[Webkit-unassigned] [Bug 250862] REGRESSION(?): validateTexFuncData is getting passed the entire WASM ArrayBuffer instead of the texture buffer

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Fri Jan 20 15:27:18 PST 2023


https://bugs.webkit.org/show_bug.cgi?id=250862

Brendan Duncan <brendanduncan at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |brendanduncan at gmail.com

--- Comment #4 from Brendan Duncan <brendanduncan at gmail.com> ---
Hi Alexey,

I don't believe this is a regression in WebGL2, but a regression from WebGL1 to WebGL2. WebGL2 has the new variant of texSubImage2D that takes the WASM heap pointer and a srcOffset to the texture data. This is the variant that is reporting too much memory for the texture.

In WebGLRenderingContextBase::validateTexFuncData, it returns byteLength, which it gets from pixels->byteLength(). byteLength is getting adjusted by srcOffset. However, pixels is still a view to the entire heap, so byteLength will be heapSize - srcOffset. From our understanding, this is causing even a small texture to report needing a large amount of memory.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20230120/4b0eba93/attachment.htm>


More information about the webkit-unassigned mailing list