On Fri, Jun 5, 2015 at 10:42 AM, Edgar E. Iglesias
<edgar.igles...@xilinx.com> wrote:
> On Thu, Jun 04, 2015 at 11:23:19AM +1000, Alistair Francis wrote:
>> Originally the version_mask PVR bits were manually set for each
>> machine. This is a hassle and difficult to read, instead set them
>> based on the CPU properties.
>>
>> Signed-off-by: Alistair Francis <alistair.fran...@xilinx.com>
>> ---
>> hw/microblaze/petalogix_ml605_mmu.c | 3 ++-
>> target-microblaze/cpu-qom.h | 1 +
>> target-microblaze/cpu.c | 4 +++-
>> 3 files changed, 6 insertions(+), 2 deletions(-)
>>
>> diff --git a/hw/microblaze/petalogix_ml605_mmu.c
>> b/hw/microblaze/petalogix_ml605_mmu.c
>> index 5f341c4..f52654c 100644
>> --- a/hw/microblaze/petalogix_ml605_mmu.c
>> +++ b/hw/microblaze/petalogix_ml605_mmu.c
>> @@ -70,7 +70,7 @@ static void machine_cpu_reset(MicroBlazeCPU *cpu)
>>
>> env->pvr.regs[10] = 0x0e000000; /* virtex 6 */
>> /* setup pvr to match kernel setting */
>> - env->pvr.regs[0] = (env->pvr.regs[0] & ~PVR0_VERSION_MASK) | (0x14 <<
>> 8);
>> + env->pvr.regs[0] |= (0x14 << 8);
>> env->pvr.regs[4] = 0xc56b8000;
>> env->pvr.regs[5] = 0xc56be000;
>> }
>> @@ -99,6 +99,7 @@ petalogix_ml605_init(MachineState *machine)
>> object_property_set_bool(OBJECT(cpu), true, "dcache-writeback",
>> &error_abort);
>> object_property_set_bool(OBJECT(cpu), true, "endi", &error_abort);
>> + object_property_set_bool(OBJECT(cpu), true, "version_mask",
>> &error_abort);
>> object_property_set_bool(OBJECT(cpu), true, "realized", &error_abort);
>>
>> /* Attach emulated BRAM through the LMB. */
>> diff --git a/target-microblaze/cpu-qom.h b/target-microblaze/cpu-qom.h
>> index 85b8f75..b6c6374 100644
>> --- a/target-microblaze/cpu-qom.h
>> +++ b/target-microblaze/cpu-qom.h
>> @@ -67,6 +67,7 @@ typedef struct MicroBlazeCPU {
>> bool usemmu;
>> bool dcache_writeback;
>> bool endi;
>> + bool version_mask;
>> } cfg;
>>
>> CPUMBState env;
>> diff --git a/target-microblaze/cpu.c b/target-microblaze/cpu.c
>> index f40df43..849c737 100644
>> --- a/target-microblaze/cpu.c
>> +++ b/target-microblaze/cpu.c
>> @@ -115,7 +115,8 @@ static void mb_cpu_realizefn(DeviceState *dev, Error
>> **errp)
>> env->pvr.regs[0] |= (cpu->cfg.stackprot ? PVR0_SPROT_MASK : 0) |
>> (cpu->cfg.usefpu ? PVR0_USE_FPU_MASK : 0) |
>> (cpu->cfg.usemmu ? PVR0_USE_MMU_MASK : 0) |
>> - (cpu->cfg.endi ? PVR0_ENDI_MASK : 0);
>> + (cpu->cfg.endi ? PVR0_ENDI_MASK : 0) |
>> + (cpu->cfg.endi ? ~PVR0_VERSION_MASK : 0);
>
> This looks wrong...
>
> My guess is that this should be:
> cpu->cfg.version_mask ? PVR0_VERSION_MASK : 0
So looking at the spec more the version mask should be a two byte
number that indicated the version. I can't see it in the DTS, so I
think it should be a settable number, not a bool.
I'm not sure why it zeroes it out though,
Thanks,
Alistair
>
> and that maybe the ml605_mmu board is trying to disable the version mask,
> e.g setting it to false.
>
> Can you double check that it matches with dts and specs?
>
>
>>
>> env->pvr.regs[2] |= (cpu->cfg.usefpu ? PVR2_USE_FPU_MASK : 0) |
>> (cpu->cfg.usefpu > 1 ? PVR2_USE_FPU2_MASK : 0);
>> @@ -176,6 +177,7 @@ static Property mb_properties[] = {
>> DEFINE_PROP_BOOL("dcache-writeback", MicroBlazeCPU,
>> cfg.dcache_writeback,
>> false),
>> DEFINE_PROP_BOOL("endi", MicroBlazeCPU, cfg.endi, false),
>> + DEFINE_PROP_BOOL("version-mask", MicroBlazeCPU, cfg.version_mask,
>> false),
>> DEFINE_PROP_END_OF_LIST(),
>> };
>>
>> --
>> 1.7.1
>>
>
