KVM: s390: guestdbg: separate defines for per code
authorDavid Hildenbrand <dahi@linux.vnet.ibm.com>
Fri, 27 May 2016 13:24:33 +0000 (15:24 +0200)
committerChristian Borntraeger <borntraeger@de.ibm.com>
Thu, 8 Sep 2016 07:07:52 +0000 (09:07 +0200)
Let's avoid working with the PER_EVENT* defines, used for control register
manipulation, when checking the u8 PER code. Introduce separate defines
based on the existing defines.

Reviewed-by: Eric Farman <farman@linux.vnet.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
arch/s390/kvm/guestdbg.c

index 31a05330d11c77fc0c34cd30ea54bc2075a2c3d9..d1f8241993a48b7af0d02ec21b118fdc63900284 100644 (file)
@@ -382,14 +382,20 @@ void kvm_s390_prepare_debug_exit(struct kvm_vcpu *vcpu)
        vcpu->guest_debug &= ~KVM_GUESTDBG_EXIT_PENDING;
 }
 
+#define PER_CODE_MASK          (PER_EVENT_MASK >> 24)
+#define PER_CODE_BRANCH                (PER_EVENT_BRANCH >> 24)
+#define PER_CODE_IFETCH                (PER_EVENT_IFETCH >> 24)
+#define PER_CODE_STORE         (PER_EVENT_STORE >> 24)
+#define PER_CODE_STORE_REAL    (PER_EVENT_STORE_REAL >> 24)
+
 #define per_bp_event(code) \
-                       (code & (PER_EVENT_IFETCH | PER_EVENT_BRANCH))
+                       (code & (PER_CODE_IFETCH | PER_CODE_BRANCH))
 #define per_write_wp_event(code) \
-                       (code & (PER_EVENT_STORE | PER_EVENT_STORE_REAL))
+                       (code & (PER_CODE_STORE | PER_CODE_STORE_REAL))
 
 static int debug_exit_required(struct kvm_vcpu *vcpu)
 {
-       u32 perc = (vcpu->arch.sie_block->perc << 24);
+       u8 perc = vcpu->arch.sie_block->perc;
        struct kvm_debug_exit_arch *debug_exit = &vcpu->run->debug.arch;
        struct kvm_hw_wp_info_arch *wp_info = NULL;
        struct kvm_hw_bp_info_arch *bp_info = NULL;
@@ -444,7 +450,7 @@ int kvm_s390_handle_per_ifetch_icpt(struct kvm_vcpu *vcpu)
        const u8 ilen = kvm_s390_get_ilen(vcpu);
        struct kvm_s390_pgm_info pgm_info = {
                .code = PGM_PER,
-               .per_code = PER_EVENT_IFETCH >> 24,
+               .per_code = PER_CODE_IFETCH,
                .per_address = __rewind_psw(vcpu->arch.sie_block->gpsw, ilen),
        };
 
@@ -458,33 +464,33 @@ int kvm_s390_handle_per_ifetch_icpt(struct kvm_vcpu *vcpu)
 
 static void filter_guest_per_event(struct kvm_vcpu *vcpu)
 {
-       u32 perc = vcpu->arch.sie_block->perc << 24;
+       const u8 perc = vcpu->arch.sie_block->perc;
        u64 peraddr = vcpu->arch.sie_block->peraddr;
        u64 addr = vcpu->arch.sie_block->gpsw.addr;
        u64 cr9 = vcpu->arch.sie_block->gcr[9];
        u64 cr10 = vcpu->arch.sie_block->gcr[10];
        u64 cr11 = vcpu->arch.sie_block->gcr[11];
        /* filter all events, demanded by the guest */
-       u32 guest_perc = perc & cr9 & PER_EVENT_MASK;
+       u8 guest_perc = perc & (cr9 >> 24) & PER_CODE_MASK;
 
        if (!guest_per_enabled(vcpu))
                guest_perc = 0;
 
        /* filter "successful-branching" events */
-       if (guest_perc & PER_EVENT_BRANCH &&
+       if (guest_perc & PER_CODE_BRANCH &&
            cr9 & PER_CONTROL_BRANCH_ADDRESS &&
            !in_addr_range(addr, cr10, cr11))
-               guest_perc &= ~PER_EVENT_BRANCH;
+               guest_perc &= ~PER_CODE_BRANCH;
 
        /* filter "instruction-fetching" events */
-       if (guest_perc & PER_EVENT_IFETCH &&
+       if (guest_perc & PER_CODE_IFETCH &&
            !in_addr_range(peraddr, cr10, cr11))
-               guest_perc &= ~PER_EVENT_IFETCH;
+               guest_perc &= ~PER_CODE_IFETCH;
 
        /* All other PER events will be given to the guest */
        /* TODO: Check altered address/address space */
 
-       vcpu->arch.sie_block->perc = guest_perc >> 24;
+       vcpu->arch.sie_block->perc = guest_perc;
 
        if (!guest_perc)
                vcpu->arch.sie_block->iprcc &= ~PGM_PER;