[Qemu-devel] [PATCH v4 02/15] memory: Access MemoryRegion with MemOp

tony.nguyen@bt.com posted 15 patches 6 years, 3 months ago
Maintainers: "Michael S. Tsirkin" <mst@redhat.com>, Christian Borntraeger <borntraeger@de.ibm.com>, David Gibson <david@gibson.dropbear.id.au>, Eduardo Habkost <ehabkost@redhat.com>, Peter Maydell <peter.maydell@linaro.org>, Bastian Koppelmann <kbastian@mail.uni-paderborn.de>, Aleksandar Rikalo <arikalo@wavecomp.com>, Claudio Fontana <claudio.fontana@huawei.com>, "Edgar E. Iglesias" <edgar.iglesias@gmail.com>, Sagar Karandikar <sagark@eecs.berkeley.edu>, Laurent Vivier <laurent@vivier.eu>, Aurelien Jarno <aurelien@aurel32.net>, Alex Williamson <alex.williamson@redhat.com>, Alistair Francis <Alistair.Francis@wdc.com>, Richard Henderson <rth@twiddle.net>, Aleksandar Markovic <amarkovic@wavecomp.com>, Stafford Horne <shorne@gmail.com>, Cornelia Huck <cohuck@redhat.com>, Palmer Dabbelt <palmer@sifive.com>, Artyom Tarasenko <atar4qemu@gmail.com>, Halil Pasic <pasic@linux.ibm.com>, Collin Walling <walling@linux.ibm.com>, Andrzej Zaborowski <balrogg@gmail.com>, Stefan Hajnoczi <stefanha@redhat.com>, Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>, David Hildenbrand <david@redhat.com>, Paolo Bonzini <pbonzini@redhat.com>
There is a newer version of this series
[Qemu-devel] [PATCH v4 02/15] memory: Access MemoryRegion with MemOp
Posted by tony.nguyen@bt.com 6 years, 3 months ago
Replacing size with size+sign+endianness (MemOp) will enable us to
collapse the two byte swaps, adjust_endianness and handle_bswap, along
the I/O path.

While interfaces are converted, callers will have existing unsigned
size coerced into a MemOp, and the callee will use this MemOp as an
unsigned size.

Signed-off-by: Tony Nguyen <tony.nguyen@bt.com>
---
 include/exec/memop.h  | 4 ++++
 include/exec/memory.h | 9 +++++----
 memory.c              | 7 +++++--
 3 files changed, 14 insertions(+), 6 deletions(-)

diff --git a/include/exec/memop.h b/include/exec/memop.h
index ac58066..09c8d20 100644
--- a/include/exec/memop.h
+++ b/include/exec/memop.h
@@ -106,4 +106,8 @@ typedef enum MemOp {
     MO_SSIZE = MO_SIZE | MO_SIGN,
 } MemOp;

+/* No-op while memory_region_dispatch_[read|write] is converted to MemOp */
+#define MEMOP_SIZE(op)  (op)    /* MemOp to size.  */
+#define SIZE_MEMOP(ul)  (ul)    /* Size to MemOp.  */
+
 #endif
diff --git a/include/exec/memory.h b/include/exec/memory.h
index bb0961d..30b1c58 100644
--- a/include/exec/memory.h
+++ b/include/exec/memory.h
@@ -19,6 +19,7 @@
 #include "exec/cpu-common.h"
 #include "exec/hwaddr.h"
 #include "exec/memattrs.h"
+#include "exec/memop.h"
 #include "exec/ramlist.h"
 #include "qemu/queue.h"
 #include "qemu/int128.h"
@@ -1731,13 +1732,13 @@ void mtree_info(bool flatview, bool dispatch_tree, bool owner);
  * @mr: #MemoryRegion to access
  * @addr: address within that region
  * @pval: pointer to uint64_t which the data is written to
- * @size: size of the access in bytes
+ * @op: encodes size of the access in bytes
  * @attrs: memory transaction attributes to use for the access
  */
 MemTxResult memory_region_dispatch_read(MemoryRegion *mr,
                                         hwaddr addr,
                                         uint64_t *pval,
-                                        unsigned size,
+                                        MemOp op,
                                         MemTxAttrs attrs);
 /**
  * memory_region_dispatch_write: perform a write directly to the specified
@@ -1746,13 +1747,13 @@ MemTxResult memory_region_dispatch_read(MemoryRegion *mr,
  * @mr: #MemoryRegion to access
  * @addr: address within that region
  * @data: data to write
- * @size: size of the access in bytes
+ * @op: encodes size of the access in bytes
  * @attrs: memory transaction attributes to use for the access
  */
 MemTxResult memory_region_dispatch_write(MemoryRegion *mr,
                                          hwaddr addr,
                                          uint64_t data,
-                                         unsigned size,
+                                         MemOp op,
                                          MemTxAttrs attrs);

