diff mbox series

[v5,4/5] hw/nvme: fix mmio read

Message ID 20210719224647.68559-5-its@irrelevant.dk
State New
Headers show
Series hw/nvme: fix mmio read | expand

Commit Message

Klaus Jensen July 19, 2021, 10:46 p.m. UTC
From: Klaus Jensen <k.jensen@samsung.com>

The new PMR test unearthed a long-standing issue with MMIO reads on
big-endian hosts.

Fix this by unconditionally storing all controller registers in little
endian.

Cc: Gollu Appalanaidu <anaidu.gollu@samsung.com>
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
---
 hw/nvme/ctrl.c | 290 +++++++++++++++++++++++++++----------------------
 1 file changed, 162 insertions(+), 128 deletions(-)

Comments

Philippe Mathieu-Daudé July 20, 2021, 12:58 p.m. UTC | #1
On 7/20/21 12:46 AM, Klaus Jensen wrote:
> From: Klaus Jensen <k.jensen@samsung.com>
> 
> The new PMR test unearthed a long-standing issue with MMIO reads on
> big-endian hosts.
> 
> Fix this by unconditionally storing all controller registers in little
> endian.
> 
> Cc: Gollu Appalanaidu <anaidu.gollu@samsung.com>
> Reported-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
> ---
>  hw/nvme/ctrl.c | 290 +++++++++++++++++++++++++++----------------------
>  1 file changed, 162 insertions(+), 128 deletions(-)
> 
> diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
> index 0449cc4dee9b..43dfaeac9f54 100644
> --- a/hw/nvme/ctrl.c
> +++ b/hw/nvme/ctrl.c
> @@ -439,10 +439,12 @@ static uint8_t nvme_sq_empty(NvmeSQueue *sq)
>  
>  static void nvme_irq_check(NvmeCtrl *n)
>  {
> +    uint32_t intms = ldl_le_p(&n->bar.intms);
> +
>      if (msix_enabled(&(n->parent_obj))) {
>          return;
>      }
> -    if (~n->bar.intms & n->irq_status) {

Why not use an inline call like the rest of this file?

       if (~ldl_le_p(&n->bar.intms) & n->irq_status) {

Anyway, not an issue.

> +    if (~intms & n->irq_status) {
>          pci_irq_assert(&n->parent_obj);
>      } else {
>          pci_irq_deassert(&n->parent_obj);
> @@ -1289,7 +1291,7 @@ static void nvme_post_cqes(void *opaque)
>          if (ret) {
>              trace_pci_nvme_err_addr_write(addr);
>              trace_pci_nvme_err_cfs();
> -            n->bar.csts = NVME_CSTS_FAILED;
> +            stl_le_p(&n->bar.csts, NVME_CSTS_FAILED);
>              break;
>          }
>          QTAILQ_REMOVE(&cq->req_list, req, entry);
> @@ -4022,7 +4024,7 @@ static uint16_t nvme_create_sq(NvmeCtrl *n, NvmeRequest *req)
>          trace_pci_nvme_err_invalid_create_sq_sqid(sqid);
>          return NVME_INVALID_QID | NVME_DNR;
>      }
> -    if (unlikely(!qsize || qsize > NVME_CAP_MQES(n->bar.cap))) {
> +    if (unlikely(!qsize || qsize > NVME_CAP_MQES(ldq_le_p(&n->bar.cap)))) {
>          trace_pci_nvme_err_invalid_create_sq_size(qsize);
>          return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR;
>      }
> @@ -4208,7 +4210,7 @@ static uint16_t nvme_cmd_effects(NvmeCtrl *n, uint8_t csi, uint32_t buf_len,
>          return NVME_INVALID_FIELD | NVME_DNR;
>      }
>  
> -    switch (NVME_CC_CSS(n->bar.cc)) {
> +    switch (NVME_CC_CSS(ldl_le_p(&n->bar.cc))) {
>      case NVME_CC_CSS_NVM:
>          src_iocs = nvme_cse_iocs_nvm;
>          /* fall through */
> @@ -4370,7 +4372,7 @@ static uint16_t nvme_create_cq(NvmeCtrl *n, NvmeRequest *req)
>          trace_pci_nvme_err_invalid_create_cq_cqid(cqid);
>          return NVME_INVALID_QID | NVME_DNR;
>      }
> -    if (unlikely(!qsize || qsize > NVME_CAP_MQES(n->bar.cap))) {
> +    if (unlikely(!qsize || qsize > NVME_CAP_MQES(ldq_le_p(&n->bar.cap)))) {
>          trace_pci_nvme_err_invalid_create_cq_size(qsize);
>          return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR;
>      }
> @@ -5163,17 +5165,19 @@ static void nvme_update_dmrsl(NvmeCtrl *n)
>  
>  static void nvme_select_iocs_ns(NvmeCtrl *n, NvmeNamespace *ns)
>  {
> +    uint32_t cc = ldl_le_p(&n->bar.cc);

This one is understandable.

