Skip to contentSkip to conversation

real host page API

PATCH In review Peter Crosthwaite 5 participants 7 messages Jun 6, 2015 – Jun 25, 2015

lore
Share & tools
lore

Peter Crosthwaite

author
#0 lore raw

cover real host page API

Hi All,

As we discussed earlier, the sole use of HOST_PAGE_ALIGN in system level code should be targeting the "real" host page alignment and not the target-specific HOST_PAGE_ALIGN.

This complete the API for REAL_HOST_PAGE_ALIGN and friends that can be use the same way as HOST_PAGE_ALIGN in system level code.

The one use in vfio/pci is then patches.

Regards,
Peter

Peter Crosthwaite (2):
cpu-all: complete "real" host page size API
vfio: cpu: Use "real" page size API

 hw/vfio/pci.c          | 7 ++++---
 include/exec/cpu-all.h | 3 +++
 translate-all.c        | 2 ++
 3 files changed, 9 insertions(+), 3 deletions(-)
Signature · 1 line
1.9.1
author
#1 lore raw

PATCH 1/2 cpu-all: complete "real" host page size API

Currently the "host" page size alignment API is really aligning to both host and target page sizes. There is the qemu_real_page_size which can be used for the actual host page size but it's missing a mask and ALIGN macro as provided for qemu_page_size. Complete the API. This allows system level code that cares about the host page size to use a consistent alignment interface without having to un-needingly align to the target page size. This also reduces system level code dependency on the cpu specific TARGET_PAGE_SIZE.

Signed-off-by Peter Crosthwaite <[email protected]>
---
 include/exec/cpu-all.h | 3 +++
 translate-all.c        | 2 ++
 2 files changed, 5 insertions(+)
include/exec/cpu-all.h +3 -0
index ac06c67..8f89669 100644
@@ -177,10 +177,13 @@ extern unsigned long reserved_va;
 
 /* ??? These should be the larger of uintptr_t and target_ulong.  */
 extern uintptr_t qemu_real_host_page_size;
+extern uintptr_t qemu_real_host_page_mask;
 extern uintptr_t qemu_host_page_size;
 extern uintptr_t qemu_host_page_mask;
 
 #define HOST_PAGE_ALIGN(addr) (((addr) + qemu_host_page_size - 1) & qemu_host_page_mask)
+#define REAL_HOST_PAGE_ALIGN(addr) (((addr) + qemu_real_host_page_size - 1) & \
+                                    qemu_real_host_page_mask)
 
 /* same as PROT_xxx */
 #define PAGE_READ      0x0001
translate-all.c +2 -0
index 62042af..d5a1977 100644
@@ -117,6 +117,7 @@ typedef struct PageDesc {
 #define V_L1_SHIFT (L1_MAP_ADDR_SPACE_BITS - TARGET_PAGE_BITS - V_L1_BITS)
 
 uintptr_t qemu_real_host_page_size;
+uintptr_t qemu_real_host_page_mask;
 uintptr_t qemu_host_page_size;
 uintptr_t qemu_host_page_mask;
 

@@ -306,6 +307,7 @@ void page_size_init(void)
     /* NOTE: we can always suppose that qemu_host_page_size >=
        TARGET_PAGE_SIZE */
     qemu_real_host_page_size = getpagesize();
+    qemu_real_host_page_mask = ~(qemu_real_host_page_size - 1);
     if (qemu_host_page_size == 0) {
         qemu_host_page_size = qemu_real_host_page_size;
     }
Signature · 1 line
1.9.1
author
#2 lore raw

PATCH 2/2 vfio: cpu: Use "real" page size API

In reply to Peter Crosthwaite

This is system level code, and should only depend on the host page size, not the target page size.

Note that HOST_PAGE_SIZE is misleadingly lead and is really aligning to both host and target page size. Hence it's replacement with REAL_HOST_PAGE_SIZE.

Signed-off-by Peter Crosthwaite <[email protected]>
---
 hw/vfio/pci.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
hw/vfio/pci.c +4 -3
index e0e339a..cf9791e 100644
@@ -2388,7 +2388,7 @@ static void vfio_map_bar(VFIOPCIDevice *vdev, int nr)
      * potentially insert a direct-mapped subregion before and after it.
      */
     if (vdev->msix && vdev->msix->table_bar == nr) {
-        size = vdev->msix->table_offset & qemu_host_page_mask;
+        size = vdev->msix->table_offset & qemu_real_host_page_mask;
     }
 
     strncat(name, " mmap", sizeof(name) - strlen(name) - 1);

@@ -2401,8 +2401,9 @@ static void vfio_map_bar(VFIOPCIDevice *vdev, int nr)
     if (vdev->msix && vdev->msix->table_bar == nr) {
         uint64_t start;
 
