gpu: nvgpu: Fix bitmap_allocator zero length bugs

Currently, function nvgpu_bitmap_allocator_init() initializes bitmap
with length=0 but will fail in alloc(). Function alloc() allocates
len=0 bitmap and next request also starts from same address. However,
rbtree only holds zero length allocation.

This patch adds length = 0 check in nvgpu_bitmap_allocator_init() and
alloc() functions.

JIRA NVGPU-3086

Change-Id: I0936977cd193f3eba00bba28edae257e40af23bf
Signed-off-by: Vedashree Vidwans <vvidwans@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2087077
Reviewed-by: Nicolas Benech <nbenech@nvidia.com>
Reviewed-by: svc-misra-checker <svc-misra-checker@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This commit is contained in:
Vedashree Vidwans
2019-04-01 14:38:41 -07:00
committed by mobile promotions
parent 7d60eb5bf0
commit 1f27acb983

View File

@@ -209,6 +209,11 @@ static u64 nvgpu_bitmap_alloc(struct nvgpu_allocator *na, u64 len)
unsigned long offs, adjusted_offs, limit;
struct nvgpu_bitmap_allocator *a = bitmap_allocator(na);
if (len == 0ULL) {
alloc_dbg(na, "len = 0, Alloc failed!");
return 0;
}
tmp_u64 = len >> a->blk_shift;
nvgpu_assert(tmp_u64 <= U32_MAX);
blks = (u32)tmp_u64;
@@ -405,6 +410,10 @@ int nvgpu_bitmap_allocator_init(struct gk20a *g, struct nvgpu_allocator *na,
return -EINVAL;
}
if (length == 0ULL) {
return -EINVAL;
}
if (base == 0ULL) {
base = blk_size;
length -= blk_size;