vfio/nvgrace-gpu: Support huge PFNMAP and wait for GPU ready post reset
[email protected]
maintainerv2 cover vfio/nvgrace-gpu: Support huge PFNMAP and wait for GPU ready post reset
From: Ankit Agrawal <[email protected]>
NVIDIA's Grace based system have large GPU device memory. The device memory is mapped as VM_PFNMAP in the VMM VMA. The nvgrace-gpu module could make use of the huge PFNMAP support added in mm [1].
To achieve this, nvgrace-gpu module is updated to implement huge_fault ops. The implementation establishes mapping according to the order request. Note that if the PFN or the VMA address is unaligned to the order, the mapping fallbacks to the PTE level.
Secondly, it is expected that the mapping not be re-established until the GPU is ready post reset. Presence of the mappings during that time could potentially leads to harmless corrected RAS events to be logged if the CPU attempts to do speculative reads on the GPU memory.
Wait for the GPU to be ready on the first fault. The GPU readiness can be checked through BAR0 registers as is already being done at the device probe.
Patch 1 updates the mapping mechanism to be done through faults.
Patch 2 splits the code to map at the various levels.
Patch 3 implements support for huge pfnmap.
Path 4-6 intercepts reset request and ensures that the GP is ready before re-establishing the mapping after reset.
Applied over 6.18-rc6.
Changelog:
v2:
- Fixed build kernel warning
- subject text changes
- Rebased to 6.18-rc6.
Ankit Agrawal (6):
vfio/nvgrace-gpu: Use faults to map device memory
vfio: export function to map the VMA vfio/nvgrace-gpu: Add support for huge pfnmap
vfio: export vfio_find_cap_start vfio/nvgrace-gpu: split the code to wait for GPU ready vfio/nvgrace-gpu: wait for the GPU mem to be ready
drivers/vfio/pci/nvgrace-gpu/main.c | 171 ++++++++++++++++++++++------ drivers/vfio/pci/vfio_pci_config.c | 3 +- drivers/vfio/pci/vfio_pci_core.c | 46 +++++--- include/linux/vfio_pci_core.h | 3 + 4 files changed, 169 insertions(+), 54 deletions(-)
Signature · 1 line
2.34.1
[email protected]
maintainerPATCH v2 1/6 vfio/nvgrace-gpu: Use faults to map device memory
From: Ankit Agrawal <[email protected]>
To make use of the huge pfnmap support and to support zap/remap sequence, fault/huge_fault ops based mapping mechanism needs to be implemented.
Currently nvgrace-gpu module relies on remap_pfn_range to do the mapping during VM bootup. Replace it to instead rely on fault and use vmf_insert_pfn to setup the mapping.
--- drivers/vfio/pci/nvgrace-gpu/main.c | 50 ++++++++++++++++++----------- 1 file changed, 31 insertions(+), 19 deletions(-)
drivers/vfio/pci/nvgrace-gpu/main.c +31 -19
@@ -130,6 +130,33 @@ static void nvgrace_gpu_close_device(struct vfio_device *core_vdev)vfio_pci_core_close_device (core_vdev); } +static vm_fault_t nvgrace_gpu_vfio_pci_fault (struct vm_fault *vmf) +{ +struct vm_area_struct *vma = vmf->vma; +struct nvgrace_gpu_pci_core_device *nvdev = vma->vm_private_data; +int index = vma->vm_pgoff >> (VFIO_PCI_OFFSET_SHIFT -PAGE_SHIFT ); +vm_fault_t ret =VM_FAULT_SIGBUS ; +struct mem_region *memregion; +unsigned long pgoff, pfn; + + memregion =nvgrace_gpu_memregion (index, nvdev); +if (!memregion) +return ret; + + pgoff = (vmf->address - vma->vm_start) >>PAGE_SHIFT ; + pfn =PHYS_PFN (memregion->memphys) + pgoff; + +down_read (&nvdev->core_device.memory_lock); + ret =vmf_insert_pfn (vmf->vma, vmf->address, pfn); +up_read (&nvdev->core_device.memory_lock); + +return ret; +} + +static const struct vm_operations_struct nvgrace_gpu_vfio_pci_mmap_ops = { + .fault = nvgrace_gpu_vfio_pci_fault, +}; +static int nvgrace_gpu_mmap (struct vfio_device *core_vdev,struct vm_area_struct *vma) {
