[PATCH v1 1/1] iommu/vt-d: Fix WARN_ON in iommu probe path

Govindapillai, Vinod vinod.govindapillai at intel.com
Tue Apr 9 09:38:20 UTC 2024


+Jani


On Tue, 2024-04-09 at 12:36 +0300, Vinod Govindapillai wrote:
> From: Lu Baolu <baolu.lu at linux.intel.com>
> 
> Commit 1a75cc710b95 ("iommu/vt-d: Use rbtree to track iommu probed
> devices") adds all devices probed by the iommu driver in a rbtree
> indexed by the source ID of each device. It assumes that each device
> has a unique source ID. This assumption is incorrect and the VT-d
> spec doesn't state this requirement either.
> 
> The reason for using a rbtree to track devices is to look up the device
> with PCI bus and devfunc in the paths of handling ATS invalidation time
> out error and the PRI I/O page faults. Both are PCI ATS feature related.
> 
> Only track the devices that have PCI ATS capabilities in the rbtree to
> avoid unnecessary WARN_ON in the iommu probe path. Otherwise, on some
> platforms below kernel splat will be displayed and the iommu probe results
> in failure.
> 
>  WARNING: CPU: 3 PID: 166 at drivers/iommu/intel/iommu.c:158 intel_iommu_probe_device+0x319/0xd90
>  Call Trace:
>   <TASK>
>   ? __warn+0x7e/0x180
>   ? intel_iommu_probe_device+0x319/0xd90
>   ? report_bug+0x1f8/0x200
>   ? handle_bug+0x3c/0x70
>   ? exc_invalid_op+0x18/0x70
>   ? asm_exc_invalid_op+0x1a/0x20
>   ? intel_iommu_probe_device+0x319/0xd90
>   ? debug_mutex_init+0x37/0x50
>   __iommu_probe_device+0xf2/0x4f0
>   iommu_probe_device+0x22/0x70
>   iommu_bus_notifier+0x1e/0x40
>   notifier_call_chain+0x46/0x150
>   blocking_notifier_call_chain+0x42/0x60
>   bus_notify+0x2f/0x50
>   device_add+0x5ed/0x7e0
>   platform_device_add+0xf5/0x240
>   mfd_add_devices+0x3f9/0x500
>   ? preempt_count_add+0x4c/0xa0
>   ? up_write+0xa2/0x1b0
>   ? __debugfs_create_file+0xe3/0x150
>   intel_lpss_probe+0x49f/0x5b0
>   ? pci_conf1_write+0xa3/0xf0
>   intel_lpss_pci_probe+0xcf/0x110 [intel_lpss_pci]
>   pci_device_probe+0x95/0x120
>   really_probe+0xd9/0x370
>   ? __pfx___driver_attach+0x10/0x10
>   __driver_probe_device+0x73/0x150
>   driver_probe_device+0x19/0xa0
>   __driver_attach+0xb6/0x180
>   ? __pfx___driver_attach+0x10/0x10
>   bus_for_each_dev+0x77/0xd0
>   bus_add_driver+0x114/0x210
>   driver_register+0x5b/0x110
>   ? __pfx_intel_lpss_pci_driver_init+0x10/0x10 [intel_lpss_pci]
>   do_one_initcall+0x57/0x2b0
>   ? kmalloc_trace+0x21e/0x280
>   ? do_init_module+0x1e/0x210
>   do_init_module+0x5f/0x210
>   load_module+0x1d37/0x1fc0
>   ? init_module_from_file+0x86/0xd0
>   init_module_from_file+0x86/0xd0
>   idempotent_init_module+0x17c/0x230
>   __x64_sys_finit_module+0x56/0xb0
>   do_syscall_64+0x6e/0x140
>   entry_SYSCALL_64_after_hwframe+0x71/0x79
> 
> Fixes: 1a75cc710b95 ("iommu/vt-d: Use rbtree to track iommu probed devices")
> Signed-off-by: Lu Baolu <baolu.lu at linux.intel.com>
> Signed-off-by: Vinod Govindapillai <vinod.govindapillai at intel.com>
> ---
>  drivers/iommu/intel/iommu.c | 11 +++++++----
>  1 file changed, 7 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
> index 50eb9aed47cc..a7ecd90303dc 100644
> --- a/drivers/iommu/intel/iommu.c
> +++ b/drivers/iommu/intel/iommu.c
> @@ -4299,9 +4299,11 @@ static struct iommu_device *intel_iommu_probe_device(struct device *dev)
>         }
>  
>         dev_iommu_priv_set(dev, info);
> -       ret = device_rbtree_insert(iommu, info);
> -       if (ret)
> -               goto free;
> +       if (pdev && pci_ats_supported(pdev)) {
> +               ret = device_rbtree_insert(iommu, info);
> +               if (ret)
> +                       goto free;
> +       }
>  
>         if (sm_supported(iommu) && !dev_is_real_dma_subdevice(dev)) {
>                 ret = intel_pasid_alloc_table(dev);
> @@ -4336,7 +4338,8 @@ static void intel_iommu_release_device(struct device *dev)
>         struct intel_iommu *iommu = info->iommu;
>  
>         mutex_lock(&iommu->iopf_lock);
> -       device_rbtree_remove(info);
> +       if (dev_is_pci(dev) && pci_ats_supported(to_pci_dev(dev)))
> +               device_rbtree_remove(info);
>         mutex_unlock(&iommu->iopf_lock);
>  
>         if (sm_supported(iommu) && !dev_is_real_dma_subdevice(dev) &&



More information about the Intel-gfx mailing list