mirror of
https://github.com/tangerineturtle1/NX2navi.git
synced 2026-11-17 17:59:58 -04:00
Free a destroyed texture's record only after the GPU is past the submission being recorded, and cache no view of it.
A device lost seen on RTX 4090, 5090 and 5070 Ti cards in 0.5.0 and 0.6.0 (#177, #61, and a 0.6.0 test run): a "read invalid" at the start of a 128x128 scene capture pass, re-rendered at area loads. destroy_tex purged the views of the image and deferred the image, then freed its record at once, on whichever thread released the texture. The replay keeps &e->pub as a render target for the rest of a command buffer (vk_exec.c resolve_target) and DLSS for the frame: a pass that began again (each indirect mesh draw ends the pass) read the freed record and cached a view of the dead image, which nothing removed. Vulkan gives a destroyed image's handle to a later image, which then got that view, and the pass loaded its attachments from freed memory. - destroy_tex marks the record gone and frees it through the deferred destruction (vk_defer_free, a host kind stamped with the submission being recorded, as the image is). - A gone texture's attachment view is made per command buffer (vk_exec.c view_of) or deferred with it (vk_exec_target_view, for DLSS and the contact shadows), never cached. - A new image drops any view still cached under its handle, and the log counts it ("cached views of a destroyed image"). No rendering changes: the views and images live exactly as long as before, the records slightly longer. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
@@ -707,6 +707,16 @@ static Tex *create_tex(const tex_t *t, const TexKey *k, int defer_init) {
|
||||
e->bytes = own.size;
|
||||
}
|
||||
e->reusable = e->alloc && gpu_only;
|
||||
// A new image's handle can be one a destroyed image had. Views are purged with the image (destroy_tex), so none
|
||||
// should be cached under it: one that is was made from a destroyed texture's record after its purge, and would
|
||||
// reach this image's memory through the dead one's. Dropped here, and counted in the log.
|
||||
if (!e->reused) {
|
||||
static volatile int32_t n_stale;
|
||||
int stale = vk_exec_image_gone(img);
|
||||
if (stale)
|
||||
log_limited(&n_stale, "new image %#llx had %d cached views of a destroyed image with its handle; dropped",
|
||||
(unsigned long long)(uintptr_t)img, stale);
|
||||
}
|
||||
VkTexImage *p = &e->pub;
|
||||
p->image = img;
|
||||
p->format = f.vk;
|
||||
@@ -833,7 +843,12 @@ static void destroy_tex(Tex *e) {
|
||||
if (e->pub.image) vk_exec_image_gone(e->pub.image); // none when another texture took it (create_tex, reuse)
|
||||
vk_defer_image(e->pub.image, e->alloc);
|
||||
free(e->snap);
|
||||
free(e);
|
||||
// The record goes once the GPU is past the submission being recorded, as its image: the replay keeps &e->pub as a
|
||||
// render target for the rest of a command buffer (vk_exec.c resolve_target) and DLSS for the frame, and this runs on
|
||||
// whichever thread released the texture. Freed here, a pass that began again read the freed record, cached a view of
|
||||
// the dead image (vk_exec_target_view), and a later image given the same handle drew through it into freed memory.
|
||||
e->pub.gone = 1;
|
||||
vk_defer_free(e);
|
||||
}
|
||||
|
||||
// Released entries with another descriptor at the same address are replaced; past RELEASED_MAX released entries the
|
||||
|
||||
Reference in New Issue
Block a user