@@ -137,10 +164,8 @@ static int nvgrace_gpu_mmap(struct vfio_device *core_vdev, container_of(core_vdev, struct nvgrace_gpu_pci_core_device, core_device.vdev);struct mem_region *memregion; -unsigned long start_pfn;u64 req_len, pgoff, end;unsigned int index; -int ret =0 ; index = vma->vm_pgoff >> (VFIO_PCI_OFFSET_SHIFT -PAGE_SHIFT );
@@ -157,7 +182,6 @@ static int nvgrace_gpu_mmap(struct vfio_device *core_vdev, ((1U << (VFIO_PCI_OFFSET_SHIFT -PAGE_SHIFT )) -1 );if (check_sub_overflow (vma->vm_end, vma->vm_start, &req_len) || -check_add_overflow (PHYS_PFN (memregion->memphys), pgoff, &start_pfn) ||check_add_overflow (PFN_PHYS (pgoff), req_len, &end))return -EOVERFLOW ;
@@ -168,6 +192,8 @@ static int nvgrace_gpu_mmap(struct vfio_device *core_vdev,if (end > memregion->memlength)return -EINVAL ; +vm_flags_set (vma,VM_IO |VM_PFNMAP |VM_DONTEXPAND |VM_DONTDUMP ); +/* * The carved out region of the device memory needs the NORMAL_NC * property. Communicate as such to the hypervisor.
@@ -184,23 +210,9 @@ static int nvgrace_gpu_mmap(struct vfio_device *core_vdev, vma->vm_page_prot =pgprot_writecombine (vma->vm_page_prot); } -/* -* Perform a PFN map to the memory and back the device BAR by the -* GPU memory. -* -* The available GPU memory size may not be power-of-2 aligned. The -* remainder is only backed by vfio_device_ops read/write handlers. -* -* During device reset, the GPU is safely disconnected to the CPU -* and access to the BAR will be immediately returned preventing -* machine check. -*/ - ret =remap_pfn_range (vma, vma->vm_start, start_pfn, - req_len, vma->vm_page_prot); -if (ret) -return ret; - vma->vm_pgoff = start_pfn; + vma->vm_ops = &nvgrace_gpu_vfio_pci_mmap_ops; + vma->vm_private_data = nvdev;return 0 ; }
Signature · 1 line
2.34.1
[email protected]
maintainerPATCH v2 2/6 vfio: export function to map the VMA
In reply to [email protected]From: Ankit Agrawal <[email protected]>
Take out the implementation to map the VMA to the PTE/PMD/PUD as a separate function.
Export the function to be used by nvgrace-gpu module.
--- drivers/vfio/pci/vfio_pci_core.c | 46 ++++++++++++++++++++------------ include/linux/vfio_pci_core.h | 2 ++ 2 files changed, 31 insertions(+), 17 deletions(-)
drivers/vfio/pci/vfio_pci_core.c +29 -17
@@ -1640,6 +1640,34 @@ static unsigned long vma_to_pfn(struct vm_area_struct *vma)return (pci_resource_start (vdev->pdev, index) >>PAGE_SHIFT ) + pgoff; } +vm_fault_t vfio_pci_map_pfn (struct vm_fault *vmf, +unsigned long pfn, +unsigned int order) +{ +vm_fault_t ret; + +switch (order) { +case 0 : + ret =vmf_insert_pfn (vmf->vma, vmf->address, pfn); +break ; +#ifdef CONFIG_ARCH_SUPPORTS_PMD_PFNMAP +case PMD_ORDER : + ret =vmf_insert_pfn_pmd (vmf, pfn, false); +break ; +#endif +#ifdef CONFIG_ARCH_SUPPORTS_PUD_PFNMAP +case PUD_ORDER : + ret =vmf_insert_pfn_pud (vmf, pfn, false); +break ; +#endif +default : + ret =VM_FAULT_FALLBACK ; + } + +return ret; +} +EXPORT_SYMBOL_GPL (vfio_pci_map_pfn); +static vm_fault_t vfio_pci_mmap_huge_fault (struct vm_fault *vmf,unsigned int order) {
