| Message ID | 32e33a0b-d572-ea7f-b1a8-3263711582e5@suse.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | x86/HVM: drop dead assignments from hvmemul_rep_{movs,stos}() | expand |
On 27/07/2023 4:25 pm, Jan Beulich wrote: > In the latter case the variable altogether is then unused and hence gets > dropped, eliminating a Misra Rule 5.3 violation. I'm afraid I mistakenly > introduced both assignments in 57a57465daaf ("x86/HVM: use available > linear->phys translations in REP MOVS/STOS handling"), likely as a > result of some re-work on the patch. > > Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
On Thu, 27 Jul 2023, Jan Beulich wrote: > In the latter case the variable altogether is then unused and hence gets > dropped, eliminating a Misra Rule 5.3 violation. I'm afraid I mistakenly > introduced both assignments in 57a57465daaf ("x86/HVM: use available > linear->phys translations in REP MOVS/STOS handling"), likely as a > result of some re-work on the patch. > > Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Stefano Stabellini <sstabellini@kernel.org> > --- a/xen/arch/x86/hvm/emulate.c > +++ b/xen/arch/x86/hvm/emulate.c > @@ -1861,7 +1861,6 @@ static int cf_check hvmemul_rep_movs( > return rc; > } > > - bytes = PAGE_SIZE - (daddr & ~PAGE_MASK); > if ( hvio->mmio_access.write_access && > (hvio->mmio_gla == (daddr & PAGE_MASK)) && > /* See comment above. */ > @@ -1988,7 +1987,7 @@ static int cf_check hvmemul_rep_stos( > container_of(ctxt, struct hvm_emulate_ctxt, ctxt); > struct vcpu *curr = current; > struct hvm_vcpu_io *hvio = &curr->arch.hvm.hvm_io; > - unsigned long addr, bytes; > + unsigned long addr; > paddr_t gpa; > p2m_type_t p2mt; > bool_t df = !!(ctxt->regs->eflags & X86_EFLAGS_DF); > @@ -1998,7 +1997,6 @@ static int cf_check hvmemul_rep_stos( > if ( rc != X86EMUL_OKAY ) > return rc; > > - bytes = PAGE_SIZE - (addr & ~PAGE_MASK); > if ( hvio->mmio_access.write_access && > (hvio->mmio_gla == (addr & PAGE_MASK)) && > /* See respective comment in MOVS processing. */ >
On 27.07.2023 20:41, Stefano Stabellini wrote: > On Thu, 27 Jul 2023, Jan Beulich wrote: >> In the latter case the variable altogether is then unused and hence gets >> dropped, eliminating a Misra Rule 5.3 violation. I'm afraid I mistakenly >> introduced both assignments in 57a57465daaf ("x86/HVM: use available >> linear->phys translations in REP MOVS/STOS handling"), likely as a >> result of some re-work on the patch. >> >> Signed-off-by: Jan Beulich <jbeulich@suse.com> > > Reviewed-by: Stefano Stabellini <sstabellini@kernel.org> Paul - any chance of an ack? Thanks, Jan
On 03/08/2023 12:46, Jan Beulich wrote: > On 27.07.2023 20:41, Stefano Stabellini wrote: >> On Thu, 27 Jul 2023, Jan Beulich wrote: >>> In the latter case the variable altogether is then unused and hence gets >>> dropped, eliminating a Misra Rule 5.3 violation. I'm afraid I mistakenly >>> introduced both assignments in 57a57465daaf ("x86/HVM: use available >>> linear->phys translations in REP MOVS/STOS handling"), likely as a >>> result of some re-work on the patch. >>> >>> Signed-off-by: Jan Beulich <jbeulich@suse.com> >> >> Reviewed-by: Stefano Stabellini <sstabellini@kernel.org> > > Paul - any chance of an ack? > Sure. Reviewed-by: Paul Durrant <paul@xen.org>
--- a/xen/arch/x86/hvm/emulate.c +++ b/xen/arch/x86/hvm/emulate.c @@ -1861,7 +1861,6 @@ static int cf_check hvmemul_rep_movs( return rc; } - bytes = PAGE_SIZE - (daddr & ~PAGE_MASK); if ( hvio->mmio_access.write_access && (hvio->mmio_gla == (daddr & PAGE_MASK)) && /* See comment above. */ @@ -1988,7 +1987,7 @@ static int cf_check hvmemul_rep_stos( container_of(ctxt, struct hvm_emulate_ctxt, ctxt); struct vcpu *curr = current; struct hvm_vcpu_io *hvio = &curr->arch.hvm.hvm_io; - unsigned long addr, bytes; + unsigned long addr; paddr_t gpa; p2m_type_t p2mt; bool_t df = !!(ctxt->regs->eflags & X86_EFLAGS_DF); @@ -1998,7 +1997,6 @@ static int cf_check hvmemul_rep_stos( if ( rc != X86EMUL_OKAY ) return rc; - bytes = PAGE_SIZE - (addr & ~PAGE_MASK); if ( hvio->mmio_access.write_access && (hvio->mmio_gla == (addr & PAGE_MASK)) && /* See respective comment in MOVS processing. */
In the latter case the variable altogether is then unused and hence gets dropped, eliminating a Misra Rule 5.3 violation. I'm afraid I mistakenly introduced both assignments in 57a57465daaf ("x86/HVM: use available linear->phys translations in REP MOVS/STOS handling"), likely as a result of some re-work on the patch. Signed-off-by: Jan Beulich <jbeulich@suse.com>