[robtsuk]

The Rust Vulkano wrapper around Vulkan uses zero time timeouts which produce a lot of these messages. The author tells me that these zero timeouts are in order to check whether a fence is signalled without blocking the thread.

If this is in fact a legitimate use I think it would be great if the logging of this condition could be removed.

[Bill Hollings]

Thanks for bringing this to our attention.

We’ll look into it.

…Bill

[GolDDranks]

I’m encountering this too: loads of…

[***MoltenVK ERROR***] VK_TIMEOUT: Vulkan fence timeout after 0 nanoseconds.

Using Rust + the Vulkano library on MacOS 10.12 Sierra + MoltenVK.

Here’s a thread mentioning the issue (it was originally about a HiDPI bug, but it’s still unclear whether that is an MoltenVK issue or not): https://github.com/vulkano-rs/vulkano/issues/738

Tomaka, the author of the library mentions:

A timeout of zero is used in order to check whether a fence is signalled without blocking the thread.
It’s a valid thing to do (the specs even talk explicitly about that case if I remember correctly) and I’d even say that it is probably a very common thing to do.

Indeed, he is probably talking about the usage mentioned in here: https://www.khronos.org/registry/vulkan/specs/1.0/man/html/vkWaitForFences.html

If timeout is zero, then vkWaitForFences does not wait, but simply returns the current state of the fences. VK_TIMEOUT will be returned in this case if the condition is not satisfied, even though no actual wait was performed.

It would be great if you didn’t log that use case out as an error, as it seems to be a valid thing to do.

[Bill Hollings]

@golddranks & @robtsuk

Yes…the excessive logging is a MoltenVK issue.

It has been corrected for the next release of MoltenVK.

…Bill

[Author]

Posts