ifdef CONFIG_USER_ONLY
+              .resetvalue = cpu->isar.id_aa64pfr0
+#else
               .readfn = id_aa64pfr0_read,
-              .writefn = arm_cp_write_ignore },
+              .writefn = arm_cp_write_ignore
+#endif
+            },
             { .name = "ID_AA64PFR1_EL1", .state = ARM_CP_STATE_AA64,
               .opc0 = 3, .opc1 = 0, .crn = 0, .crm = 4, .opc2 = 1,
               .access = PL1_R, .type = ARM_CP_CONST,
--
2.20.1


Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] target/arm: don't bother with id_aa64pfr0_read for USER_ONLY

Peter Maydell-5
On Fri, 6 Dec 2019 at 12:22, Alex Bennée <[hidden email]> wrote:
>
> For system emulation we need to check the state of the GIC before we
> report the value. However this isn't relevant to exporting of the
> value to linux-user and indeed breaks the exported value as set by
> modify_arm_cp_regs.
>
> [AJB: the other option would be just to set reset value anyway and not
> ifdef out the readfn as the register will become const anyway]

If you want it to be const it would be clearer to define it
with ARM_CP_CONST... I'm not sure what an ARM_CP_NO_RAW without
a readfn or a fieldoffset will do on reads.

thanks
-- PMM

Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] target/arm: don't bother with id_aa64pfr0_read for USER_ONLY

Richard Henderson-3
On 12/6/19 7:29 AM, Peter Maydell wrote:

> On Fri, 6 Dec 2019 at 12:22, Alex Bennée <[hidden email]> wrote:
>>
>> For system emulation we need to check the state of the GIC before we
>> report the value. However this isn't relevant to exporting of the
>> value to linux-user and indeed breaks the exported value as set by
>> modify_arm_cp_regs.
>>
>> [AJB: the other option would be just to set reset value anyway and not
>> ifdef out the readfn as the register will become const anyway]
>
> If you want it to be const it would be clearer to define it
> with ARM_CP_CONST... I'm not sure what an ARM_CP_NO_RAW without
> a readfn or a fieldoffset will do on reads.

Yep, and the accessfn should be ifdefed away as well.


r~

Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] target/arm: don't bother with id_aa64pfr0_read for USER_ONLY

Alex Bennée-2
In reply to this post by Peter Maydell-5

Peter Maydell <[hidden email]> writes:

> On Fri, 6 Dec 2019 at 12:22, Alex Bennée <[hidden email]> wrote:
>>
>> For system emulation we need to check the state of the GIC before we
>> report the value. However this isn't relevant to exporting of the
>> value to linux-user and indeed breaks the exported value as set by
>> modify_arm_cp_regs.
>>
>> [AJB: the other option would be just to set reset value anyway and not
>> ifdef out the readfn as the register will become const anyway]
>
> If you want it to be const it would be clearer to define it
> with ARM_CP_CONST... I'm not sure what an ARM_CP_NO_RAW without
> a readfn or a fieldoffset will do on reads.

Well the modify_arm_cp_regs ensures it is ARM_CP_CONST when it changes
the definition. It's just ensuring the reset value is set so it can be
masked/fixed.

However the ifdef approach does reduce the amount of unused stuff in the
linux-user build.

>
> thanks
> -- PMM


--
Alex Bennée