@@ -1662,23 +1690,7 @@ static vm_fault_t vfio_pci_mmap_huge_fault(struct vm_fault *vmf,if (vdev->pm_runtime_engaged || !__vfio_pci_memory_enabled (vdev)) goto out_unlock; -switch (order) { -case 0 : - ret =vmf_insert_pfn (vma, vmf->address, pfn); -break ; -#ifdef CONFIG_ARCH_SUPPORTS_PMD_PFNMAP -case PMD_ORDER : - ret =vmf_insert_pfn_pmd (vmf, pfn, false); -break ; -#endif -#ifdef CONFIG_ARCH_SUPPORTS_PUD_PFNMAP -case PUD_ORDER : - ret =vmf_insert_pfn_pud (vmf, pfn, false); -break ; -#endif -default : - ret =VM_FAULT_FALLBACK ; - } + ret =vfio_pci_map_pfn (vmf, pfn, order); out_unlock:up_read (&vdev->memory_lock);
include/linux/vfio_pci_core.h +2 -0
@@ -119,6 +119,8 @@ ssize_t vfio_pci_core_read(struct vfio_device *core_vdev, char __user *buf, size_t count, loff_t *ppos);ssize_t vfio_pci_core_write (struct vfio_device *core_vdev,const char __user *buf,size_t count,loff_t *ppos); +vm_fault_t vfio_pci_map_pfn (struct vm_fault *vmf,unsigned long pfn, +unsigned int order);int vfio_pci_core_mmap (struct vfio_device *core_vdev,struct vm_area_struct *vma);void vfio_pci_core_request (struct vfio_device *core_vdev,unsigned int count);int vfio_pci_core_match (struct vfio_device *core_vdev,char *buf);
Signature · 1 line
2.34.1
[email protected]
maintainerPATCH v2 3/6 vfio/nvgrace-gpu: Add support for huge pfnmap
In reply to [email protected]From: Ankit Agrawal <[email protected]>
NVIDIA's Grace based systems have large device memory. The device memory is mapped as VM_PFNMAP in the VMM VMA. The nvgrace-gpu module could make use of the huge PFNMAP support added in mm [1].
To achieve this, nvgrace-gpu module is updated to implement huge_fault ops. The implementation establishes mapping according to the order request. Note that if the PFN or the VMA address is unaligned to the order, the mapping fallbacks to the PTE level.
--- drivers/vfio/pci/nvgrace-gpu/main.c | 44 +++++++++++++++++++++++------ 1 file changed, 35 insertions(+), 9 deletions(-)
drivers/vfio/pci/nvgrace-gpu/main.c +35 -9
@@ -130,33 +130,59 @@ static void nvgrace_gpu_close_device(struct vfio_device *core_vdev)vfio_pci_core_close_device (core_vdev); } -static vm_fault_t nvgrace_gpu_vfio_pci_fault (struct vm_fault *vmf) +static vm_fault_t nvgrace_gpu_vfio_pci_huge_fault (struct vm_fault *vmf, +unsigned int order) {struct vm_area_struct *vma = vmf->vma;struct nvgrace_gpu_pci_core_device *nvdev = vma->vm_private_data;int index = vma->vm_pgoff >> (VFIO_PCI_OFFSET_SHIFT -PAGE_SHIFT );vm_fault_t ret =VM_FAULT_SIGBUS ;struct mem_region *memregion; -unsigned long pgoff, pfn; +unsigned long pgoff, pfn, addr; memregion =nvgrace_gpu_memregion (index, nvdev);if (!memregion)return ret; - pgoff = (vmf->address - vma->vm_start) >>PAGE_SHIFT ; + addr = vmf->address & ~((PAGE_SIZE << order) -1 ); + pgoff = (addr - vma->vm_start) >>PAGE_SHIFT ; pfn =PHYS_PFN (memregion->memphys) + pgoff; +if (order && (addr < vma->vm_start || + addr + (PAGE_SIZE << order) > vma->vm_end || + pfn & ((1 << order) -1 ))) +return VM_FAULT_FALLBACK ; +down_read (&nvdev->core_device.memory_lock); - ret =vmf_insert_pfn (vmf->vma, vmf->address, pfn); + ret =vfio_pci_map_pfn (vmf, pfn, order);up_read (&nvdev->core_device.memory_lock);return ret; } +static vm_fault_t nvgrace_gpu_vfio_pci_fault (struct vm_fault *vmf) +{ +return nvgrace_gpu_vfio_pci_huge_fault (vmf,0 ); +} +static const struct