Felix Kuehling

What if you change if(0) to if(!a_locked)? That way the compiler would be forced to run hsa_amd_memory_lock before checkLocked(a). Without that, it doesn't know that the checkLocked call depends on...

What you are proposing is like making pthread_mutex_lock timeout after some arbitrary time to workaround a deadlock in an application. Kernels can run for an arbitrary length of time. They...

Probably ROCr isn't handling the HW_EXCEPTION events KFD sends to notify it of the GPU reset. That's assuming that GPU reset is enabled when KFD detects a hang. You should...

Do you mean ATS/PRI or ATC? ATC is an address translation cache, which allows the GPU to cache IOMMU address translations. This can sometimes improve performance under virtualization and is...

I guess my question is, which of these two aspects of ATS and ATC are you interested in: 1. Performance and caching of GPA (guest physical address) to SPA (system...

I got some more information from a colleague who works on our virtualization support: > For the whole ATS to work properly , it requires root complex also support the...

You could test this under bare metal by booting with the IOMMU in device-isolation mode. In this mode each device gets isolated in its own DMA address space, so they...

You should see address translation requests for system memory accesses from the GPU. ATS should result in fewer such requests because the GPU is caching the translations. If you're not...

If you download the package drivers, the kernel module will be built against your current kernel by DKMS. The module will be loaded automatically at boot. You can confirm the...

What's the problem with AMDKFD_IOC_WAIT_EVENTS? I just fixed a problem in this ioctl. See this patch: https://lore.kernel.org/all/[email protected]/