qemu-devel
[Top][All Lists]
Advanced

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

Re: [Qemu-devel] [PATCH v1 for-2.12 07/15] s390x: handle exceptions duri


From: David Hildenbrand
Subject: Re: [Qemu-devel] [PATCH v1 for-2.12 07/15] s390x: handle exceptions during s390_cpu_virt_mem_rw() correctly (TCG)
Date: Wed, 29 Nov 2017 16:54:30 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.4.0

> This looks a bit odd if you don't realize that kvm already handled the
> exception. But I don't really have a better idea.
> 
>>          }
>>          return;
>>      }
> 
>> @@ -645,9 +657,12 @@ void ioinst_handle_chsc(S390CPU *cpu, uint32_t ipb, 
>> uintptr_t ra)
>>      if (!s390_cpu_virt_mem_write(cpu, addr + len, reg, res,
>>                                   be16_to_cpu(res->len))) {
>>          setcc(cpu, 0);    /* Command execution complete */
>> +    } else {
>> +        s390_cpu_virt_mem_handle_exc(cpu, ra);
>>      }
>>  }
>>  
>> +
> 
> unrelated whitespace

ack, will drop.

> 
>>  #define SCHM_REG1_RES(_reg) (_reg & 0x000000000ffffffc)
>>  #define SCHM_REG1_MBK(_reg) ((_reg & 0x00000000f0000000) >> 28)
>>  #define SCHM_REG1_UPD(_reg) ((_reg & 0x0000000000000002) >> 1)
>> diff --git a/target/s390x/mmu_helper.c b/target/s390x/mmu_helper.c
>> index 31e3f3f415..39da9aeef4 100644
>> --- a/target/s390x/mmu_helper.c
>> +++ b/target/s390x/mmu_helper.c
>> @@ -22,6 +22,7 @@
>>  #include "internal.h"
>>  #include "kvm_s390x.h"
>>  #include "sysemu/kvm.h"
>> +#include "exec/exec-all.h"
>>  #include "trace.h"
>>  #include "hw/s390x/storage-keys.h"
>>  
>> @@ -458,7 +459,7 @@ static int translate_pages(S390CPU *cpu, vaddr addr, int 
>> nr_pages,
>>          }
>>          if (!address_space_access_valid(&address_space_memory, pages[i],
>>                                          TARGET_PAGE_SIZE, is_write)) {
>> -            program_interrupt(env, PGM_ADDRESSING, ILEN_AUTO);
>> +            trigger_pgm_exception(env, PGM_ADDRESSING, ILEN_AUTO);
> 
> Why did you change this?

"For TCG, there was one case where a cpu loop exit was triggered. Fix
that up."

Wasn't worth a separate patch :)

> 
>>              return -EFAULT;
>>          }
>>          addr += TARGET_PAGE_SIZE;
>> @@ -478,6 +479,9 @@ static int translate_pages(S390CPU *cpu, vaddr addr, int 
>> nr_pages,
>>   *
>>   * Copy from/to guest memory using logical addresses. Note that we inject a
>>   * program interrupt in case there is an error while accessing the memory.
>> + *
>> + * This function will always return (also for TCG), make sure to call
>> + * s390_cpu_virt_mem_handle_exc() to properly exit the CPU loop.
>>   */
>>  int s390_cpu_virt_mem_rw(S390CPU *cpu, vaddr laddr, uint8_t ar, void 
>> *hostbuf,
>>                           int len, bool is_write)
>> @@ -514,6 +518,19 @@ int s390_cpu_virt_mem_rw(S390CPU *cpu, vaddr laddr, 
>> uint8_t ar, void *hostbuf,
>>      return ret;
>>  }
>>  
>> +void s390_cpu_virt_mem_handle_exc(S390CPU *cpu, uintptr_t ra)
>> +{
>> +    /* KVM will handle the interrupt automatically, TCG has to exit the TB 
>> */
>> +    #ifdef CONFIG_TCG
> 
> Please move the #ifdef/#endif to the beginning of the line.

Oh, yes.

> 
>> +    if (tcg_enabled()) {
>> +        if (ra) {
>> +            cpu_restore_state(CPU(cpu), ra);
>> +        }
>> +        cpu_loop_exit(CPU(cpu));
>> +    }
>> +    #endif
>> +}
>> +
>>  /**
>>   * Translate a real address into a physical (absolute) address.
>>   * @param raddr  the real address
> 


-- 

Thanks,

David / dhildenb



reply via email to

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