vm_operations_struct nvgrace_gpu_vfio_pci_mmap_ops = { .fault = nvgrace_gpu_vfio_pci_fault, +#ifdef CONFIG_ARCH_SUPPORTS_HUGE_PFNMAP + .huge_fault = nvgrace_gpu_vfio_pci_huge_fault, +#endif }; +static size_t nvgrace_gpu_aligned_devmem_size (size_t memlength) +{ +#ifdef CONFIG_ARCH_SUPPORTS_PMD_PFNMAP +return ALIGN (memlength,PMD_SIZE ); +#endif +#ifdef CONFIG_ARCH_SUPPORTS_PUD_PFNMAP +return ALIGN (memlength,PUD_SIZE ); +#endif +return memlength; +} +static int nvgrace_gpu_mmap (struct vfio_device *core_vdev,struct vm_area_struct *vma) {
@@ -186,10 +212,10 @@ static int nvgrace_gpu_mmap(struct vfio_device *core_vdev,return -EOVERFLOW ;/* -* Check that the mapping request does not go beyond available device -* memory size +* Check that the mapping request does not go beyond the exposed +* device memory size. */ -if (end > memregion->memlength) +if (end >nvgrace_gpu_aligned_devmem_size (memregion->memlength))return -EINVAL ;vm_flags_set (vma,VM_IO |VM_PFNMAP |VM_DONTEXPAND |VM_DONTDUMP );
@@ -210,7 +236,6 @@ static int nvgrace_gpu_mmap(struct vfio_device *core_vdev, vma->vm_page_prot =pgprot_writecombine (vma->vm_page_prot); } - vma->vm_ops = &nvgrace_gpu_vfio_pci_mmap_ops; vma->vm_private_data = nvdev;
@@ -260,7 +285,8 @@ nvgrace_gpu_ioctl_get_region_info(struct vfio_device *core_vdev, sparse->nr_areas =1 ; sparse->areas[0 ].offset =0 ; - sparse->areas[0 ].size = memregion->memlength; + sparse->areas[0 ].size = +nvgrace_gpu_aligned_devmem_size (memregion->memlength); sparse->header.id =VFIO_REGION_INFO_CAP_SPARSE_MMAP ; sparse->header.version =1 ;
Signature · 1 line
2.34.1
[email protected]
maintainerPATCH v2 4/6 vfio: export vfio_find_cap_start
In reply to [email protected]From: Ankit Agrawal <[email protected]>
Export vfio_find_cap_start to be used by the nvgrace-gpu module. This would be used to determine GPU FLR requests.
--- drivers/vfio/pci/vfio_pci_config.c | 3 ++- include/linux/vfio_pci_core.h | 1 + 2 files changed, 3 insertions(+), 1 deletion(-)
drivers/vfio/pci/vfio_pci_config.c +2 -1
@@ -1114,7 +1114,7 @@ int __init vfio_pci_init_perm_bits(void)return ret; } -static int vfio_find_cap_start (struct vfio_pci_core_device *vdev,int pos) +int vfio_find_cap_start (struct vfio_pci_core_device *vdev,int pos) {u8 cap; int base = (pos >= PCI_CFG_SPACE_SIZE) ? PCI_CFG_SPACE_SIZE :
@@ -1130,6 +1130,7 @@ static int vfio_find_cap_start(struct vfio_pci_core_device *vdev, int pos)return pos; } +EXPORT_SYMBOL_GPL (vfio_find_cap_start); static int vfio_msi_config_read(struct vfio_pci_core_device *vdev, int pos, int count, struct perm_bits *perm,
include/linux/vfio_pci_core.h +1 -0
@@ -132,6 +132,7 @@ void vfio_pci_core_finish_enable(struct vfio_pci_core_device *vdev);int vfio_pci_core_setup_barmap (struct vfio_pci_core_device *vdev,int bar);pci_ers_result_t vfio_pci_core_aer_err_detected (struct pci_dev *pdev,pci_channel_state_t state); +int vfio_find_cap_start (struct vfio_pci_core_device *vdev,int pos); ssize_t vfio_pci_core_do_io_rw(struct vfio_pci_core_device *vdev, bool test_mem, void __iomem *io, char __user *buf, loff_t off, size_t count, size_t x_start,
Signature · 1 line
2.34.1
[email protected]
maintainerPATCH v2 5/6 vfio/nvgrace-gpu: split the code to wait for GPU ready
In reply to [email protected]From: Ankit Agrawal <[email protected]>
Split the function that check for the GPU device being ready on the probe.
