RE: [PATCH v2 1/1] vfio/nvgrace-gpu: Add Blackwell-Next GPU readiness check via CXL DVSEC

[Tian, Kevin]

> From: Ankit Agrawal <ankita@xxxxxxxxxx>
> Sent: Thursday, April 9, 2026 9:37 PM
> 
> +static inline unsigned long nvgrace_gpu_cxl_mem_active_timeout_ms(u8
> timeout)
> +{
> +	return 1000UL << (2 * min_t(u8, timeout, 4));
> +}

this is local to this file and simple. it's ok to remove the 'nvgrace_gpu_'
prefix so the caller line is shorter. Later it may be moved to cxl header
file if other callers exist.

> +
> +static int nvgrace_gpu_wait_device_ready_bw_next(struct
> nvgrace_gpu_pci_core_device *nvdev)

it's a new method in Blackwell-Next, but the mechanism is CXL standard.

call it nvgrace_gpu_wait_device_ready_cxl()?

> +{
> +	struct pci_dev *pdev = nvdev->core_device.pdev;
> +	int pcie_dvsec = nvdev->cxl_dvsec;

s/pcie_dvsec/cxl_dvsec/

> -static int nvgrace_gpu_probe_check_device_ready(struct pci_dev *pdev)
> +static int nvgrace_gpu_probe_check_device_ready(struct
> nvgrace_gpu_pci_core_device *nvdev)
>  {
> +	struct pci_dev *pdev = nvdev->core_device.pdev;
>  	void __iomem *io;
>  	int ret;
> 
> @@ -1165,7 +1221,7 @@ static int
> nvgrace_gpu_probe_check_device_ready(struct pci_dev *pdev)
>  		goto iomap_exit;
>  	}
> 
> -	ret = nvgrace_gpu_wait_device_ready(io);
> +	ret = nvgrace_gpu_wait_device_ready(nvdev, io);
> 
>  	pci_iounmap(pdev, io);

bar0 mapping is not required in the new method.