From: Alex Bennée <[email protected]> This allows us to keep the MSI data in plain host order all the way from the MemoryRegionOps write method to the final KVM_SIGNAL_MSI ioctl. This fixes a theoretical bug on big-endian hosts because we were using different size byte swaps which would have truncated the data.
Signed-off-by: Alex Bennée <[email protected]> Reviewed-by: Peter Maydell <[email protected]> Message-id: [email protected] Cc: [email protected] Signed-off-by: Peter Maydell <[email protected]> (cherry picked from commit cf10273aff8198ab1c7e2a00e7e5fe51c80b04e7) Signed-off-by: Michael Tokarev <[email protected]> diff --git a/hw/intc/arm_gicv3_its_common.c b/hw/intc/arm_gicv3_its_common.c index 70dbee83a6..6a8fb0e4f8 100644 --- a/hw/intc/arm_gicv3_its_common.c +++ b/hw/intc/arm_gicv3_its_common.c @@ -81,7 +81,7 @@ static MemTxResult gicv3_its_trans_write(void *opaque, hwaddr offset, if (offset == 0x0040 && ((size == 2) || (size == 4))) { GICv3ITSState *s = ARM_GICV3_ITS_COMMON(opaque); GICv3ITSCommonClass *c = ARM_GICV3_ITS_COMMON_GET_CLASS(s); - int ret = c->send_msi(s, le64_to_cpu(value), attrs.requester_id); + int ret = c->send_msi(s, value, attrs.requester_id); if (ret <= 0) { qemu_log_mask(LOG_GUEST_ERROR, diff --git a/hw/intc/arm_gicv3_its_kvm.c b/hw/intc/arm_gicv3_its_kvm.c index e198974560..362b0a5cc0 100644 --- a/hw/intc/arm_gicv3_its_kvm.c +++ b/hw/intc/arm_gicv3_its_kvm.c @@ -58,7 +58,7 @@ static int kvm_its_send_msi(GICv3ITSState *s, uint32_t value, uint16_t devid) msi.address_lo = extract64(s->gits_translater_gpa, 0, 32); msi.address_hi = extract64(s->gits_translater_gpa, 32, 32); - msi.data = le32_to_cpu(value); + msi.data = value; msi.flags = KVM_MSI_VALID_DEVID; msi.devid = devid; memset(msi.pad, 0, sizeof(msi.pad)); -- 2.47.3
