| Message ID | 13744c7ba2af61bc39b68e9ab05ab8482f928334.1514896042.git.robin.murphy@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Tue, Jan 02, 2018 at 12:33:14PM +0000, Robin Murphy wrote: > For PCI devices behind an aliasing PCIe-to-PCI/X bridge, the bridge > alias to DevFn 0.0 on the subordinate bus may match the original RID of > the device, resulting in the same SID being present in the device's > fwspec twice. This causes trouble later in arm_smmu_write_strtab_ent() > when we wind up visiting the STE a second time and find it already live. > > Avoid the issue by giving arm_smmu_install_ste_for_dev() the cleverness > to skip over duplicates. It seems mildly counterintuitive compared to > preventing the duplicates from existing in the first place, but since > the DT and ACPI probe paths build their fwspecs differently, this is > actually the cleanest and most self-contained way to deal with it. > > Fixes: 8f78515425da ("iommu/arm-smmu: Implement of_xlate() for SMMUv3") > Reported-by: Tomasz Nowicki <tomasz.nowicki@caviumnetworks.com> > Tested-by: Tomasz Nowicki <Tomasz.Nowicki@cavium.com> > Tested-by: Jayachandran C. <jnair@caviumnetworks.com> > Signed-off-by: Robin Murphy <robin.murphy@arm.com> > --- > drivers/iommu/arm-smmu-v3.c | 9 ++++++++- > 1 file changed, 8 insertions(+), 1 deletion(-) Cheers, Robin. I'll pick this up and send to Alex later this week along with a couple of other fixes I have kicking around. Will
Hi Robin, Thank you for fixing this. Regards, Tomasz On 02.01.2018 13:33, Robin Murphy wrote: > For PCI devices behind an aliasing PCIe-to-PCI/X bridge, the bridge > alias to DevFn 0.0 on the subordinate bus may match the original RID of > the device, resulting in the same SID being present in the device's > fwspec twice. This causes trouble later in arm_smmu_write_strtab_ent() > when we wind up visiting the STE a second time and find it already live. > > Avoid the issue by giving arm_smmu_install_ste_for_dev() the cleverness > to skip over duplicates. It seems mildly counterintuitive compared to > preventing the duplicates from existing in the first place, but since > the DT and ACPI probe paths build their fwspecs differently, this is > actually the cleanest and most self-contained way to deal with it. > > Fixes: 8f78515425da ("iommu/arm-smmu: Implement of_xlate() for SMMUv3") > Reported-by: Tomasz Nowicki <tomasz.nowicki@caviumnetworks.com> > Tested-by: Tomasz Nowicki <Tomasz.Nowicki@cavium.com> > Tested-by: Jayachandran C. <jnair@caviumnetworks.com> > Signed-off-by: Robin Murphy <robin.murphy@arm.com> > --- > drivers/iommu/arm-smmu-v3.c | 9 ++++++++- > 1 file changed, 8 insertions(+), 1 deletion(-) > > diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c > index 9ce3cde575a8..57c92aa3122e 100644 > --- a/drivers/iommu/arm-smmu-v3.c > +++ b/drivers/iommu/arm-smmu-v3.c > @@ -1752,7 +1752,7 @@ static __le64 *arm_smmu_get_step_for_sid(struct arm_smmu_device *smmu, u32 sid) > > static void arm_smmu_install_ste_for_dev(struct iommu_fwspec *fwspec) > { > - int i; > + int i, j; > struct arm_smmu_master_data *master = fwspec->iommu_priv; > struct arm_smmu_device *smmu = master->smmu; > > @@ -1760,6 +1760,13 @@ static void arm_smmu_install_ste_for_dev(struct iommu_fwspec *fwspec) > u32 sid = fwspec->ids[i]; > __le64 *step = arm_smmu_get_step_for_sid(smmu, sid); > > + /* Bridged PCI devices may end up with duplicated IDs */ > + for (j = 0; j < i; j++) > + if (fwspec->ids[j] == sid) > + break; > + if (j < i) > + continue; > + > arm_smmu_write_strtab_ent(smmu, sid, step, &master->ste); > } > } >
