The dev_iommu_priv_set() must be called after probe_device(). This fixes
a NULL pointer deference bug when booting a system with kernel cmdline
"intel_iommu=on,igfx_off", where the dev_iommu_priv_set() is abused.
The following stacktrace was produced:
[ 0.000000] Command line: BOOT_IMAGE=/isolinux/bzImage console=tty1 intel_iommu=on,igfx_off
...
[ 3.341682] DMAR: Host address width 39
[ 3.341684] DMAR: DRHD base: 0x000000fed90000 flags: 0x0
[ 3.341702] DMAR: dmar0: reg_base_addr fed90000 ver 1:0 cap 1c0000c40660462 ecap 19e2ff0505e
[ 3.341705] DMAR: DRHD base: 0x000000fed91000 flags: 0x1
[ 3.341711] DMAR: dmar1: reg_base_addr fed91000 ver 1:0 cap d2008c40660462 ecap f050da
[ 3.341713] DMAR: RMRR base: 0x0000009aa9f000 end: 0x0000009aabefff
[ 3.341716] DMAR: RMRR base: 0x0000009d000000 end: 0x0000009f7fffff
[ 3.341726] DMAR: No ATSR found
[ 3.341772] BUG: kernel NULL pointer dereference, address: 0000000000000038
[ 3.341774] #PF: supervisor write access in kernel mode
[ 3.341776] #PF: error_code(0x0002) - not-present page
[ 3.341777] PGD 0 P4D 0
[ 3.341780] Oops: 0002 [#1] SMP PTI
[ 3.341783] CPU: 1 PID: 1 Comm: swapper/0 Not tainted 5.9.0-devel+ #2
[ 3.341785] Hardware name: LENOVO 20HGS0TW00/20HGS0TW00, BIOS N1WET46S (1.25s ) 03/30/2018
[ 3.341790] RIP: 0010:intel_iommu_init+0xed0/0x1136
[ 3.341792] Code: fe e9 61 02 00 00 bb f4 ff ff ff e9 57 02 00 00 48 63 d1 48 c1 e2 04 48
03 50 20 48 8b 12 48 85 d2 74 0b 48 8b 92 d0 02 00 00 48 89 7a 38 ff c1
e9 15 f5 ff ff 48 c7 c7 60 99 ac a7 49 c7 c7 a0
[ 3.341796] RSP: 0000:ffff96d180073dd0 EFLAGS: 00010282
[ 3.341798] RAX: ffff8c91037a7d20 RBX: 0000000000000000 RCX: 0000000000000000
[ 3.341800] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffffffffffffff
[ 3.341802] RBP: ffff96d180073e90 R08: 0000000000000001 R09: ffff8c91039fe3c0
[ 3.341804] R10: 0000000000000226 R11: 0000000000000226 R12: 000000000000000b
[ 3.341806] R13: ffff8c910367c650 R14: ffffffffa8426d60 R15: 0000000000000000
[ 3.341808] FS: 0000000000000000(0000) GS:ffff8c9107480000(0000) knlGS:0000000000000000
[ 3.341810] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 3.341812] CR2: 0000000000000038 CR3: 00000004b100a001 CR4: 00000000003706e0
[ 3.341814] Call Trace:
[ 3.341820] ? _raw_spin_unlock_irqrestore+0x1f/0x30
[ 3.341824] ? call_rcu+0x10e/0x320
[ 3.341828] ? trace_hardirqs_on+0x2c/0xd0
[ 3.341831] ? rdinit_setup+0x2c/0x2c
[ 3.341834] ? e820__memblock_setup+0x8b/0x8b
[ 3.341836] pci_iommu_init+0x16/0x3f
[ 3.341839] do_one_initcall+0x46/0x1e4
[ 3.341842] kernel_init_freeable+0x169/0x1b2
[ 3.341845] ? rest_init+0x9f/0x9f
[ 3.341847] kernel_init+0xa/0x101
[ 3.341849] ret_from_fork+0x22/0x30
[ 3.341851] Modules linked in:
[ 3.341854] CR2: 0000000000000038
[ 3.341860] ---[ end trace 3653722a6f936f18 ]---
Fixes: 01b9d4e21148c ("iommu/vt-d: Use dev_iommu_priv_get/set()")
Reported-by: Torsten Hilbrich <[email protected]>
Reported-by: Wendy Wang <[email protected]>
Link: https://lore.kernel.org/linux-iommu/[email protected]/
Signed-off-by: Lu Baolu <[email protected]>
---
drivers/iommu/intel/iommu.c | 100 ++++++++++++++++++++----------------
1 file changed, 55 insertions(+), 45 deletions(-)
diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