Move the code to wait for the GPU to be ready through BAR0 register reads to a separate function. This would help reuse the code.
--- drivers/vfio/pci/nvgrace-gpu/main.c | 33 ++++++++++++++++++----------- 1 file changed, 21 insertions(+), 12 deletions(-)
drivers/vfio/pci/nvgrace-gpu/main.c +21 -12
@@ -130,6 +130,24 @@ static void nvgrace_gpu_close_device(struct vfio_device *core_vdev)vfio_pci_core_close_device (core_vdev); } +static int nvgrace_gpu_wait_device_ready (void __iomem *io) +{ +unsigned long timeout = jiffies +msecs_to_jiffies (POLL_TIMEOUT_MS ); +int ret = -ETIME ; + +do { +if ((ioread32 (io +C2C_LINK_BAR0_OFFSET ) ==STATUS_READY ) && + (ioread32 (io +HBM_TRAINING_BAR0_OFFSET ) ==STATUS_READY )) { + ret =0 ; + goto ready_check_exit; + } +msleep (POLL_QUANTUM_MS ); + }while (!time_after (jiffies, timeout)); + +ready_check_exit: +return ret; +} +static vm_fault_t nvgrace_gpu_vfio_pci_huge_fault (struct vm_fault *vmf,unsigned int order) {
@@ -931,9 +949,8 @@ static bool nvgrace_gpu_has_mig_hw_bug(struct pci_dev *pdev)* Ensure that the BAR0 region is enabled before accessing the * registers. */ -static int nvgrace_gpu_wait_device_ready (struct pci_dev *pdev) +static int nvgrace_gpu_check_device_ready (struct pci_dev *pdev) { -unsigned long timeout = jiffies +msecs_to_jiffies (POLL_TIMEOUT_MS );void __iomem *io;int ret = -ETIME ;
@@ -951,16 +968,8 @@ static int nvgrace_gpu_wait_device_ready(struct pci_dev *pdev) goto iomap_exit; } -do { -if ((ioread32 (io +C2C_LINK_BAR0_OFFSET ) ==STATUS_READY ) && - (ioread32 (io +HBM_TRAINING_BAR0_OFFSET ) ==STATUS_READY )) { - ret =0 ; - goto reg_check_exit; - } -msleep (POLL_QUANTUM_MS ); - }while (!time_after (jiffies, timeout)); + ret =nvgrace_gpu_wait_device_ready (io); -reg_check_exit:pci_iounmap (pdev, io); iomap_exit:pci_release_selected_regions (pdev,1 <<0 );
@@ -977,7 +986,7 @@ static int nvgrace_gpu_probe(struct pci_dev *pdev,u64 memphys, memlength;int ret; - ret =nvgrace_gpu_wait_device_ready (pdev); + ret =nvgrace_gpu_check_device_ready (pdev);if (ret)return ret;
Signature · 1 line
2.34.1
[email protected]
maintainerPATCH v2 6/6 vfio/nvgrace-gpu: wait for the GPU mem to be ready
In reply to [email protected]From: Ankit Agrawal <[email protected]>
Speculative prefetches from CPU to GPU memory until the GPU is not ready after reset can cause harmless corrected RAS events to be logged. It is thus expected that the mapping not be re-established until the GPU is ready post reset.
Wait for the GPU to be ready on the first fault before establishing CPU mapping to the GPU memory. The GPU readiness can be checked through BAR0 registers as is already being done at the device probe.
The state is checked on the first fault/huge_fault request using a flag. Unset the flag on every reset request.
So intercept the following calls to the GPU reset, unset gpu_mem_mapped. Then use it to determine whether to wait before mapping.
