This patch backport the previous commit from pbrobinson in devel
to the f21 branch (without the changelog entry)
Tested with fedpkg prep to verify that the patches applies
---
arm-fix-iommu-exynos.patch | 48 +++++++++++++++++++++++++++++++++++++++
arm-fix-iommu-omap.patch | 48 +++++++++++++++++++++++++++++++++++++++
arm-fix-iommu-rockchip.patch | 51 ++++++++++++++++++++++++++++++++++++++++++
kernel.spec | 9 +++++++
4 files changed, 156 insertions(+), 0 deletions(-)
create mode 100644 arm-fix-iommu-exynos.patch
create mode 100644 arm-fix-iommu-omap.patch
create mode 100644 arm-fix-iommu-rockchip.patch
diff --git a/arm-fix-iommu-exynos.patch b/arm-fix-iommu-exynos.patch
new file mode 100644
index 0000000..447716f
--- /dev/null
+++ b/arm-fix-iommu-exynos.patch
@@ -0,0 +1,48 @@
+From: Thierry Reding <treding at nvidia.com>
+
+The Exynos System MMU driver unconditionally executes code and registers
+a struct iommu_ops with the platform bus irrespective of whether it runs
+on an Exynos SoC or not. This causes problems in multi-platform kernels
+where drivers for other SoCs will no longer be able to register their
+own struct iommu_ops or even try to use a struct iommu_ops for an IOMMU
+that obviously isn't there.
+
+The smallest fix I could think of is to check for the existence of any
+Exynos System MMU devices in the device tree and skip initialization
+otherwise.
+
+This fixes a problem on Tegra20 where the DRM driver will try to use the
+obviously non-existent Exynos System MMU.
+
+Reported-by: Nicolas Chauvet <kwizart at gmail.com>
+Cc: Kukjin Kim <kgene at kernel.org>
+Signed-off-by: Thierry Reding <treding at nvidia.com>
+---
+Changes in v2:
+- drop reference to struct device_node
+
+ drivers/iommu/exynos-iommu.c | 7 +++++++
+ 1 file changed, 7 insertions(+)
+
+diff --git a/drivers/iommu/exynos-iommu.c b/drivers/iommu/exynos-iommu.c
+index 7ce52737c7a1..dc14fec4ede1 100644
+--- a/drivers/iommu/exynos-iommu.c
++++ b/drivers/iommu/exynos-iommu.c
+@@ -1186,8 +1186,15 @@ static const struct iommu_ops exynos_iommu_ops = {
+
+ static int __init exynos_iommu_init(void)
+ {
++ struct device_node *np;
+ int ret;
+
++ np = of_find_matching_node(NULL, sysmmu_of_match);
++ if (!np)
++ return 0;
++
++ of_node_put(np);
++
+ lv2table_kmem_cache = kmem_cache_create("exynos-iommu-lv2table",
+ LV2TABLE_SIZE, LV2TABLE_SIZE, 0, NULL);
+ if (!lv2table_kmem_cache) {
+--
+2.1.3
diff --git a/arm-fix-iommu-omap.patch b/arm-fix-iommu-omap.patch
new file mode 100644
index 0000000..cf864ae
--- /dev/null
+++ b/arm-fix-iommu-omap.patch
@@ -0,0 +1,48 @@
+From: Thierry Reding <treding at nvidia.com>
+
+The OMAP IOMMU driver unconditionally executes code and registers a
+struct iommu_ops with the platform bus irrespective of whether it runs
+on an OMAP SoC or not. This causes problems in multi-platform kernels
+where drivers for other SoCs will no longer be able to register their
+own struct iommu_ops or even try to use a struct iommu_ops for an IOMMU
+that obviously isn't there.
+
+The smallest fix I could think of is to check for the existence of any
+OMAP IOMMU devices in the device tree and skip initialization otherwise.
+
+This fixes a problem on Tegra20 where the DRM driver will try to use the
+obviously non-existent OMAP IOMMU.