>      ns->iocs = nvme_cse_iocs_none;
>      switch (ns->csi) {
>      case NVME_CSI_NVM:
> -        if (NVME_CC_CSS(n->bar.cc) != NVME_CC_CSS_ADMIN_ONLY) {
> +        if (NVME_CC_CSS(cc) != NVME_CC_CSS_ADMIN_ONLY) {
>              ns->iocs = nvme_cse_iocs_nvm;
>          }
>          break;
>      case NVME_CSI_ZONED:
> -        if (NVME_CC_CSS(n->bar.cc) == NVME_CC_CSS_CSI) {
> +        if (NVME_CC_CSS(cc) == NVME_CC_CSS_CSI) {
>              ns->iocs = nvme_cse_iocs_zoned;
> -        } else if (NVME_CC_CSS(n->bar.cc) == NVME_CC_CSS_NVM) {
> +        } else if (NVME_CC_CSS(cc) == NVME_CC_CSS_NVM) {
>              ns->iocs = nvme_cse_iocs_nvm;
>          }
>          break;
> @@ -5510,7 +5514,7 @@ static void nvme_process_sq(void *opaque)
>          if (nvme_addr_read(n, addr, (void *)&cmd, sizeof(cmd))) {
>              trace_pci_nvme_err_addr_read(addr);
>              trace_pci_nvme_err_cfs();
> -            n->bar.csts = NVME_CSTS_FAILED;
> +            stl_le_p(&n->bar.csts, NVME_CSTS_FAILED);
>              break;
>          }
>          nvme_inc_sq_head(sq);
> @@ -5565,8 +5569,6 @@ static void nvme_ctrl_reset(NvmeCtrl *n)
>      n->aer_queued = 0;
>      n->outstanding_aers = 0;
>      n->qs_created = false;
> -
> -    n->bar.cc = 0;

This change is not documented, is it related to the fix?

>  }
>  
>  static void nvme_ctrl_shutdown(NvmeCtrl *n)
> @@ -5605,7 +5607,12 @@ static void nvme_select_iocs(NvmeCtrl *n)
>  
>  static int nvme_start_ctrl(NvmeCtrl *n)
>  {
> -    uint32_t page_bits = NVME_CC_MPS(n->bar.cc) + 12;
> +    uint64_t cap = ldq_le_p(&n->bar.cap);
> +    uint32_t cc = ldl_le_p(&n->bar.cc);
> +    uint32_t aqa = ldl_le_p(&n->bar.aqa);
> +    uint64_t asq = ldq_le_p(&n->bar.asq);
> +    uint64_t acq = ldq_le_p(&n->bar.acq);
> +    uint32_t page_bits = NVME_CC_MPS(cc) + 12;
>      uint32_t page_size = 1 << page_bits;

My brain overflowed at this point, too many changes to track :/

Would it make sense to split it? (per big function body maybe?)


Note, using so many manual endian accesses seems fragile. Maybe we
could find a way to write a pair of macros taking 'n' + bar 'fieldname'
as input, using the correct size swapping using sizeof_field(fieldname)?

Something like (untested):

  #define ld_bar(fieldname) \
          ldn_le_p(&n->bar.fieldname, \
                   sizeof_field(NvmeBar, fieldname));

  #define st_bar(fieldname, val) \
          stn_le_p(&n->bar.fieldname, \
                   sizeof_field(NvmeBar, fieldname), val);

Using as:

  uint64_t cap = ld_bar(cap);

Or if you prefer:

  #define BAR_LD(bar, fieldname) \
          ldn_le_p(pbar.fieldname, \
                   sizeof_field(NvmeBar, fieldname));

as:

  uint64_t cap = BAR_LD(&n->bar, cap);

Regards,

Phil.
Peter Maydell July 20, 2021, 1:33 p.m. UTC | #2
On Tue, 20 Jul 2021 at 13:58, Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
>
> On 7/20/21 12:46 AM, Klaus Jensen wrote:
> > From: Klaus Jensen <k.jensen@samsung.com>
> >
> > The new PMR test unearthed a long-standing issue with MMIO reads on
> > big-endian hosts.
> >
> > Fix this by unconditionally storing all controller registers in little
> > endian.

> My brain overflowed at this point, too many changes to track :/
>
> Would it make sense to split it? (per big function body maybe?)

I did review of (a previous version of) the patch by:
 (1) check that after the patch is applied there aren't any
     direct accesses to n->bar.anything which don't go via a
     ld*/st* (ie it didn't miss anything)
 (2) check that the accesses which use ldq/stq are consistently
     doing so for those fields and that they are uint64_t
 (3) read through the patch and check all the changes are ok
     (either using inline calls to the accessors, or fishing
     the value out at the start of the function)

It is a big patch, but I'm not sure splitting it apart helps much.

thanks
-- PMM
Peter Maydell July 20, 2021, 1:37 p.m. UTC | #3
On Mon, 19 Jul 2021 at 23:47, Klaus Jensen <its@irrelevant.dk> wrote:
>
> From: Klaus Jensen <k.jensen@samsung.com>
>
> The new PMR test unearthed a long-standing issue with MMIO reads on
> big-endian hosts.
>
> Fix this by unconditionally storing all controller registers in little
> endian.
>
> Cc: Gollu Appalanaidu <anaidu.gollu@samsung.com>
> Reported-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
> ---
>  hw/nvme/ctrl.c | 290 +++++++++++++++++++++++++++----------------------
>  1 file changed, 162 insertions(+), 128 deletions(-)

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM
Philippe Mathieu-Daudé July 20, 2021, 1:39 p.m. UTC | #4
On 7/20/21 3:33 PM, Peter Maydell wrote:
> On Tue, 20 Jul 2021 at 13:58, Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
>>
>> On 7/20/21 12:46 AM, Klaus Jensen wrote:
>>> From: Klaus Jensen <k.jensen@samsung.com>
>>>
>>> The new PMR test unearthed a long-standing issue with MMIO reads on
>>> big-endian hosts.
>>>
>>> Fix this by unconditionally storing all controller registers in little
>>> endian.
> 
>> My brain overflowed at this point, too many changes to track :/
>>
>> Would it make sense to split it? (per big function body maybe?)
> 
> I did review of (a previous version of) the patch by:
>  (1) check that after the patch is applied there aren't any
>      direct accesses to n->bar.anything which don't go via a
>      ld*/st* (ie it didn't miss anything)

OK

>  (2) check that the accesses which use ldq/stq are consistently
>      doing so for those fields and that they are uint64_t

OK

>  (3) read through the patch and check all the changes are ok
>      (either using inline calls to the accessors, or fishing
>      the value out at the start of the function)

OK, perfect then :)

> It is a big patch, but I'm not sure splitting it apart helps much.