 /**
diff --git a/memory.c b/memory.c
index 5d8c9a9..6982e19 100644
--- a/memory.c
+++ b/memory.c
@@ -1439,10 +1439,11 @@ static MemTxResult memory_region_dispatch_read1(MemoryRegion *mr,
 MemTxResult memory_region_dispatch_read(MemoryRegion *mr,
                                         hwaddr addr,
                                         uint64_t *pval,
-                                        unsigned size,
+                                        MemOp op,
                                         MemTxAttrs attrs)
 {
     MemTxResult r;
+    unsigned size = MEMOP_SIZE(op);

     if (!memory_region_access_valid(mr, addr, size, false, attrs)) {
         *pval = unassigned_mem_read(mr, addr, size);
@@ -1483,9 +1484,11 @@ static bool memory_region_dispatch_write_eventfds(MemoryRegion *mr,
 MemTxResult memory_region_dispatch_write(MemoryRegion *mr,
                                          hwaddr addr,
                                          uint64_t data,
-                                         unsigned size,
+                                         MemOp op,
                                          MemTxAttrs attrs)
 {
+    unsigned size = MEMOP_SIZE(op);
+
     if (!memory_region_access_valid(mr, addr, size, true, attrs)) {
         unassigned_mem_write(mr, addr, data, size);
         return MEMTX_DECODE_ERROR;
--
1.8.3.1



Re: [Qemu-devel] [PATCH v4 02/15] memory: Access MemoryRegion with MemOp
Posted by Philippe Mathieu-Daudé 6 years, 3 months ago
On 7/25/19 11:52 AM, tony.nguyen@bt.com wrote:
> Replacing size with size+sign+endianness (MemOp) will enable us to
> collapse the two byte swaps, adjust_endianness and handle_bswap, along
> the I/O path.
> 
> While interfaces are converted, callers will have existing unsigned
> size coerced into a MemOp, and the callee will use this MemOp as an
> unsigned size.
> 
> Signed-off-by: Tony Nguyen <tony.nguyen@bt.com>
> ---
>  include/exec/memop.h  | 4 ++++
>  include/exec/memory.h | 9 +++++----
>  memory.c              | 7 +++++--
>  3 files changed, 14 insertions(+), 6 deletions(-)
> 
> diff --git a/include/exec/memop.h b/include/exec/memop.h
> index ac58066..09c8d20 100644
> --- a/include/exec/memop.h
> +++ b/include/exec/memop.h
> @@ -106,4 +106,8 @@ typedef enum MemOp {
>      MO_SSIZE = MO_SIZE | MO_SIGN,
>  } MemOp;
> 
> +/* No-op while memory_region_dispatch_[read|write] is converted to MemOp */
> +#define MEMOP_SIZE(op)  (op)    /* MemOp to size.  */
> +#define SIZE_MEMOP(ul)  (ul)    /* Size to MemOp.  */

SIZE_MEMOP() is never used until patch #10 "memory: Access MemoryRegion
with MemOp semantics", it would be clearer to only introduce the
MEMOP_SIZE() no-op here, and directly introduce the correct SIZE_MEMOP()
macro in patch #10.

