qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH v1 1/8] s390x/tcg: factor out and fix DATA excep


From: David Hildenbrand
Subject: Re: [Qemu-devel] [PATCH v1 1/8] s390x/tcg: factor out and fix DATA exception injection
Date: Wed, 22 Aug 2018 18:52:27 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1

On 22.08.2018 18:31, Richard Henderson wrote:
> On 08/22/2018 12:57 AM, David Hildenbrand wrote:
>> The DXC is to be stored in the low core, and only in the FPC in case AFP
>> is enabled in CR0.
>>
>> Signed-off-by: David Hildenbrand <address@hidden>
>> ---
>>  target/s390x/cpu.h         |  1 +
>>  target/s390x/excp_helper.c | 28 ++++++++++++++++++++++++++++
>>  target/s390x/fpu_helper.c  | 13 +++----------
>>  target/s390x/helper.h      |  1 +
>>  target/s390x/tcg_s390x.h   |  2 ++
>>  target/s390x/translate.c   | 19 +++++++++----------
>>  6 files changed, 44 insertions(+), 20 deletions(-)
>>
>> diff --git a/target/s390x/cpu.h b/target/s390x/cpu.h
>> index 6f8861e554..5e50c3a303 100644
>> --- a/target/s390x/cpu.h
>> +++ b/target/s390x/cpu.h
>> @@ -322,6 +322,7 @@ extern const struct VMStateDescription vmstate_s390_cpu;
>>  #define CR0_LOWPROT             0x0000000010000000ULL
>>  #define CR0_SECONDARY           0x0000000004000000ULL
>>  #define CR0_EDAT                0x0000000000800000ULL
>> +#define CR0_AFP                 0x0000000000040000ULL
>>  #define CR0_EMERGENCY_SIGNAL_SC 0x0000000000004000ULL
>>  #define CR0_EXTERNAL_CALL_SC    0x0000000000002000ULL
>>  #define CR0_CKC_SC              0x0000000000000800ULL
>> diff --git a/target/s390x/excp_helper.c b/target/s390x/excp_helper.c
>> index f0ce60cff2..f5cab94ec5 100644
>> --- a/target/s390x/excp_helper.c
>> +++ b/target/s390x/excp_helper.c
>> @@ -21,11 +21,13 @@
>>  #include "qemu/osdep.h"
>>  #include "cpu.h"
>>  #include "internal.h"
>> +#include "exec/helper-proto.h"
>>  #include "qemu/timer.h"
>>  #include "exec/exec-all.h"
>>  #include "exec/cpu_ldst.h"
>>  #include "hw/s390x/ioinst.h"
>>  #include "exec/address-spaces.h"
>> +#include "tcg_s390x.h"
>>  #ifndef CONFIG_USER_ONLY
>>  #include "sysemu/sysemu.h"
>>  #include "hw/s390x/s390_flic.h"
>> @@ -48,6 +50,32 @@
>>      do { } while (0)
>>  #endif
>>  
>> +void QEMU_NORETURN tcg_s390_data_exception(CPUS390XState *env, uint32_t dxc,
>> +                                           uintptr_t ra)
>> +{
>> +    CPUState *cs = CPU(s390_env_get_cpu(env));
>> +
>> +    g_assert(!(dxc & ~0xff));
>> +#if !defined(CONFIG_USER_ONLY)
>> +    /* Store the DXC into the lowcore */
>> +    stw_phys(cs->as, env->psa + offsetof(LowCore, data_exc_code), dxc);
>> +#endif
>> +
>> +    /* Store the DXC into the FPC if AFP is enabled */
>> +    if (env->cregs[0] & CR0_AFP) {
>> +        env->fpc = (env->fpc & ~0xff00) | (dxc << 8);
>> +    }
>> +    s390_program_interrupt(env, PGM_DATA, ILEN_AUTO, ra);
> 
> We should probably export tcg_s390_program_interrupt, so ...
> 
>> +
>> +    /* the following is not necessary, but allows us to use noreturn */
>> +    cpu_loop_exit_restore(cs, ra);
> 
> ... you don't have to do this.
> But even if not, g_assert_not_reached() is better here.

If that also works, fine.

> 
> I'll note that s390_program_interrupt is buggy for tcg, in that it logs
> env->psw.addr before cpu_loop_restore, and therefore psw.addr is essentially
> random data.

Indeed, for that it works fairly reliably.

I'll send a fix with the next version.

Thanks!

> 
> 
> r~
> 


-- 

Thanks,

David / dhildenb



reply via email to

[Prev in Thread] Current Thread [Next in Thread]