Since it is reviewed, not needed.

Thanks a lot!

Phil.
diff mbox series

Patch

diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
index 0449cc4dee9b..43dfaeac9f54 100644
--- a/hw/nvme/ctrl.c
+++ b/hw/nvme/ctrl.c
@@ -439,10 +439,12 @@  static uint8_t nvme_sq_empty(NvmeSQueue *sq)
 
 static void nvme_irq_check(NvmeCtrl *n)
 {
+    uint32_t intms = ldl_le_p(&n->bar.intms);
+
     if (msix_enabled(&(n->parent_obj))) {
         return;
     }
-    if (~n->bar.intms & n->irq_status) {
+    if (~intms & n->irq_status) {
         pci_irq_assert(&n->parent_obj);
     } else {
         pci_irq_deassert(&n->parent_obj);
@@ -1289,7 +1291,7 @@  static void nvme_post_cqes(void *opaque)
         if (ret) {
             trace_pci_nvme_err_addr_write(addr);
             trace_pci_nvme_err_cfs();
-            n->bar.csts = NVME_CSTS_FAILED;
+            stl_le_p(&n->bar.csts, NVME_CSTS_FAILED);
             break;
         }
         QTAILQ_REMOVE(&cq->req_list, req, entry);
@@ -4022,7 +4024,7 @@  static uint16_t nvme_create_sq(NvmeCtrl *n, NvmeRequest *req)
         trace_pci_nvme_err_invalid_create_sq_sqid(sqid);
         return NVME_INVALID_QID | NVME_DNR;
     }
-    if (unlikely(!qsize || qsize > NVME_CAP_MQES(n->bar.cap))) {
+    if (unlikely(!qsize || qsize > NVME_CAP_MQES(ldq_le_p(&n->bar.cap)))) {
         trace_pci_nvme_err_invalid_create_sq_size(qsize);
         return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR;
     }
@@ -4208,7 +4210,7 @@  static uint16_t nvme_cmd_effects(NvmeCtrl *n, uint8_t csi, uint32_t buf_len,
         return NVME_INVALID_FIELD | NVME_DNR;
     }
 
-    switch (NVME_CC_CSS(n->bar.cc)) {
+    switch (NVME_CC_CSS(ldl_le_p(&n->bar.cc))) {
     case NVME_CC_CSS_NVM:
         src_iocs = nvme_cse_iocs_nvm;
         /* fall through */
@@ -4370,7 +4372,7 @@  static uint16_t nvme_create_cq(NvmeCtrl *n, NvmeRequest *req)
         trace_pci_nvme_err_invalid_create_cq_cqid(cqid);
         return NVME_INVALID_QID | NVME_DNR;
     }
-    if (unlikely(!qsize || qsize > NVME_CAP_MQES(n->bar.cap))) {
+    if (unlikely(!qsize || qsize > NVME_CAP_MQES(ldq_le_p(&n->bar.cap)))) {
         trace_pci_nvme_err_invalid_create_cq_size(qsize);
         return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR;
     }
@@ -5163,17 +5165,19 @@  static void nvme_update_dmrsl(NvmeCtrl *n)
 
 static void nvme_select_iocs_ns(NvmeCtrl *n, NvmeNamespace *ns)
 {
+    uint32_t cc = ldl_le_p(&n->bar.cc);
+
     ns->iocs = nvme_cse_iocs_none;
     switch (ns->csi) {
     case NVME_CSI_NVM:
-        if (NVME_CC_CSS(n->bar.cc) != NVME_CC_CSS_ADMIN_ONLY) {
+        if (NVME_CC_CSS(cc) != NVME_CC_CSS_ADMIN_ONLY) {
             ns->iocs = nvme_cse_iocs_nvm;
         }
         break;
     case NVME_CSI_ZONED:
-        if (NVME_CC_CSS(n->bar.cc) == NVME_CC_CSS_CSI) {
+        if (NVME_CC_CSS(cc) == NVME_CC_CSS_CSI) {
             ns->iocs = nvme_cse_iocs_zoned;
-        } else if (NVME_CC_CSS(n->bar.cc) == NVME_CC_CSS_NVM) {
+        } else if (NVME_CC_CSS(cc) == NVME_CC_CSS_NVM) {
             ns->iocs = nvme_cse_iocs_nvm;
         }
         break;
@@ -5510,7 +5514,7 @@  static void nvme_process_sq(void *opaque)
         if (nvme_addr_read(n, addr, (void *)&cmd, sizeof(cmd))) {
             trace_pci_nvme_err_addr_read(addr);
             trace_pci_nvme_err_cfs();
-            n->bar.csts = NVME_CSTS_FAILED;
+            stl_le_p(&n->bar.csts, NVME_CSTS_FAILED);
             break;
         }
         nvme_inc_sq_head(sq);
@@ -5565,8 +5569,6 @@  static void nvme_ctrl_reset(NvmeCtrl *n)
     n->aer_queued = 0;
     n->outstanding_aers = 0;
     n->qs_created = false;
-
-    n->bar.cc = 0;
 }
 
 static void nvme_ctrl_shutdown(NvmeCtrl *n)
@@ -5605,7 +5607,12 @@  static void nvme_select_iocs(NvmeCtrl *n)
 