> +
>  #endif
> diff --git a/include/exec/memory.h b/include/exec/memory.h
> index bb0961d..30b1c58 100644
> --- a/include/exec/memory.h
> +++ b/include/exec/memory.h
> @@ -19,6 +19,7 @@
>  #include "exec/cpu-common.h"
>  #include "exec/hwaddr.h"
>  #include "exec/memattrs.h"
> +#include "exec/memop.h"
>  #include "exec/ramlist.h"
>  #include "qemu/queue.h"
>  #include "qemu/int128.h"
> @@ -1731,13 +1732,13 @@ void mtree_info(bool flatview, bool dispatch_tree, bool owner);
>   * @mr: #MemoryRegion to access
>   * @addr: address within that region
>   * @pval: pointer to uint64_t which the data is written to
> - * @size: size of the access in bytes
> + * @op: encodes size of the access in bytes
>   * @attrs: memory transaction attributes to use for the access
>   */
>  MemTxResult memory_region_dispatch_read(MemoryRegion *mr,
>                                          hwaddr addr,
>                                          uint64_t *pval,
> -                                        unsigned size,
> +                                        MemOp op,
>                                          MemTxAttrs attrs);
>  /**
>   * memory_region_dispatch_write: perform a write directly to the specified
> @@ -1746,13 +1747,13 @@ MemTxResult memory_region_dispatch_read(MemoryRegion *mr,
>   * @mr: #MemoryRegion to access
>   * @addr: address within that region
>   * @data: data to write
> - * @size: size of the access in bytes
> + * @op: encodes size of the access in bytes
>   * @attrs: memory transaction attributes to use for the access
>   */
>  MemTxResult memory_region_dispatch_write(MemoryRegion *mr,
>                                           hwaddr addr,
>                                           uint64_t data,
> -                                         unsigned size,
> +                                         MemOp op,
>                                           MemTxAttrs attrs);
> 
>  /**
> diff --git a/memory.c b/memory.c
> index 5d8c9a9..6982e19 100644
> --- a/memory.c
> +++ b/memory.c
> @@ -1439,10 +1439,11 @@ static MemTxResult memory_region_dispatch_read1(MemoryRegion *mr,
>  MemTxResult memory_region_dispatch_read(MemoryRegion *mr,
>                                          hwaddr addr,
>                                          uint64_t *pval,
> -                                        unsigned size,
> +                                        MemOp op,
>                                          MemTxAttrs attrs)
>  {
>      MemTxResult r;
> +    unsigned size = MEMOP_SIZE(op);
> 
>      if (!memory_region_access_valid(mr, addr, size, false, attrs)) {
>          *pval = unassigned_mem_read(mr, addr, size);
> @@ -1483,9 +1484,11 @@ static bool memory_region_dispatch_write_eventfds(MemoryRegion *mr,
>  MemTxResult memory_region_dispatch_write(MemoryRegion *mr,
>                                           hwaddr addr,
>                                           uint64_t data,
> -                                         unsigned size,
> +                                         MemOp op,
>                                           MemTxAttrs attrs)
>  {
> +    unsigned size = MEMOP_SIZE(op);
> +
>      if (!memory_region_access_valid(mr, addr, size, true, attrs)) {
>          unassigned_mem_write(mr, addr, data, size);
>          return MEMTX_DECODE_ERROR;
> --
> 1.8.3.1
> 
> 
> 