-        start = HOST_PAGE_ALIGN((uint64_t)vdev->msix->table_offset +
-                                (vdev->msix->entries * PCI_MSIX_ENTRY_SIZE));
+        start = REAL_HOST_PAGE_ALIGN((uint64_t)vdev->msix->table_offset +
+                                     (vdev->msix->entries *
+                                      PCI_MSIX_ENTRY_SIZE));
 
         size = start < bar->region.size ? bar->region.size - start : 0;
         strncat(name, " msix-hi", sizeof(name) - strlen(name) - 1);
Signature · 1 line
1.9.1
Tested-by on the v1 series
#3 lore raw
In reply to Peter Crosthwaite
Quoted 10 lines · Peter Crosthwaite

Hi All,

As we discussed earlier, the sole use of HOST_PAGE_ALIGN in system level code should be targeting the "real" host page alignment and not the target-specific HOST_PAGE_ALIGN.

This complete the API for REAL_HOST_PAGE_ALIGN and friends that can be use the same way as HOST_PAGE_ALIGN in system level code.

The one use in vfio/pci is then patches.

Looks good to me.

Tested-by Alexey Kardashevskiy <[email protected]>
Quoted 13 lines

Regards,
Peter

Peter Crosthwaite (2):
cpu-all: complete "real" host page size API
vfio: cpu: Use "real" page size API

  hw/vfio/pci.c          | 7 ++++---
  include/exec/cpu-all.h | 3 +++
  translate-all.c        | 2 ++
  3 files changed, 9 insertions(+), 3 deletions(-)
Signature · 1 line
Alexey
#4 lore raw
In reply to Peter Crosthwaite

Ping!

Can this go via the TCG queue, is there something for VFIO, or can Peter pick it up as a between-the-maintainerships straggler?

Regards,
Peter

Quoted 27 lines · Peter Crosthwaite

Hi All,

As we discussed earlier, the sole use of HOST_PAGE_ALIGN in system level code should be targeting the "real" host page alignment and not the target-specific HOST_PAGE_ALIGN.

This complete the API for REAL_HOST_PAGE_ALIGN and friends that can be use the same way as HOST_PAGE_ALIGN in system level code.

The one use in vfio/pci is then patches.

Regards,
Peter

Peter Crosthwaite (2):
cpu-all: complete "real" host page size API
vfio: cpu: Use "real" page size API

 hw/vfio/pci.c          | 7 ++++---
 include/exec/cpu-all.h | 3 +++
 translate-all.c        | 2 ++
 3 files changed, 9 insertions(+), 3 deletions(-)

--
1.9.1

Acked-by
#5 lore raw
Quoted 37 lines · Peter Crosthwaite

Ping!

Can this go via the TCG queue, is there something for VFIO, or can Peter pick it up as a between-the-maintainerships straggler?

Regards,
Peter

Quoted 27 lines · Peter Crosthwaite

Hi All,

As we discussed earlier, the sole use of HOST_PAGE_ALIGN in system level code should be targeting the "real" host page alignment and not the target-specific HOST_PAGE_ALIGN.

This complete the API for REAL_HOST_PAGE_ALIGN and friends that can be use the same way as HOST_PAGE_ALIGN in system level code.

The one use in vfio/pci is then patches.

Regards,
Peter

Peter Crosthwaite (2):
cpu-all: complete "real" host page size API
vfio: cpu: Use "real" page size API

 hw/vfio/pci.c          | 7 ++++---
 include/exec/cpu-all.h | 3 +++
 translate-all.c        | 2 ++
 3 files changed, 9 insertions(+), 3 deletions(-)

--
1.9.1

Acked-by Paolo Bonzini <[email protected]>

Alex, can you pick it up or shall I?

Paolo

#6 lore raw
Quoted 35 lines · Paolo Bonzini
Quoted 33 lines · Peter Crosthwaite

Ping!

Can this go via the TCG queue, is there something for VFIO, or can Peter pick it up as a between-the-maintainerships straggler?

Regards,
Peter

Quoted 23 lines · Peter Crosthwaite

Hi All,

As we discussed earlier, the sole use of HOST_PAGE_ALIGN in system level code should be targeting the "real" host page alignment and not the target-specific HOST_PAGE_ALIGN.

This complete the API for REAL_HOST_PAGE_ALIGN and friends that can be use the same way as HOST_PAGE_ALIGN in system level code.

The one use in vfio/pci is then patches.

Regards,
Peter

Peter Crosthwaite (2):
cpu-all: complete "real" host page size API
vfio: cpu: Use "real" page size API

 hw/vfio/pci.c          | 7 ++++---
 include/exec/cpu-all.h | 3 +++
 translate-all.c        | 2 ++
 3 files changed, 9 insertions(+), 3 deletions(-)
Quoted 8 lines
Quoted 4 lines
Quoted 4 lines

--
1.9.1

Acked-by Paolo Bonzini <[email protected]>

Alex, can you pick it up or shall I?

I'll take it, I need to send a pull request for your coverity fix before hard freeze anyway. Thanks,

Alex