 static int nvme_start_ctrl(NvmeCtrl *n)
 {
-    uint32_t page_bits = NVME_CC_MPS(n->bar.cc) + 12;
+    uint64_t cap = ldq_le_p(&n->bar.cap);
+    uint32_t cc = ldl_le_p(&n->bar.cc);
+    uint32_t aqa = ldl_le_p(&n->bar.aqa);
+    uint64_t asq = ldq_le_p(&n->bar.asq);
+    uint64_t acq = ldq_le_p(&n->bar.acq);
+    uint32_t page_bits = NVME_CC_MPS(cc) + 12;
     uint32_t page_size = 1 << page_bits;
 
     if (unlikely(n->cq[0])) {
@@ -5616,73 +5623,72 @@  static int nvme_start_ctrl(NvmeCtrl *n)
         trace_pci_nvme_err_startfail_sq();
         return -1;
     }
-    if (unlikely(!n->bar.asq)) {
+    if (unlikely(!asq)) {
         trace_pci_nvme_err_startfail_nbarasq();
         return -1;
     }
-    if (unlikely(!n->bar.acq)) {
+    if (unlikely(!acq)) {
         trace_pci_nvme_err_startfail_nbaracq();
         return -1;
     }
-    if (unlikely(n->bar.asq & (page_size - 1))) {
-        trace_pci_nvme_err_startfail_asq_misaligned(n->bar.asq);
+    if (unlikely(asq & (page_size - 1))) {
+        trace_pci_nvme_err_startfail_asq_misaligned(asq);
         return -1;
     }
-    if (unlikely(n->bar.acq & (page_size - 1))) {
-        trace_pci_nvme_err_startfail_acq_misaligned(n->bar.acq);
+    if (unlikely(acq & (page_size - 1))) {
+        trace_pci_nvme_err_startfail_acq_misaligned(acq);
         return -1;
     }
-    if (unlikely(!(NVME_CAP_CSS(n->bar.cap) & (1 << NVME_CC_CSS(n->bar.cc))))) {
-        trace_pci_nvme_err_startfail_css(NVME_CC_CSS(n->bar.cc));
+    if (unlikely(!(NVME_CAP_CSS(cap) & (1 << NVME_CC_CSS(cc))))) {
+        trace_pci_nvme_err_startfail_css(NVME_CC_CSS(cc));
         return -1;
     }
-    if (unlikely(NVME_CC_MPS(n->bar.cc) <
-                 NVME_CAP_MPSMIN(n->bar.cap))) {
+    if (unlikely(NVME_CC_MPS(cc) < NVME_CAP_MPSMIN(cap))) {
         trace_pci_nvme_err_startfail_page_too_small(
-                    NVME_CC_MPS(n->bar.cc),
-                    NVME_CAP_MPSMIN(n->bar.cap));
+                    NVME_CC_MPS(cc),
+                    NVME_CAP_MPSMIN(cap));
         return -1;
     }
-    if (unlikely(NVME_CC_MPS(n->bar.cc) >
-                 NVME_CAP_MPSMAX(n->bar.cap))) {
+    if (unlikely(NVME_CC_MPS(cc) >
+                 NVME_CAP_MPSMAX(cap))) {
         trace_pci_nvme_err_startfail_page_too_large(
-                    NVME_CC_MPS(n->bar.cc),
-                    NVME_CAP_MPSMAX(n->bar.cap));
+                    NVME_CC_MPS(cc),
+                    NVME_CAP_MPSMAX(cap));
         return -1;
     }
-    if (unlikely(NVME_CC_IOCQES(n->bar.cc) <
+    if (unlikely(NVME_CC_IOCQES(cc) <
                  NVME_CTRL_CQES_MIN(n->id_ctrl.cqes))) {
         trace_pci_nvme_err_startfail_cqent_too_small(
-                    NVME_CC_IOCQES(n->bar.cc),
-                    NVME_CTRL_CQES_MIN(n->bar.cap));
+                    NVME_CC_IOCQES(cc),
+                    NVME_CTRL_CQES_MIN(cap));
         return -1;
     }
-    if (unlikely(NVME_CC_IOCQES(n->bar.cc) >
+    if (unlikely(NVME_CC_IOCQES(cc) >
                  NVME_CTRL_CQES_MAX(n->id_ctrl.cqes))) {
         trace_pci_nvme_err_startfail_cqent_too_large(
-                    NVME_CC_IOCQES(n->bar.cc),
-                    NVME_CTRL_CQES_MAX(n->bar.cap));
+                    NVME_CC_IOCQES(cc),
+                    NVME_CTRL_CQES_MAX(cap));
         return -1;
     }
-    if (unlikely(NVME_CC_IOSQES(n->bar.cc) <
+    if (unlikely(NVME_CC_IOSQES(cc) <
                  NVME_CTRL_SQES_MIN(n->id_ctrl.sqes))) {
         trace_pci_nvme_err_startfail_sqent_too_small(
-                    NVME_CC_IOSQES(n->bar.cc),
-                    NVME_CTRL_SQES_MIN(n->bar.cap));
+                    NVME_CC_IOSQES(cc),
+                    NVME_CTRL_SQES_MIN(cap));
         return -1;
     }
-    if (unlikely(NVME_CC_IOSQES(n->bar.cc) >
+    if (unlikely(NVME_CC_IOSQES(cc) >
                  NVME_CTRL_SQES_MAX(n->id_ctrl.sqes))) {
         trace_pci_nvme_err_startfail_sqent_too_large(
-                    NVME_CC_IOSQES(n->bar.cc),
-                    NVME_CTRL_SQES_MAX(n->bar.cap));
+                    NVME_CC_IOSQES(cc),
+                    NVME_CTRL_SQES_MAX(cap));
         return -1;
     }
-    if (unlikely(!NVME_AQA_ASQS(n->bar.aqa))) {
+    if (unlikely(!NVME_AQA_ASQS(aqa))) {
         trace_pci_nvme_err_startfail_asqent_sz_zero();
         return -1;
     }
-    if (unlikely(!NVME_AQA_ACQS(n->bar.aqa))) {
+    if (unlikely(!NVME_AQA_ACQS(aqa))) {
         trace_pci_nvme_err_startfail_acqent_sz_zero();
         return -1;
     }
@@ -5690,12 +5696,10 @@  static int nvme_start_ctrl(NvmeCtrl *n)
     n->page_bits = page_bits;
     n->page_size = page_size;
     n->max_prp_ents = n->page_size / sizeof(uint64_t);
-    n->cqe_size = 1 << NVME_CC_IOCQES(n->bar.cc);
-    n->sqe_size = 1 << NVME_CC_IOSQES(n->bar.cc);
-    nvme_init_cq(&n->admin_cq, n, n->bar.acq, 0, 0,
-                 NVME_AQA_ACQS(n->bar.aqa) + 1, 1);
-    nvme_init_sq(&n->admin_sq, n, n->bar.asq, 0, 0,
-                 NVME_AQA_ASQS(n->bar.aqa) + 1);
+    n->cqe_size = 1 << NVME_CC_IOCQES(cc);
+    n->sqe_size = 1 << NVME_CC_IOSQES(cc);
+    nvme_init_cq(&n->admin_cq, n, acq, 0, 0, NVME_AQA_ACQS(aqa) + 1, 1);
+    nvme_init_sq(&n->admin_sq, n, asq, 0, 0, NVME_AQA_ASQS(aqa) + 1);
 