1. VFIO_DEVICE_RESET ioctl call
2. FLR through config space.
--- drivers/vfio/pci/nvgrace-gpu/main.c | 52 +++++++++++++++++++++++++++++ 1 file changed, 52 insertions(+)
drivers/vfio/pci/nvgrace-gpu/main.c +52 -0
@@ -58,6 +58,8 @@ struct nvgrace_gpu_pci_core_device {/* Lock to control device memory kernel mapping */ struct mutex remap_lock;bool has_mig_hw_bug; +/* Any GPU memory mapped to the VMA */ +bool gpu_mem_mapped; };static void nvgrace_gpu_init_fake_bar_emu_regs (struct vfio_device *core_vdev)
@@ -102,6 +104,8 @@ static int nvgrace_gpu_open_device(struct vfio_device *core_vdev)mutex_init (&nvdev->remap_lock); } + nvdev->gpu_mem_mapped = false; +vfio_pci_core_finish_enable (vdev);return 0 ;
@@ -158,6 +162,24 @@ static vm_fault_t nvgrace_gpu_vfio_pci_huge_fault(struct vm_fault *vmf,struct mem_region *memregion;unsigned long pgoff, pfn, addr; +/* +* If the GPU memory is accessed by the CPU while the GPU is +* not ready after reset, it can cause harmless corrected RAS +* events to be logged. Make sure the GPU is ready before +* establishing the mappings. +*/ +if (!nvdev->gpu_mem_mapped) { +struct vfio_pci_core_device *vdev = &nvdev->core_device; + +if (!vdev->barmap[0 ]) +return VM_FAULT_SIGBUS ; + +if (nvgrace_gpu_wait_device_ready (vdev->barmap[0 ])) +return VM_FAULT_SIGBUS ; + + nvdev->gpu_mem_mapped = true; + } + memregion =nvgrace_gpu_memregion (index, nvdev);if (!memregion)return ret;
@@ -354,7 +376,17 @@ static long nvgrace_gpu_ioctl(struct vfio_device *core_vdev,case VFIO_DEVICE_IOEVENTFD :return -ENOTTY ;case VFIO_DEVICE_RESET : +struct nvgrace_gpu_pci_core_device *nvdev = + container_of(core_vdev, struct nvgrace_gpu_pci_core_device, + core_device.vdev);nvgrace_gpu_init_fake_bar_emu_regs (core_vdev); + +/* +* GPU memory is exposed as device BAR2 (region 4,5). +* This would be zapped during GPU reset. Unset +* nvdev->gpu_mem_mapped to reflect just that. +*/ + nvdev->gpu_mem_mapped = false; fallthrough;default :return vfio_pci_core_ioctl (core_vdev, cmd, arg);
@@ -439,11 +471,14 @@ nvgrace_gpu_write_config_emu(struct vfio_device *core_vdev,struct nvgrace_gpu_pci_core_device *nvdev = container_of(core_vdev, struct nvgrace_gpu_pci_core_device, core_device.vdev); +struct vfio_pci_core_device *vdev = + container_of(core_vdev, struct vfio_pci_core_device, vdev);u64 pos = *ppos &VFIO_PCI_OFFSET_MASK ;struct mem_region *memregion =NULL ;size_t register_offset;loff_t copy_offset;size_t copy_count; +int cap_start =vfio_find_cap_start (vdev, pos); if (vfio_pci_core_range_intersect_range(pos, count, PCI_BASE_ADDRESS_2, sizeof(u64), ©_offset,
@@ -462,6 +497,23 @@ nvgrace_gpu_write_config_emu(struct vfio_device *core_vdev,return copy_count; } +if (vfio_pci_core_range_intersect_range (pos, count, cap_start +PCI_EXP_DEVCTL , +sizeof (u16), ©_offset, + ©_count, ®ister_offset)) { +__le16 val16; + +if (copy_from_user ((void *)&val16, buf, copy_count)) +return -EFAULT ; + +/* +* GPU memory is exposed as device BAR2 (region 4,5). +* This would be zapped during GPU reset. Unset +* nvdev->gpu_mem_mapped to reflect just that. +*/ +if (val16 &cpu_to_le16 (PCI_EXP_DEVCTL_BCR_FLR )) + nvdev->gpu_mem_mapped = false; + } +return vfio_pci_core_write (core_vdev, buf, count, ppos); }
Signature · 1 line
2.34.1