+
+Reported-by: Nicolas Chauvet <kwizart at gmail.com>
+Cc: Tony Lindgren <tony at atomide.com>
+Cc: Suman Anna <s-anna at ti.com>
+Cc: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
+Signed-off-by: Thierry Reding <treding at nvidia.com>
+---
+Changes in v2:
+- do not fix up module exit function since it's dead code
+- drop reference to struct device_node
+
+ drivers/iommu/omap-iommu.c | 7 +++++++
+ 1 file changed, 7 insertions(+)
+
+diff --git a/drivers/iommu/omap-iommu.c b/drivers/iommu/omap-iommu.c
+index f59f857b702e..a4ba851825c2 100644
+--- a/drivers/iommu/omap-iommu.c
++++ b/drivers/iommu/omap-iommu.c
+@@ -1376,6 +1376,13 @@ static int __init omap_iommu_init(void)
+ struct kmem_cache *p;
+ const unsigned long flags = SLAB_HWCACHE_ALIGN;
+ size_t align = 1 << 10; /* L2 pagetable alignement */
++ struct device_node *np;
++
++ np = of_find_matching_node(NULL, omap_iommu_of_match);
++ if (!np)
++ return 0;
++
++ of_node_put(np);
+
+ p = kmem_cache_create("iopte_cache", IOPTE_TABLE_SIZE, align, flags,
+ iopte_cachep_ctor);
+--
+2.1.3
diff --git a/arm-fix-iommu-rockchip.patch b/arm-fix-iommu-rockchip.patch
new file mode 100644
index 0000000..def4a0d
--- /dev/null
+++ b/arm-fix-iommu-rockchip.patch
@@ -0,0 +1,51 @@
+From: Thierry Reding <treding at nvidia.com>
+
+The Rockchip IOMMU driver unconditionally executes code and registers a
+struct iommu_ops with the platform bus irrespective of whether it runs
+on a Rockchip SoC or not. This causes problems in multi-platform kernels
+where drivers for other SoCs will no longer be able to register their
+own struct iommu_ops or even try to use a struct iommu_ops for an IOMMU
+that obviously isn't there.
+
+The smallest fix I could think of is to check for the existence of any
+Rockchip IOMMU devices in the device tree and skip initialization
+otherwise.
+
+This fixes a problem on Tegra20 where the DRM driver will try to use the
+obviously non-existent Rockchip IOMMU.
+Reported-by: Nicolas Chauvet <kwizart at gmail.com>
+Cc: Heiko Stuebner <heiko at sntech.de>
+Cc: Daniel Kurtz <djkurtz at chromium.org>
+Reviewed-by: Heiko Stuebner <heiko at sntech.de>
+Tested-by: Heiko Stuebner <heiko at sntech.de>
+Signed-off-by: Thierry Reding <treding at nvidia.com>
+---
+Changes in v2:
+- do not fix up module exit function since it's dead code
+- drop reference to struct device_node
+
+ drivers/iommu/rockchip-iommu.c | 7 +++++++
+ 1 file changed, 7 insertions(+)
+
+diff --git a/drivers/iommu/rockchip-iommu.c b/drivers/iommu/rockchip-iommu.c
+index 6a8b1ec4a48a..9f74fddcd304 100644
+--- a/drivers/iommu/rockchip-iommu.c
++++ b/drivers/iommu/rockchip-iommu.c
+@@ -1015,8 +1015,15 @@ static struct platform_driver rk_iommu_driver = {
+
+ static int __init rk_iommu_init(void)
+ {
++ struct device_node *np;
+ int ret;
+
++ np = of_find_matching_node(NULL, rk_iommu_dt_ids);
++ if (!np)
++ return 0;
++
++ of_node_put(np);
++
+ ret = bus_set_iommu(&platform_bus_type, &rk_iommu_ops);
+ if (ret)
+ return ret;
+--
+2.1.3
diff --git a/kernel.spec b/kernel.spec
index 2d62fa8..496ca99 100644
--- a/kernel.spec
+++ b/kernel.spec
@@ -590,6 +590,11 @@ Patch21026: pinctrl-pinctrl-single-must-be-initialized-early.patch
Patch21028: arm-i.MX6-Utilite-device-dtb.patch
+# IOMMU crash fixes - https://lists.linuxfoundation.org/pipermail/iommu/2015-February/012329.html
+Patch21030: arm-fix-iommu-omap.patch
+Patch21031: arm-fix-iommu-exynos.patch
+Patch21032: arm-fix-iommu-rockchip.patch
+
Patch21100: arm-highbank-l2-reverts.patch
#rhbz 754518
@@ -1246,6 +1251,10 @@ ApplyPatch pinctrl-pinctrl-single-must-be-initialized-early.patch
ApplyPatch arm-i.MX6-Utilite-device-dtb.patch
+ApplyPatch arm-fix-iommu-omap.patch
+ApplyPatch arm-fix-iommu-exynos.patch
+ApplyPatch arm-fix-iommu-rockchip.patch
+
ApplyPatch arm-highbank-l2-reverts.patch
#
--
1.7.2.1
_______________________________________________
kernel mailing list
kernel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/kernel