     nvme_set_timestamp(n, 0ULL);
 
@@ -5708,22 +5712,33 @@  static int nvme_start_ctrl(NvmeCtrl *n)
 
 static void nvme_cmb_enable_regs(NvmeCtrl *n)
 {
-    NVME_CMBLOC_SET_CDPCILS(n->bar.cmbloc, 1);
-    NVME_CMBLOC_SET_CDPMLS(n->bar.cmbloc, 1);
-    NVME_CMBLOC_SET_BIR(n->bar.cmbloc, NVME_CMB_BIR);
+    uint32_t cmbloc = ldl_le_p(&n->bar.cmbloc);
+    uint32_t cmbsz = ldl_le_p(&n->bar.cmbsz);
 
-    NVME_CMBSZ_SET_SQS(n->bar.cmbsz, 1);
-    NVME_CMBSZ_SET_CQS(n->bar.cmbsz, 0);
-    NVME_CMBSZ_SET_LISTS(n->bar.cmbsz, 1);
-    NVME_CMBSZ_SET_RDS(n->bar.cmbsz, 1);
-    NVME_CMBSZ_SET_WDS(n->bar.cmbsz, 1);
-    NVME_CMBSZ_SET_SZU(n->bar.cmbsz, 2); /* MBs */
-    NVME_CMBSZ_SET_SZ(n->bar.cmbsz, n->params.cmb_size_mb);
+    NVME_CMBLOC_SET_CDPCILS(cmbloc, 1);
+    NVME_CMBLOC_SET_CDPMLS(cmbloc, 1);
+    NVME_CMBLOC_SET_BIR(cmbloc, NVME_CMB_BIR);
+    stl_le_p(&n->bar.cmbloc, cmbloc);
+
+    NVME_CMBSZ_SET_SQS(cmbsz, 1);
+    NVME_CMBSZ_SET_CQS(cmbsz, 0);
+    NVME_CMBSZ_SET_LISTS(cmbsz, 1);
+    NVME_CMBSZ_SET_RDS(cmbsz, 1);
+    NVME_CMBSZ_SET_WDS(cmbsz, 1);
+    NVME_CMBSZ_SET_SZU(cmbsz, 2); /* MBs */
+    NVME_CMBSZ_SET_SZ(cmbsz, n->params.cmb_size_mb);
+    stl_le_p(&n->bar.cmbsz, cmbsz);
 }
 
 static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
                            unsigned size)
 {
+    uint64_t cap = ldq_le_p(&n->bar.cap);
+    uint32_t cc = ldl_le_p(&n->bar.cc);
+    uint32_t intms = ldl_le_p(&n->bar.intms);
+    uint32_t csts = ldl_le_p(&n->bar.csts);
+    uint32_t pmrsts = ldl_le_p(&n->bar.pmrsts);
+
     if (unlikely(offset & (sizeof(uint32_t) - 1))) {
         NVME_GUEST_ERR(pci_nvme_ub_mmiowr_misaligned32,
                        "MMIO write not 32-bit aligned,"
@@ -5747,9 +5762,10 @@  static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
                            " when MSI-X is enabled");
             /* should be ignored, fall through for now */
         }
-        n->bar.intms |= data & 0xffffffff;
+        intms |= data;
+        stl_le_p(&n->bar.intms, intms);
         n->bar.intmc = n->bar.intms;
-        trace_pci_nvme_mmio_intm_set(data & 0xffffffff, n->bar.intmc);
+        trace_pci_nvme_mmio_intm_set(data & 0xffffffff, intms);
         nvme_irq_check(n);
         break;
     case NVME_REG_INTMC:
@@ -5759,44 +5775,55 @@  static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
                            " when MSI-X is enabled");
             /* should be ignored, fall through for now */
         }
-        n->bar.intms &= ~(data & 0xffffffff);
+        intms &= ~data;
+        stl_le_p(&n->bar.intms, intms);
         n->bar.intmc = n->bar.intms;
-        trace_pci_nvme_mmio_intm_clr(data & 0xffffffff, n->bar.intmc);
+        trace_pci_nvme_mmio_intm_clr(data & 0xffffffff, intms);
         nvme_irq_check(n);
         break;
     case NVME_REG_CC:
         trace_pci_nvme_mmio_cfg(data & 0xffffffff);
+
         /* Windows first sends data, then sends enable bit */
-        if (!NVME_CC_EN(data) && !NVME_CC_EN(n->bar.cc) &&
-            !NVME_CC_SHN(data) && !NVME_CC_SHN(n->bar.cc))
+        if (!NVME_CC_EN(data) && !NVME_CC_EN(cc) &&
+            !NVME_CC_SHN(data) && !NVME_CC_SHN(cc))
         {
-            n->bar.cc = data;
+            cc = data;
         }
 
