Patchwork Align PCI capabilities in pci_find_space

login
register
mail settings
Submitter mjr@cs.wisc.edu
Date Sept. 25, 2012, 8:59 p.m.
Message ID <1348606761-44739-1-git-send-email-mjr@cs.wisc.edu>
Download mbox | patch
Permalink /patch/186891/
State New
Headers show

Comments

mjr@cs.wisc.edu - Sept. 25, 2012, 8:59 p.m.
From: Matt Renzelmann <mjr@cs.wisc.edu>

The current implementation of pci_find_space does not properly align
PCI capabilities in the PCI configuration space.  This patch fixes
this issue.

Signed-off-by: Matt Renzelmann <mjr@cs.wisc.edu>
---

This is my first patch to QEMU so I hope I'm not screwing up too much.
The purpose of this patch is to mask off the low-order two bits--Linux
masks these while scanning the PCI configuration space, for example,
so we need to make sure QEMU's behavior matches the standard.

No current QEMU hardware is likely using this but it may be important
later.

 hw/pci.c |   14 ++++++++++----
 1 files changed, 10 insertions(+), 4 deletions(-)
Alex Williamson - Sept. 25, 2012, 9:37 p.m.
On Tue, 2012-09-25 at 15:59 -0500, mjr@cs.wisc.edu wrote:
> From: Matt Renzelmann <mjr@cs.wisc.edu>
> 
> The current implementation of pci_find_space does not properly align
> PCI capabilities in the PCI configuration space.  This patch fixes
> this issue.
> 
> Signed-off-by: Matt Renzelmann <mjr@cs.wisc.edu>
> ---
> 
> This is my first patch to QEMU so I hope I'm not screwing up too much.
> The purpose of this patch is to mask off the low-order two bits--Linux
> masks these while scanning the PCI configuration space, for example,
> so we need to make sure QEMU's behavior matches the standard.
> 
> No current QEMU hardware is likely using this but it may be important
> later.
> 
>  hw/pci.c |   14 ++++++++++----
>  1 files changed, 10 insertions(+), 4 deletions(-)
> 
> diff --git a/hw/pci.c b/hw/pci.c
> index e149305..8771b7e 100644
> --- a/hw/pci.c
> +++ b/hw/pci.c
> @@ -1571,11 +1571,17 @@ static int pci_find_space(PCIDevice *pdev, uint8_t size)
>      int config_size = pci_config_size(pdev);
>      int offset = PCI_CONFIG_HEADER_SIZE;
>      int i;
> -    for (i = PCI_CONFIG_HEADER_SIZE; i < config_size; ++i)
> -        if (pdev->used[i])
> -            offset = i + 1;
> -        else if (i - offset + 1 == size)
> +    int masked;
> +
> +    for (i = PCI_CONFIG_HEADER_SIZE; i < config_size; ++i) {
> +        masked = i & (~3);
> +        if (pdev->used[i]) {
> +            offset = masked + 4;
> +        } else if (i - offset + 1 == size) {
>              return offset;
> +        }
> +    }
> +
>      return 0;
>  }
>  

I think you could just search every 4th byte.  In fact, this whole used
byte-map could be turned into a single uint64_t bitmap for standard
config space.  Thanks,

Alex

Patch

diff --git a/hw/pci.c b/hw/pci.c
index e149305..8771b7e 100644
--- a/hw/pci.c
+++ b/hw/pci.c
@@ -1571,11 +1571,17 @@  static int pci_find_space(PCIDevice *pdev, uint8_t size)
     int config_size = pci_config_size(pdev);
     int offset = PCI_CONFIG_HEADER_SIZE;
     int i;
-    for (i = PCI_CONFIG_HEADER_SIZE; i < config_size; ++i)
-        if (pdev->used[i])
-            offset = i + 1;
-        else if (i - offset + 1 == size)
+    int masked;
+
+    for (i = PCI_CONFIG_HEADER_SIZE; i < config_size; ++i) {
+        masked = i & (~3);
+        if (pdev->used[i]) {
+            offset = masked + 4;
+        } else if (i - offset + 1 == size) {
             return offset;
+        }
+    }
+
     return 0;
 }