drivers/iommu/s390-iommu.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
zpci_get_iommu_ctrs() returns counter information to be reported as part
of device statistics; these counters are stored as part of the s390_domain.
The problem, however, is that the identity domain is not backed by an
s390_domain and so the conversion via to_s390_domain() yields a bad address
that is zero'd initially and read on-demand later via a sysfs read.
These counters aren't necessary for the identity domain; just return NULL
in this case.
This issue was discovered via KASAN with reports that look like:
BUG: KASAN: global-out-of-bounds in zpci_fmb_enable_device
when using the identity domain for a device on s390.
Cc: stable@vger.kernel.org
Fixes: 64af12c6ec3a ("iommu/s390: implement iommu passthrough via identity domain")
Reported-by: Cam Miller <cam@linux.ibm.com>
Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
---
drivers/iommu/s390-iommu.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c
index 9c80d61deb2c..d7370347c910 100644
--- a/drivers/iommu/s390-iommu.c
+++ b/drivers/iommu/s390-iommu.c
@@ -1032,7 +1032,8 @@ struct zpci_iommu_ctrs *zpci_get_iommu_ctrs(struct zpci_dev *zdev)
lockdep_assert_held(&zdev->dom_lock);
- if (zdev->s390_domain->type == IOMMU_DOMAIN_BLOCKED)
+ if (zdev->s390_domain->type == IOMMU_DOMAIN_BLOCKED ||
+ zdev->s390_domain->type == IOMMU_DOMAIN_IDENTITY)
return NULL;
s390_domain = to_s390_domain(zdev->s390_domain);
--
2.50.1
On Wed, Aug 27, 2025 at 05:08:27PM -0400, Matthew Rosato wrote: > zpci_get_iommu_ctrs() returns counter information to be reported as part > of device statistics; these counters are stored as part of the s390_domain. > The problem, however, is that the identity domain is not backed by an > s390_domain and so the conversion via to_s390_domain() yields a bad address > that is zero'd initially and read on-demand later via a sysfs read. > These counters aren't necessary for the identity domain; just return NULL > in this case. > > This issue was discovered via KASAN with reports that look like: > BUG: KASAN: global-out-of-bounds in zpci_fmb_enable_device > when using the identity domain for a device on s390. > > Cc: stable@vger.kernel.org > Fixes: 64af12c6ec3a ("iommu/s390: implement iommu passthrough via identity domain") > Reported-by: Cam Miller <cam@linux.ibm.com> > Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com> > --- > drivers/iommu/s390-iommu.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) Applied for -rc, thanks.
Tested-by: Cam Miller <cam@linux.ibm.com> On 8/27/25 17:08, Matthew Rosato wrote: > zpci_get_iommu_ctrs() returns counter information to be reported as part > of device statistics; these counters are stored as part of the s390_domain. > The problem, however, is that the identity domain is not backed by an > s390_domain and so the conversion via to_s390_domain() yields a bad address > that is zero'd initially and read on-demand later via a sysfs read. > These counters aren't necessary for the identity domain; just return NULL > in this case. > > This issue was discovered via KASAN with reports that look like: > BUG: KASAN: global-out-of-bounds in zpci_fmb_enable_device > when using the identity domain for a device on s390. > > Cc: stable@vger.kernel.org > Fixes: 64af12c6ec3a ("iommu/s390: implement iommu passthrough via identity domain") > Reported-by: Cam Miller <cam@linux.ibm.com> > Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com> > --- > drivers/iommu/s390-iommu.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c > index 9c80d61deb2c..d7370347c910 100644 > --- a/drivers/iommu/s390-iommu.c > +++ b/drivers/iommu/s390-iommu.c > @@ -1032,7 +1032,8 @@ struct zpci_iommu_ctrs *zpci_get_iommu_ctrs(struct zpci_dev *zdev) > > lockdep_assert_held(&zdev->dom_lock); > > - if (zdev->s390_domain->type == IOMMU_DOMAIN_BLOCKED) > + if (zdev->s390_domain->type == IOMMU_DOMAIN_BLOCKED || > + zdev->s390_domain->type == IOMMU_DOMAIN_IDENTITY) > return NULL; > > s390_domain = to_s390_domain(zdev->s390_domain);
On 8/27/2025 2:08 PM, Matthew Rosato wrote: > zpci_get_iommu_ctrs() returns counter information to be reported as part > of device statistics; these counters are stored as part of the s390_domain. > The problem, however, is that the identity domain is not backed by an > s390_domain and so the conversion via to_s390_domain() yields a bad address > that is zero'd initially and read on-demand later via a sysfs read. > These counters aren't necessary for the identity domain; just return NULL > in this case. > > This issue was discovered via KASAN with reports that look like: > BUG: KASAN: global-out-of-bounds in zpci_fmb_enable_device > when using the identity domain for a device on s390. > > Cc: stable@vger.kernel.org > Fixes: 64af12c6ec3a ("iommu/s390: implement iommu passthrough via identity domain") > Reported-by: Cam Miller <cam@linux.ibm.com> > Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com> > --- > drivers/iommu/s390-iommu.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c > index 9c80d61deb2c..d7370347c910 100644 > --- a/drivers/iommu/s390-iommu.c > +++ b/drivers/iommu/s390-iommu.c > @@ -1032,7 +1032,8 @@ struct zpci_iommu_ctrs *zpci_get_iommu_ctrs(struct zpci_dev *zdev) > > lockdep_assert_held(&zdev->dom_lock); > > - if (zdev->s390_domain->type == IOMMU_DOMAIN_BLOCKED) > + if (zdev->s390_domain->type == IOMMU_DOMAIN_BLOCKED || > + zdev->s390_domain->type == IOMMU_DOMAIN_IDENTITY) > return NULL; > > s390_domain = to_s390_domain(zdev->s390_domain); > Reviewed-by: Farhan Ali <alifm@linux.ibm.com>
On Wed, 2025-08-27 at 17:08 -0400, Matthew Rosato wrote: > zpci_get_iommu_ctrs() returns counter information to be reported as part > of device statistics; these counters are stored as part of the s390_domain. > The problem, however, is that the identity domain is not backed by an > s390_domain and so the conversion via to_s390_domain() yields a bad address > that is zero'd initially and read on-demand later via a sysfs read. > These counters aren't necessary for the identity domain; just return NULL > in this case. > > This issue was discovered via KASAN with reports that look like: > BUG: KASAN: global-out-of-bounds in zpci_fmb_enable_device > when using the identity domain for a device on s390. > > Cc: stable@vger.kernel.org > Fixes: 64af12c6ec3a ("iommu/s390: implement iommu passthrough via identity domain") > Reported-by: Cam Miller <cam@linux.ibm.com> > Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com> > --- > drivers/iommu/s390-iommu.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c > index 9c80d61deb2c..d7370347c910 100644 > --- a/drivers/iommu/s390-iommu.c > +++ b/drivers/iommu/s390-iommu.c > @@ -1032,7 +1032,8 @@ struct zpci_iommu_ctrs *zpci_get_iommu_ctrs(struct zpci_dev *zdev) > > lockdep_assert_held(&zdev->dom_lock); > > - if (zdev->s390_domain->type == IOMMU_DOMAIN_BLOCKED) > + if (zdev->s390_domain->type == IOMMU_DOMAIN_BLOCKED || > + zdev->s390_domain->type == IOMMU_DOMAIN_IDENTITY) > return NULL; > > s390_domain = to_s390_domain(zdev->s390_domain); Looks good to me and thanks for the fix! Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com>
© 2016 - 2025 Red Hat, Inc.