[02/17] hw/block/nvme: handle dma errors

From: Klaus Jensen <k.jensen@samsung.com>

Handling DMA errors gracefully is required for the device to pass the
block/011 test ("disable PCI device while doing I/O") in the blktests
suite.

With this patch the device passes the test by retrying "critical"
transfers (posting of completion entries and processing of submission
queue entries).

If DMA errors occur at any other point in the execution of the command
(say, while mapping the PRPs), the command is aborted with a Data
Transfer Error status code.

Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Acked-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
---
 hw/block/nvme.c       | 43 ++++++++++++++++++++++++++++++++-----------
 hw/block/trace-events |  2 ++
 include/block/nvme.h  |  2 +-
 3 files changed, 35 insertions(+), 12 deletions(-)

Hi Klaus,

On 9/4/20 4:19 PM, Klaus Jensen wrote:
> From: Klaus Jensen <k.jensen@samsung.com>
> 
> Handling DMA errors gracefully is required for the device to pass the
> block/011 test ("disable PCI device while doing I/O") in the blktests
> suite.
> 
> With this patch the device passes the test by retrying "critical"
> transfers (posting of completion entries and processing of submission
> queue entries).
> 
> If DMA errors occur at any other point in the execution of the command
> (say, while mapping the PRPs), the command is aborted with a Data
> Transfer Error status code.
> 
> Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
> Acked-by: Keith Busch <kbusch@kernel.org>
> Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
> ---
>  hw/block/nvme.c       | 43 ++++++++++++++++++++++++++++++++-----------
>  hw/block/trace-events |  2 ++
>  include/block/nvme.h  |  2 +-
>  3 files changed, 35 insertions(+), 12 deletions(-)
> 
> diff --git a/hw/block/nvme.c b/hw/block/nvme.c
> index 63078f600920..49bcdf31ced6 100644
> --- a/hw/block/nvme.c
> +++ b/hw/block/nvme.c
> @@ -140,14 +140,14 @@ static inline void *nvme_addr_to_cmb(NvmeCtrl *n, hwaddr addr)
>      return &n->cmbuf[addr - n->ctrl_mem.addr];
>  }
>  
> -static void nvme_addr_read(NvmeCtrl *n, hwaddr addr, void *buf, int size)
> +static int nvme_addr_read(NvmeCtrl *n, hwaddr addr, void *buf, int size)

If this get merged first:
https://www.mail-archive.com/qemu-devel@nongnu.org/msg737483.html
then please return MemTxResult, ...

>  {
>      if (n->bar.cmbsz && nvme_addr_is_cmb(n, addr)) {
>          memcpy(buf, nvme_addr_to_cmb(n, addr), size);
> -        return;
> +        return 0;
>      }
>  
> -    pci_dma_read(&n->parent_obj, addr, buf, size);
> +    return pci_dma_read(&n->parent_obj, addr, buf, size);
>  }
>  
>  static int nvme_check_sqid(NvmeCtrl *n, uint16_t sqid)
> @@ -253,7 +253,7 @@ static uint16_t nvme_map_addr_cmb(NvmeCtrl *n, QEMUIOVector *iov, hwaddr addr,
>      trace_pci_nvme_map_addr_cmb(addr, len);
>  
>      if (!nvme_addr_is_cmb(n, addr) || !nvme_addr_is_cmb(n, addr + len - 1)) {
> -        return NVME_DATA_TRAS_ERROR;
> +        return NVME_DATA_TRANSFER_ERROR;
>      }
>  
>      qemu_iovec_add(iov, nvme_addr_to_cmb(n, addr), len);
> @@ -307,6 +307,7 @@ static uint16_t nvme_map_prp(NvmeCtrl *n, uint64_t prp1, uint64_t prp2,
>      int num_prps = (len >> n->page_bits) + 1;
>      uint16_t status;
>      bool prp_list_in_cmb = false;
> +    int ret;
>  
>      QEMUSGList *qsg = &req->qsg;
>      QEMUIOVector *iov = &req->iov;
> @@ -347,7 +348,11 @@ static uint16_t nvme_map_prp(NvmeCtrl *n, uint64_t prp1, uint64_t prp2,
>  
>              nents = (len + n->page_size - 1) >> n->page_bits;
>              prp_trans = MIN(n->max_prp_ents, nents) * sizeof(uint64_t);
> -            nvme_addr_read(n, prp2, (void *)prp_list, prp_trans);
> +            ret = nvme_addr_read(n, prp2, (void *)prp_list, prp_trans);
> +            if (ret) {

... and check it (other cases following).

