On 08/20/2010 09:54 PM, [email protected] wrote:
From: Alexander Larsson<[email protected]>
This is smaller than a byte array, and allows us to skip full
blocks of 32 ids in one check.
---
Hi,
why not use a linked list, for free surfaces, in a static
UINT32[n_surfaces] array? Any reason besides space, which is 4k?
static _inline void FreeSurface(PDev *pdev, UINT32 surface_id)
{
- pdev->Res.surfaces_used[surface_id] = 0;
+ UINT32 i, bitmask;
+
+ i = surface_id / 32;
+ bitmask = (~0x80000000)>> (surface_id % 32);
+
+ pdev->Res.surfaces_used[i]&= bitmask;
}
shouldn't it be ~(0x80000000 << (surface_id % 32))?
static UINT32 GetFreeSurface(PDev *pdev)
{
- UINT32 x;
-
- //not effective, fix me
- for (x = 1; x< pdev->n_surfaces; ++x) {
- if (!pdev->Res.surfaces_used[x]) {
- pdev->Res.surfaces_used[x] = 1;
- return x;
+ UINT32 start = 1;
+ UINT32 bit;
+ UINT32 i, end_i;
+ UINT32 bitmask;
+ UINT32 used_mask;
+ UINT32 res = 0;
+
+ i = start / 32;
+ end_i = (pdev->n_surfaces + 31) / 32;
+ bit = start % 32;
+ bitmask = (0x80000000)>> (bit % 32);
+
+ while (i< end_i) {
+ used_mask = pdev->Res.surfaces_used[i];
+ /* Avoid checks early if no unused */
+ if (used_mask == 0xffffffff) {
shouldn't it be used_mask != 0xffffffff?
_______________________________________________
Spice-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/spice-devel
