On Mon, 2 Jul 2018 15:04:11 +0200 Claudio Imbrenda <imbre...@linux.vnet.ibm.com> wrote:
> The macro CMMA_BLOCK_SIZE was defined but not used, and a hardcoded > value was instead used in the code. > > This patch fixes the value of CMMA_BLOCK_SIZE and uses it in the > appropriate place in the code, and fixes another case of hardcoded > value in the KVM backend, replacing it with the more appropriate > constant KVM_S390_CMMA_SIZE_MAX. > > Signed-off-by: Claudio Imbrenda <imbre...@linux.vnet.ibm.com> > --- > hw/s390x/s390-stattrib-kvm.c | 3 ++- > hw/s390x/s390-stattrib.c | 5 +++-- > 2 files changed, 5 insertions(+), 3 deletions(-) > diff --git a/hw/s390x/s390-stattrib.c b/hw/s390x/s390-stattrib.c > index 70b9555..a788cdb 100644 > --- a/hw/s390x/s390-stattrib.c > +++ b/hw/s390x/s390-stattrib.c > @@ -20,7 +20,8 @@ > #include "qapi/error.h" > #include "qapi/qmp/qdict.h" > > -#define CMMA_BLOCK_SIZE (1 << 10) > +/* 512kB cover 2GB of guest memory */ > +#define CMMA_BLOCK_SIZE (1 << 19) Hm, this doesn't apply to current master anymore. Do we want to switch to something based on the IEC binary prefix definitions here, or keep the 1 << 19? > > #define STATTR_FLAG_EOS 0x01ULL > #define STATTR_FLAG_MORE 0x02ULL > @@ -202,7 +203,7 @@ static int cmma_save(QEMUFile *f, void *opaque, int final) > S390StAttribClass *sac = S390_STATTRIB_GET_CLASS(sas); > uint8_t *buf; > int r, cx, reallen = 0, ret = 0; > - uint32_t buflen = 1 << 19; /* 512kB cover 2GB of guest memory */ > + uint32_t buflen = CMMA_BLOCK_SIZE; > uint64_t start_gfn = sas->migration_cur_gfn; > > buf = g_try_malloc(buflen);