Re: [Qemu-devel] [PATCH v4 02/15] memory: Access MemoryRegion with MemOp
Posted by tony.nguyen@bt.com 6 years, 3 months ago
On 7/25/19 9:45 PM, Philippe Mathieu-Daudé wrote:
>On 7/25/19 11:52 AM, tony.nguyen@bt.com wrote:
>> Replacing size with size+sign+endianness (MemOp) will enable us to
>> collapse the two byte swaps, adjust_endianness and handle_bswap, along
>> the I/O path.
>>
>> While interfaces are converted, callers will have existing unsigned
>> size coerced into a MemOp, and the callee will use this MemOp as an
>> unsigned size.
>>
>> Signed-off-by: Tony Nguyen <tony.nguyen@bt.com>
>> ---
>>  include/exec/memop.h  | 4 ++++
>>  include/exec/memory.h | 9 +++++----
>>  memory.c              | 7 +++++--
>>  3 files changed, 14 insertions(+), 6 deletions(-)
>>
>> diff --git a/include/exec/memop.h b/include/exec/memop.h
>> index ac58066..09c8d20 100644
>> --- a/include/exec/memop.h
>> +++ b/include/exec/memop.h
>> @@ -106,4 +106,8 @@ typedef enum MemOp {
>>      MO_SSIZE = MO_SIZE | MO_SIGN,
>>  } MemOp;
>>
>> +/* No-op while memory_region_dispatch_[read|write] is converted to MemOp */
>> +#define MEMOP_SIZE(op)  (op)    /* MemOp to size.  */
>> +#define SIZE_MEMOP(ul)  (ul)    /* Size to MemOp.  */
>
>SIZE_MEMOP() is never used until patch #10 "memory: Access MemoryRegion
>with MemOp semantics", it would be clearer to only introduce the
>MEMOP_SIZE() no-op here, and directly introduce the correct SIZE_MEMOP()
>macro in patch #10.

SIZE_MEMOP() is used, and is the main change, in patches #3 to #10. Perhaps you
meant MEMOP_SIZE()?

Either way, you have raised an issue :)

There is a lack of clarity in how the two macros are used to update the
interfaces.?

Re: [Qemu-devel] [PATCH v4 02/15] memory: Access MemoryRegion with MemOp
Posted by Philippe Mathieu-Daudé 6 years, 3 months ago
On 7/26/19 8:03 AM, tony.nguyen@bt.com wrote:
> On 7/25/19 9:45 PM, Philippe Mathieu-Daudé wrote: 
>>On 7/25/19 11:52 AM, tony.nguyen@bt.com wrote:
>>> Replacing size with size+sign+endianness (MemOp) will enable us to
>>> collapse the two byte swaps, adjust_endianness and handle_bswap, along
>>> the I/O path.
>>> 
>>> While interfaces are converted, callers will have existing unsigned
>>> size coerced into a MemOp, and the callee will use this MemOp as an
>>> unsigned size.
>>> 
>>> Signed-off-by: Tony Nguyen <tony.nguyen@bt.com>
>>> ---
>>>  include/exec/memop.h  | 4 ++++
>>>  include/exec/memory.h | 9 +++++----
>>>  memory.c              | 7 +++++--
>>>  3 files changed, 14 insertions(+), 6 deletions(-)
>>> 
>>> diff --git a/include/exec/memop.h b/include/exec/memop.h
>>> index ac58066..09c8d20 100644
>>> --- a/include/exec/memop.h
>>> +++ b/include/exec/memop.h
>>> @@ -106,4 +106,8 @@ typedef enum MemOp {
>>>      MO_SSIZE = MO_SIZE | MO_SIGN,
>>>  } MemOp;
>>> 
>>> +/* No-op while memory_region_dispatch_[read|write] is converted to
> MemOp */
>>> +#define MEMOP_SIZE(op)  (op)    /* MemOp to size.  */
>>> +#define SIZE_MEMOP(ul)  (ul)    /* Size to MemOp.  */
>>
>>SIZE_MEMOP() is never used until patch #10 "memory: Access MemoryRegion
>>with MemOp semantics", it would be clearer to only introduce the
>>MEMOP_SIZE() no-op here, and directly introduce the correct SIZE_MEMOP()
>>macro in patch #10.
> 
> SIZE_MEMOP() is used, and is the main change, in patches #3 to #10.
> Perhaps you
> meant MEMOP_SIZE()?

Eh, I inverted the macro names... :( Thanks for correcting me.

> Either way, you have raised an issue :)
> 
> There is a lack of clarity in how the two macros are used to update the
> interfaces.​
> 
>