-        if (NVME_CC_EN(data) && !NVME_CC_EN(n->bar.cc)) {
-            n->bar.cc = data;
+        if (NVME_CC_EN(data) && !NVME_CC_EN(cc)) {
+            cc = data;
+
+            /* flush CC since nvme_start_ctrl() needs the value */
+            stl_le_p(&n->bar.cc, cc);
             if (unlikely(nvme_start_ctrl(n))) {
                 trace_pci_nvme_err_startfail();
-                n->bar.csts = NVME_CSTS_FAILED;
+                csts = NVME_CSTS_FAILED;
             } else {
                 trace_pci_nvme_mmio_start_success();
-                n->bar.csts = NVME_CSTS_READY;
+                csts = NVME_CSTS_READY;
             }
-        } else if (!NVME_CC_EN(data) && NVME_CC_EN(n->bar.cc)) {
+        } else if (!NVME_CC_EN(data) && NVME_CC_EN(cc)) {
             trace_pci_nvme_mmio_stopped();
             nvme_ctrl_reset(n);
-            n->bar.csts &= ~NVME_CSTS_READY;
+            cc = 0;
+            csts &= ~NVME_CSTS_READY;
         }
-        if (NVME_CC_SHN(data) && !(NVME_CC_SHN(n->bar.cc))) {
+
+        if (NVME_CC_SHN(data) && !(NVME_CC_SHN(cc))) {
             trace_pci_nvme_mmio_shutdown_set();
             nvme_ctrl_shutdown(n);
-            n->bar.cc = data;
-            n->bar.csts |= NVME_CSTS_SHST_COMPLETE;
-        } else if (!NVME_CC_SHN(data) && NVME_CC_SHN(n->bar.cc)) {
+            cc = data;
+            csts |= NVME_CSTS_SHST_COMPLETE;
+        } else if (!NVME_CC_SHN(data) && NVME_CC_SHN(cc)) {
             trace_pci_nvme_mmio_shutdown_cleared();
-            n->bar.csts &= ~NVME_CSTS_SHST_COMPLETE;
-            n->bar.cc = data;
+            csts &= ~NVME_CSTS_SHST_COMPLETE;
+            cc = data;
         }
+
+        stl_le_p(&n->bar.cc, cc);
+        stl_le_p(&n->bar.csts, csts);
+
         break;
     case NVME_REG_CSTS:
         if (data & (1 << 4)) {
@@ -5818,26 +5845,24 @@  static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
         }
         break;
     case NVME_REG_AQA:
-        n->bar.aqa = data & 0xffffffff;
+        stl_le_p(&n->bar.aqa, data);
         trace_pci_nvme_mmio_aqattr(data & 0xffffffff);
         break;
     case NVME_REG_ASQ:
-        n->bar.asq = size == 8 ? data :
-            (n->bar.asq & ~0xffffffffULL) | (data & 0xffffffff);
+        stn_le_p(&n->bar.asq, size, data);
         trace_pci_nvme_mmio_asqaddr(data);
         break;
     case NVME_REG_ASQ + 4:
-        n->bar.asq = (n->bar.asq & 0xffffffff) | (data << 32);
-        trace_pci_nvme_mmio_asqaddr_hi(data, n->bar.asq);
+        stl_le_p((uint8_t *)&n->bar.asq + 4, data);
+        trace_pci_nvme_mmio_asqaddr_hi(data, ldq_le_p(&n->bar.asq));
         break;
     case NVME_REG_ACQ:
         trace_pci_nvme_mmio_acqaddr(data);
-        n->bar.acq = size == 8 ? data :
-            (n->bar.acq & ~0xffffffffULL) | (data & 0xffffffff);
+        stn_le_p(&n->bar.acq, size, data);
         break;
     case NVME_REG_ACQ + 4:
-        n->bar.acq = (n->bar.acq & 0xffffffff) | (data << 32);
-        trace_pci_nvme_mmio_acqaddr_hi(data, n->bar.acq);
+        stl_le_p((uint8_t *)&n->bar.acq + 4, data);
+        trace_pci_nvme_mmio_acqaddr_hi(data, ldq_le_p(&n->bar.acq));
         break;
     case NVME_REG_CMBLOC:
         NVME_GUEST_ERR(pci_nvme_ub_mmiowr_cmbloc_reserved,
@@ -5849,21 +5874,23 @@  static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
                        "invalid write to read only CMBSZ, ignored");
         return;
     case NVME_REG_CMBMSC:
-        if (!NVME_CAP_CMBS(n->bar.cap)) {
+        if (!NVME_CAP_CMBS(cap)) {
             return;
         }
 
-        n->bar.cmbmsc = size == 8 ? data :
-            (n->bar.cmbmsc & ~0xffffffff) | (data & 0xffffffff);
+        stn_le_p(&n->bar.cmbmsc, size, data);
         n->cmb.cmse = false;
 
         if (NVME_CMBMSC_CRE(data)) {
             nvme_cmb_enable_regs(n);
 
             if (NVME_CMBMSC_CMSE(data)) {
-                hwaddr cba = NVME_CMBMSC_CBA(data) << CMBMSC_CBA_SHIFT;
+                uint64_t cmbmsc = ldq_le_p(&n->bar.cmbmsc);
+                hwaddr cba = NVME_CMBMSC_CBA(cmbmsc) << CMBMSC_CBA_SHIFT;
                 if (cba + int128_get64(n->cmb.mem.size) < cba) {
-                    NVME_CMBSTS_SET_CBAI(n->bar.cmbsts, 1);
+                    uint32_t cmbsts = ldl_le_p(&n->bar.cmbsts);
+                    NVME_CMBSTS_SET_CBAI(cmbsts, 1);
+                    stl_le_p(&n->bar.cmbsts, cmbsts);
                     return;
                 }
 