On Tue, Jan 02, 2018 at 01:00:36PM +0000, Will Deacon wrote: > On Tue, Jan 02, 2018 at 12:33:14PM +0000, Robin Murphy wrote: > > For PCI devices behind an aliasing PCIe-to-PCI/X bridge, the bridge > > alias to DevFn 0.0 on the subordinate bus may match the original RID of > > the device, resulting in the same SID being present in the device's > > fwspec twice. This causes trouble later in arm_smmu_write_strtab_ent() > > when we wind up visiting the STE a second time and find it already live. > > > > Avoid the issue by giving arm_smmu_install_ste_for_dev() the cleverness > > to skip over duplicates. It seems mildly counterintuitive compared to > > preventing the duplicates from existing in the first place, but since > > the DT and ACPI probe paths build their fwspecs differently, this is > > actually the cleanest and most self-contained way to deal with it. > > > > Fixes: 8f78515425da ("iommu/arm-smmu: Implement of_xlate() for SMMUv3") > > Reported-by: Tomasz Nowicki <tomasz.nowicki@caviumnetworks.com> > > Tested-by: Tomasz Nowicki <Tomasz.Nowicki@cavium.com> > > Tested-by: Jayachandran C. <jnair@caviumnetworks.com> > > Signed-off-by: Robin Murphy <robin.murphy@arm.com> > > --- > > drivers/iommu/arm-smmu-v3.c | 9 ++++++++- > > 1 file changed, 8 insertions(+), 1 deletion(-) > > Cheers, Robin. I'll pick this up and send to Alex later this week along with > a couple of other fixes I have kicking around. Thanks. Even though the Fixes: tag points to an older commit, we really need the fix in 4.14 where another commit (10631d724def drm/pci: Deprecate drm_pci_init/exit completely) exposed the issue. It it is not already planned, can you please cc stable? JC.
diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c index 9ce3cde575a8..57c92aa3122e 100644 --- a/drivers/iommu/arm-smmu-v3.c +++ b/drivers/iommu/arm-smmu-v3.c @@ -1752,7 +1752,7 @@ static __le64 *arm_smmu_get_step_for_sid(struct arm_smmu_device *smmu, u32 sid) static void arm_smmu_install_ste_for_dev(struct iommu_fwspec *fwspec) { - int i; + int i, j; struct arm_smmu_master_data *master = fwspec->iommu_priv; struct arm_smmu_device *smmu = master->smmu; @@ -1760,6 +1760,13 @@ static void arm_smmu_install_ste_for_dev(struct iommu_fwspec *fwspec) u32 sid = fwspec->ids[i]; __le64 *step = arm_smmu_get_step_for_sid(smmu, sid); + /* Bridged PCI devices may end up with duplicated IDs */ + for (j = 0; j < i; j++) + if (fwspec->ids[j] == sid) + break; + if (j < i) + continue; + arm_smmu_write_strtab_ent(smmu, sid, step, &master->ste); } }
For PCI devices behind an aliasing PCIe-to-PCI/X bridge, the bridge alias to DevFn 0.0 on the subordinate bus may match the original RID of the device, resulting in the same SID being present in the device's fwspec twice. This causes trouble later in arm_smmu_write_strtab_ent() when we wind up visiting the STE a second time and find it already live. Avoid the issue by giving arm_smmu_install_ste_for_dev() the cleverness to skip over duplicates. It seems mildly counterintuitive compared to preventing the duplicates from existing in the first place, but since the DT and ACPI probe paths build their fwspecs differently, this is actually the cleanest and most self-contained way to deal with it. Fixes: 8f78515425da ("iommu/arm-smmu: Implement of_xlate() for SMMUv3") Reported-by: Tomasz Nowicki <tomasz.nowicki@caviumnetworks.com> Tested-by: Tomasz Nowicki <Tomasz.Nowicki@cavium.com> Tested-by: Jayachandran C. <jnair@caviumnetworks.com> Signed-off-by: Robin Murphy <robin.murphy@arm.com> --- drivers/iommu/arm-smmu-v3.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)