Message ID | 20230602004824.20731-11-vikram.garhwal@amd.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | dynamic node programming using overlay dtbo | expand |
On 02.06.2023 02:48, Vikram Garhwal wrote: > @@ -189,6 +194,8 @@ int iommu_add_dt_device(struct dt_device_node *np) > if ( rc < 0 ) > iommu_fwspec_free(dev); > > +fail: > + spin_unlock(&dtdevs_lock); Nit: Labels indented by at least on blank please (see ./CODING_STYLE). Jan
Hi,
On 02/06/2023 01:48, Vikram Garhwal wrote:
> Protect iommu_add_dt_device() with dtdevs_lock to prevent concurrent access add.
The commit message is a bit light. What sort of concurrent access add
are you talking about? Is it of the same node? Different node?
Also, is it a existing issue or something that will become one with a
follow-up patch in your series?
Cheers,
diff --git a/xen/drivers/passthrough/device_tree.c b/xen/drivers/passthrough/device_tree.c index 52e370db01..8cc413f867 100644 --- a/xen/drivers/passthrough/device_tree.c +++ b/xen/drivers/passthrough/device_tree.c @@ -146,6 +146,8 @@ int iommu_add_dt_device(struct dt_device_node *np) if ( dev_iommu_fwspec_get(dev) ) return 0; + spin_lock(&dtdevs_lock); + /* * According to the Documentation/devicetree/bindings/iommu/iommu.txt * from Linux. @@ -158,7 +160,10 @@ int iommu_add_dt_device(struct dt_device_node *np) * these callback implemented. */ if ( !ops->add_device || !ops->dt_xlate ) - return -EINVAL; + { + rc = -EINVAL; + goto fail; + } if ( !dt_device_is_available(iommu_spec.np) ) break; @@ -189,6 +194,8 @@ int iommu_add_dt_device(struct dt_device_node *np) if ( rc < 0 ) iommu_fwspec_free(dev); +fail: + spin_unlock(&dtdevs_lock); return rc; }