Skip to content

Commit

Permalink
drm/nouveau/imem/nv50: move slow-path locking into rd/wr functions
Browse files Browse the repository at this point in the history
This is to simplify upcoming changes.  The slow-path is something that
currently occurs during bootstrap of the BAR2 VMM, while backing up an
object during suspend/resume, or when BAR2 address space runs out.

The latter is a real problem that can happen at runtime, and occurs in
Fedora 26 already (due to some change that causes a lot of channels to
be created at login), so ideally we'd prefer not to make it any slower.

We'd also like suspend/resume speed to not suffer.

Upcoming commits will solve those problems in a better way, making the
extra overhead of moving the locking here a non-issue.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
  • Loading branch information
Ben Skeggs committed Nov 2, 2017
1 parent f584bde commit af515ec
Showing 1 changed file with 6 additions and 8 deletions.
14 changes: 6 additions & 8 deletions drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c
Original file line number Diff line number Diff line change
Expand Up @@ -31,8 +31,6 @@

struct nv50_instmem {
struct nvkm_instmem base;
unsigned long lock_flags;
spinlock_t lock;
u64 addr;
};

Expand All @@ -57,12 +55,15 @@ nv50_instobj_wr32_slow(struct nvkm_memory *memory, u64 offset, u32 data)
struct nvkm_device *device = imem->base.subdev.device;
u64 base = (iobj->mem->offset + offset) & 0xffffff00000ULL;
u64 addr = (iobj->mem->offset + offset) & 0x000000fffffULL;
unsigned long flags;

spin_lock_irqsave(&imem->base.lock, flags);
if (unlikely(imem->addr != base)) {
nvkm_wr32(device, 0x001700, base >> 16);
imem->addr = base;
}
nvkm_wr32(device, 0x700000 + addr, data);
spin_unlock_irqrestore(&imem->base.lock, flags);
}

static u32
Expand All @@ -74,12 +75,15 @@ nv50_instobj_rd32_slow(struct nvkm_memory *memory, u64 offset)
u64 base = (iobj->mem->offset + offset) & 0xffffff00000ULL;
u64 addr = (iobj->mem->offset + offset) & 0x000000fffffULL;
u32 data;
unsigned long flags;

spin_lock_irqsave(&imem->base.lock, flags);
if (unlikely(imem->addr != base)) {
nvkm_wr32(device, 0x001700, base >> 16);
imem->addr = base;
}
data = nvkm_rd32(device, 0x700000 + addr);
spin_unlock_irqrestore(&imem->base.lock, flags);
return data;
}

Expand Down Expand Up @@ -127,8 +131,6 @@ nv50_instobj_map(struct nvkm_memory *memory, struct nvkm_vma *vma, u64 offset)
static void
nv50_instobj_release(struct nvkm_memory *memory)
{
struct nv50_instmem *imem = nv50_instobj(memory)->imem;
spin_unlock_irqrestore(&imem->lock, imem->lock_flags);
}

static void __iomem *
Expand All @@ -137,15 +139,12 @@ nv50_instobj_acquire(struct nvkm_memory *memory)
struct nv50_instobj *iobj = nv50_instobj(memory);
struct nv50_instmem *imem = iobj->imem;
struct nvkm_vm *vm;
unsigned long flags;

if (!iobj->map && (vm = nvkm_bar_bar2_vmm(imem->base.subdev.device)))
nv50_instobj_kmap(iobj, vm);
if (!IS_ERR_OR_NULL(iobj->map))
return iobj->map;

spin_lock_irqsave(&imem->lock, flags);
imem->lock_flags = flags;
return NULL;
}

Expand Down Expand Up @@ -254,7 +253,6 @@ nv50_instmem_new(struct nvkm_device *device, int index,
if (!(imem = kzalloc(sizeof(*imem), GFP_KERNEL)))
return -ENOMEM;
nvkm_instmem_ctor(&nv50_instmem, device, index, &imem->base);
spin_lock_init(&imem->lock);
*pimem = &imem->base;
return 0;
}

0 comments on commit af515ec

Please sign in to comment.