@@ -5877,7 +5904,7 @@  static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
 
         return;
     case NVME_REG_CMBMSC + 4:
-        n->bar.cmbmsc = (n->bar.cmbmsc & 0xffffffff) | (data << 32);
+        stl_le_p((uint8_t *)&n->bar.cmbmsc + 4, data);
         return;
 
     case NVME_REG_PMRCAP:
@@ -5885,19 +5912,20 @@  static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
                        "invalid write to PMRCAP register, ignored");
         return;
     case NVME_REG_PMRCTL:
-        if (!NVME_CAP_PMRS(n->bar.cap)) {
+        if (!NVME_CAP_PMRS(cap)) {
             return;
         }
 
-        n->bar.pmrctl = data;
+        stl_le_p(&n->bar.pmrctl, data);
         if (NVME_PMRCTL_EN(data)) {
             memory_region_set_enabled(&n->pmr.dev->mr, true);
-            n->bar.pmrsts = 0;
+            pmrsts = 0;
         } else {
             memory_region_set_enabled(&n->pmr.dev->mr, false);
-            NVME_PMRSTS_SET_NRDY(n->bar.pmrsts, 1);
+            NVME_PMRSTS_SET_NRDY(pmrsts, 1);
             n->pmr.cmse = false;
         }
+        stl_le_p(&n->bar.pmrsts, pmrsts);
         return;
     case NVME_REG_PMRSTS:
         NVME_GUEST_ERR(pci_nvme_ub_mmiowr_pmrsts_readonly,
@@ -5912,18 +5940,20 @@  static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
                        "invalid write to PMRSWTP register, ignored");
         return;
     case NVME_REG_PMRMSCL:
-        if (!NVME_CAP_PMRS(n->bar.cap)) {
+        if (!NVME_CAP_PMRS(cap)) {
             return;
         }
 
-        n->bar.pmrmscl = data & 0xffffffff;
+        stl_le_p(&n->bar.pmrmscl, data);
         n->pmr.cmse = false;
 
-        if (NVME_PMRMSCL_CMSE(n->bar.pmrmscl)) {
-            hwaddr cba = n->bar.pmrmscu |
-                (NVME_PMRMSCL_CBA(n->bar.pmrmscl) << PMRMSCL_CBA_SHIFT);
+        if (NVME_PMRMSCL_CMSE(data)) {
+            uint64_t pmrmscu = ldl_le_p(&n->bar.pmrmscu);
+            hwaddr cba = pmrmscu << 32 |
+                (NVME_PMRMSCL_CBA(data) << PMRMSCL_CBA_SHIFT);
             if (cba + int128_get64(n->pmr.dev->mr.size) < cba) {
-                NVME_PMRSTS_SET_CBAI(n->bar.pmrsts, 1);
+                NVME_PMRSTS_SET_CBAI(pmrsts, 1);
+                stl_le_p(&n->bar.pmrsts, pmrsts);
                 return;
             }
 
@@ -5933,11 +5963,11 @@  static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
 
         return;
     case NVME_REG_PMRMSCU:
-        if (!NVME_CAP_PMRS(n->bar.cap)) {
+        if (!NVME_CAP_PMRS(cap)) {
             return;
         }
 
-        n->bar.pmrmscu = data & 0xffffffff;
+        stl_le_p(&n->bar.pmrmscu, data);
         return;
     default:
         NVME_GUEST_ERR(pci_nvme_ub_mmiowr_invalid,
@@ -5952,7 +5982,6 @@  static uint64_t nvme_mmio_read(void *opaque, hwaddr addr, unsigned size)
 {
     NvmeCtrl *n = (NvmeCtrl *)opaque;
     uint8_t *ptr = (uint8_t *)&n->bar;
-    uint64_t val = 0;
 
     trace_pci_nvme_mmio_read(addr, size);
 
@@ -5982,13 +6011,11 @@  static uint64_t nvme_mmio_read(void *opaque, hwaddr addr, unsigned size)
      * made it to persistent media
      */
     if (addr == NVME_REG_PMRSTS &&
-        (NVME_PMRCAP_PMRWBM(n->bar.pmrcap) & 0x02)) {
+        (NVME_PMRCAP_PMRWBM(ldl_le_p(&n->bar.pmrcap)) & 0x02)) {
         memory_region_msync(&n->pmr.dev->mr, 0, n->pmr.dev->size);
     }
 
-    memcpy(&val, ptr + addr, size);
-
-    return val;
+    return ldn_le_p(ptr + addr, size);
 }
 
 static void nvme_process_db(NvmeCtrl *n, hwaddr addr, int val)
@@ -6246,6 +6273,7 @@  static void nvme_init_state(NvmeCtrl *n)
 static void nvme_init_cmb(NvmeCtrl *n, PCIDevice *pci_dev)
 {
     uint64_t cmb_size = n->params.cmb_size_mb * MiB;
+    uint64_t cap = ldq_le_p(&n->bar.cap);
 
     n->cmb.buf = g_malloc0(cmb_size);
     memory_region_init_io(&n->cmb.mem, OBJECT(n), &nvme_cmb_ops, n,
@@ -6255,7 +6283,8 @@  static void nvme_init_cmb(NvmeCtrl *n, PCIDevice *pci_dev)
                      PCI_BASE_ADDRESS_MEM_TYPE_64 |
                      PCI_BASE_ADDRESS_MEM_PREFETCH, &n->cmb.mem);
 
