From: Klaus Jensen <[email protected]>

Prior to reading the shadow doorbell cq head, we have to update the
eventidx. Otherwise, we risk that the driver will skip an mmio doorbell
write. This happens on riscv64, as reported by Guenter.

Adding the missing update to the cq eventidx fixes the issue.

Fixes: 3f7fe8de3d49 ("hw/nvme: Implement shadow doorbell buffer support")
Cc: [email protected]
Cc: [email protected]
Reported-by: Guenter Roeck <[email protected]>
Reviewed-by: Keith Busch <[email protected]>
Signed-off-by: Klaus Jensen <[email protected]>
---
 hw/nvme/ctrl.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
index 28e02ec7baa6..226480033771 100644
--- a/hw/nvme/ctrl.c
+++ b/hw/nvme/ctrl.c
@@ -1334,6 +1334,15 @@ static inline void nvme_blk_write(BlockBackend *blk, 
int64_t offset,
     }
 }
 
+static void nvme_update_cq_eventidx(const NvmeCQueue *cq)
+{
+    uint32_t v = cpu_to_le32(cq->head);
+
+    trace_pci_nvme_update_cq_eventidx(cq->cqid, cq->head);
+
+    pci_dma_write(PCI_DEVICE(cq->ctrl), cq->ei_addr, &v, sizeof(v));
+}
+
 static void nvme_update_cq_head(NvmeCQueue *cq)
 {
     uint32_t v;
@@ -1358,6 +1367,7 @@ static void nvme_post_cqes(void *opaque)
         hwaddr addr;
 
         if (n->dbbuf_enabled) {
+            nvme_update_cq_eventidx(cq);
             nvme_update_cq_head(cq);
         }
 
-- 
2.39.0


Reply via email to