> +                trace_pci_nvme_err_addr_read(prp2);
> +                return NVME_DATA_TRANSFER_ERROR;
> +            }
>              while (len != 0) {
>                  uint64_t prp_ent = le64_to_cpu(prp_list[i]);
>  
> @@ -364,8 +369,12 @@ static uint16_t nvme_map_prp(NvmeCtrl *n, uint64_t prp1, uint64_t prp2,
>                      i = 0;
>                      nents = (len + n->page_size - 1) >> n->page_bits;
>                      prp_trans = MIN(n->max_prp_ents, nents) * sizeof(uint64_t);
> -                    nvme_addr_read(n, prp_ent, (void *)prp_list,
> -                        prp_trans);
> +                    ret = nvme_addr_read(n, prp_ent, (void *)prp_list,
> +                                         prp_trans);
> +                    if (ret) {
> +                        trace_pci_nvme_err_addr_read(prp_ent);
> +                        return NVME_DATA_TRANSFER_ERROR;
> +                    }
>                      prp_ent = le64_to_cpu(prp_list[i]);
>                  }
>  
> @@ -457,6 +466,7 @@ static void nvme_post_cqes(void *opaque)
>      NvmeCQueue *cq = opaque;
>      NvmeCtrl *n = cq->ctrl;
>      NvmeRequest *req, *next;
> +    int ret;
>  
>      QTAILQ_FOREACH_SAFE(req, &cq->req_list, entry, next) {
>          NvmeSQueue *sq;
> @@ -466,15 +476,21 @@ static void nvme_post_cqes(void *opaque)
>              break;
>          }
>  
> -        QTAILQ_REMOVE(&cq->req_list, req, entry);
>          sq = req->sq;
>          req->cqe.status = cpu_to_le16((req->status << 1) | cq->phase);
>          req->cqe.sq_id = cpu_to_le16(sq->sqid);
>          req->cqe.sq_head = cpu_to_le16(sq->head);
>          addr = cq->dma_addr + cq->tail * n->cqe_size;
> +        ret = pci_dma_write(&n->parent_obj, addr, (void *)&req->cqe,
> +                            sizeof(req->cqe));
> +        if (ret) {
> +            trace_pci_nvme_err_addr_write(addr);
> +            timer_mod(cq->timer, qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) +
> +                      500 * SCALE_MS);
> +            break;
> +        }
> +        QTAILQ_REMOVE(&cq->req_list, req, entry);
>          nvme_inc_cq_tail(cq);
> -        pci_dma_write(&n->parent_obj, addr, (void *)&req->cqe,
> -            sizeof(req->cqe));
>          nvme_req_exit(req);
>          QTAILQ_INSERT_TAIL(&sq->req_list, req, entry);
>      }
> @@ -1611,7 +1627,12 @@ static void nvme_process_sq(void *opaque)
>  
>      while (!(nvme_sq_empty(sq) || QTAILQ_EMPTY(&sq->req_list))) {
>          addr = sq->dma_addr + sq->head * n->sqe_size;
> -        nvme_addr_read(n, addr, (void *)&cmd, sizeof(cmd));
> +        if (nvme_addr_read(n, addr, (void *)&cmd, sizeof(cmd))) {
> +            trace_pci_nvme_err_addr_read(addr);
> +            timer_mod(sq->timer, qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) +
> +                      500 * SCALE_MS);
> +            break;
> +        }
>          nvme_inc_sq_head(sq);
>  
>          req = QTAILQ_FIRST(&sq->req_list);
> diff --git a/hw/block/trace-events b/hw/block/trace-events
> index 72cf2d15cb8e..50d5702e6b80 100644
> --- a/hw/block/trace-events
> +++ b/hw/block/trace-events
> @@ -86,6 +86,8 @@ pci_nvme_mmio_shutdown_cleared(void) "shutdown bit cleared"
>  
>  # nvme traces for error conditions
>  pci_nvme_err_mdts(uint16_t cid, size_t len) "cid %"PRIu16" len %zu"
> +pci_nvme_err_addr_read(uint64_t addr) "addr 0x%"PRIx64""
> +pci_nvme_err_addr_write(uint64_t addr) "addr 0x%"PRIx64""
>  pci_nvme_err_invalid_dma(void) "PRP/SGL is too small for transfer size"
>  pci_nvme_err_invalid_prplist_ent(uint64_t prplist) "PRP list entry is null or not page aligned: 0x%"PRIx64""
>  pci_nvme_err_invalid_prp2_align(uint64_t prp2) "PRP2 is not page aligned: 0x%"PRIx64""
> diff --git a/include/block/nvme.h b/include/block/nvme.h
> index 65e68a82c897..c8d0a3473f0d 100644
> --- a/include/block/nvme.h
> +++ b/include/block/nvme.h
> @@ -630,7 +630,7 @@ enum NvmeStatusCodes {
>      NVME_INVALID_OPCODE         = 0x0001,
>      NVME_INVALID_FIELD          = 0x0002,
>      NVME_CID_CONFLICT           = 0x0003,
> -    NVME_DATA_TRAS_ERROR        = 0x0004,
> +    NVME_DATA_TRANSFER_ERROR    = 0x0004,
>      NVME_POWER_LOSS_ABORT       = 0x0005,
>      NVME_INTERNAL_DEV_ERROR     = 0x0006,
>      NVME_CMD_ABORT_REQ          = 0x0007,
>