-    NVME_CAP_SET_CMBS(n->bar.cap, 1);
+    NVME_CAP_SET_CMBS(cap, 1);
+    stq_le_p(&n->bar.cap, cap);
 
     if (n->params.legacy_cmb) {
         nvme_cmb_enable_regs(n);
@@ -6265,14 +6294,17 @@  static void nvme_init_cmb(NvmeCtrl *n, PCIDevice *pci_dev)
 
 static void nvme_init_pmr(NvmeCtrl *n, PCIDevice *pci_dev)
 {
-    NVME_PMRCAP_SET_RDS(n->bar.pmrcap, 1);
-    NVME_PMRCAP_SET_WDS(n->bar.pmrcap, 1);
-    NVME_PMRCAP_SET_BIR(n->bar.pmrcap, NVME_PMR_BIR);
-    /* Turn on bit 1 support */
-    NVME_PMRCAP_SET_PMRWBM(n->bar.pmrcap, 0x02);
-    NVME_PMRCAP_SET_CMSS(n->bar.pmrcap, 1);
+    uint32_t pmrcap = ldl_le_p(&n->bar.pmrcap);
 
-    pci_register_bar(pci_dev, NVME_PMRCAP_BIR(n->bar.pmrcap),
+    NVME_PMRCAP_SET_RDS(pmrcap, 1);
+    NVME_PMRCAP_SET_WDS(pmrcap, 1);
+    NVME_PMRCAP_SET_BIR(pmrcap, NVME_PMR_BIR);
+    /* Turn on bit 1 support */
+    NVME_PMRCAP_SET_PMRWBM(pmrcap, 0x02);
+    NVME_PMRCAP_SET_CMSS(pmrcap, 1);
+    stl_le_p(&n->bar.pmrcap, pmrcap);
+
+    pci_register_bar(pci_dev, NVME_PMR_BIR,
                      PCI_BASE_ADDRESS_SPACE_MEMORY |
                      PCI_BASE_ADDRESS_MEM_TYPE_64 |
                      PCI_BASE_ADDRESS_MEM_PREFETCH, &n->pmr.dev->mr);
@@ -6362,6 +6394,7 @@  static void nvme_init_ctrl(NvmeCtrl *n, PCIDevice *pci_dev)
 {
     NvmeIdCtrl *id = &n->id_ctrl;
     uint8_t *pci_conf = pci_dev->config;
+    uint64_t cap = ldq_le_p(&n->bar.cap);
 
     id->vid = cpu_to_le16(pci_get_word(pci_conf + PCI_VENDOR_ID));
     id->ssvid = cpu_to_le16(pci_get_word(pci_conf + PCI_SUBSYSTEM_VENDOR_ID));
@@ -6440,17 +6473,18 @@  static void nvme_init_ctrl(NvmeCtrl *n, PCIDevice *pci_dev)
         id->cmic |= NVME_CMIC_MULTI_CTRL;
     }
 
-    NVME_CAP_SET_MQES(n->bar.cap, 0x7ff);
-    NVME_CAP_SET_CQR(n->bar.cap, 1);
-    NVME_CAP_SET_TO(n->bar.cap, 0xf);
-    NVME_CAP_SET_CSS(n->bar.cap, NVME_CAP_CSS_NVM);
-    NVME_CAP_SET_CSS(n->bar.cap, NVME_CAP_CSS_CSI_SUPP);
-    NVME_CAP_SET_CSS(n->bar.cap, NVME_CAP_CSS_ADMIN_ONLY);
-    NVME_CAP_SET_MPSMAX(n->bar.cap, 4);
-    NVME_CAP_SET_CMBS(n->bar.cap, n->params.cmb_size_mb ? 1 : 0);
-    NVME_CAP_SET_PMRS(n->bar.cap, n->pmr.dev ? 1 : 0);
+    NVME_CAP_SET_MQES(cap, 0x7ff);
+    NVME_CAP_SET_CQR(cap, 1);
+    NVME_CAP_SET_TO(cap, 0xf);
+    NVME_CAP_SET_CSS(cap, NVME_CAP_CSS_NVM);
+    NVME_CAP_SET_CSS(cap, NVME_CAP_CSS_CSI_SUPP);
+    NVME_CAP_SET_CSS(cap, NVME_CAP_CSS_ADMIN_ONLY);
+    NVME_CAP_SET_MPSMAX(cap, 4);
+    NVME_CAP_SET_CMBS(cap, n->params.cmb_size_mb ? 1 : 0);
+    NVME_CAP_SET_PMRS(cap, n->pmr.dev ? 1 : 0);
+    stq_le_p(&n->bar.cap, cap);
 
-    n->bar.vs = NVME_SPEC_VER;
+    stl_le_p(&n->bar.vs, NVME_SPEC_VER);
     n->bar.intmc = n->bar.intms = 0;
 }
 
@@ -6601,7 +6635,7 @@  static void nvme_set_smart_warning(Object *obj, Visitor *v, const char *name,
 
     cap = NVME_SMART_SPARE | NVME_SMART_TEMPERATURE | NVME_SMART_RELIABILITY
           | NVME_SMART_MEDIA_READ_ONLY | NVME_SMART_FAILED_VOLATILE_MEDIA;
-    if (NVME_CAP_PMRS(n->bar.cap)) {
+    if (NVME_CAP_PMRS(ldq_le_p(&n->bar.cap))) {
         cap |= NVME_SMART_PMR_UNRELIABLE;
     }