index 50431c7b2e71..777b9be60a0e 100644
--- a/drivers/iommu/intel/iommu.c
+++ b/drivers/iommu/intel/iommu.c
@@ -366,7 +366,6 @@ static int iommu_skip_te_disable;
int intel_iommu_gfx_mapped;
EXPORT_SYMBOL_GPL(intel_iommu_gfx_mapped);
-#define DUMMY_DEVICE_DOMAIN_INFO ((struct device_domain_info *)(-1))
#define DEFER_DEVICE_DOMAIN_INFO ((struct device_domain_info *)(-2))
struct device_domain_info *get_domain_info(struct device *dev)
{
@@ -376,8 +375,7 @@ struct device_domain_info *get_domain_info(struct device *dev)
return NULL;
info = dev_iommu_priv_get(dev);
- if (unlikely(info == DUMMY_DEVICE_DOMAIN_INFO ||
- info == DEFER_DEVICE_DOMAIN_INFO))
+ if (unlikely(info == DEFER_DEVICE_DOMAIN_INFO))
return NULL;
return info;
@@ -773,11 +771,6 @@ struct context_entry *iommu_context_addr(struct intel_iommu *iommu, u8 bus,
return &context[devfn];
}
-static int iommu_dummy(struct device *dev)
-{
- return dev_iommu_priv_get(dev) == DUMMY_DEVICE_DOMAIN_INFO;
-}
-
static bool attach_deferred(struct device *dev)
{
return dev_iommu_priv_get(dev) == DEFER_DEVICE_DOMAIN_INFO;
@@ -810,6 +803,53 @@ is_downstream_to_pci_bridge(struct device *dev, struct device *bridge)
return false;
}
+static bool quirk_ioat_snb_local_iommu(struct pci_dev *pdev)
+{
+ struct dmar_drhd_unit *drhd;
+ u32 vtbar;
+ int rc;
+
+ /* We know that this device on this chipset has its own IOMMU.
+ * If we find it under a different IOMMU, then the BIOS is lying
+ * to us. Hope that the IOMMU for this device is actually
+ * disabled, and it needs no translation...
+ */
+ rc = pci_bus_read_config_dword(pdev->bus, PCI_DEVFN(0, 0), 0xb0, &vtbar);
+ if (rc) {
+ /* "can't" happen */
+ dev_info(&pdev->dev, "failed to run vt-d quirk\n");
+ return false;
+ }
+ vtbar &= 0xffff0000;
+
+ /* we know that the this iommu should be at offset 0xa000 from vtbar */
+ drhd = dmar_find_matched_drhd_unit(pdev);
+ if (!drhd || drhd->reg_base_addr - vtbar != 0xa000) {
+ pr_warn_once(FW_BUG "BIOS assigned incorrect VT-d unit for Intel(R) QuickData Technology device\n");
+ add_taint(TAINT_FIRMWARE_WORKAROUND, LOCKDEP_STILL_OK);
+ return true;
+ }
+
+ return false;
+}
+
+static bool iommu_is_dummy(struct intel_iommu *iommu, struct device *dev)
+{
+ if (!iommu || iommu->drhd->ignored)
+ return true;
+
+ if (dev_is_pci(dev)) {
+ struct pci_dev *pdev = to_pci_dev(dev);
+
+ if (pdev->vendor == PCI_VENDOR_ID_INTEL &&
+ pdev->device == PCI_DEVICE_ID_INTEL_IOAT_SNB &&
+ quirk_ioat_snb_local_iommu(pdev))
+ return true;
+ }
+
+ return false;
+}
+
struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn)
{
struct dmar_drhd_unit *drhd = NULL;
@@ -819,7 +859,7 @@ struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn)
u16 segment = 0;
int i;
- if (!dev || iommu_dummy(dev))
+ if (!dev)
return NULL;
if (dev_is_pci(dev)) {
@@ -836,7 +876,7 @@ struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn)
dev = &ACPI_COMPANION(dev)->dev;
rcu_read_lock();
- for_each_active_iommu(iommu, drhd) {
+ for_each_iommu(iommu, drhd) {
if (pdev && segment != drhd->segment)
continue;
@@ -872,6 +912,9 @@ struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn)
}
iommu = NULL;
out:
+ if (iommu_is_dummy(iommu, dev))
+ iommu = NULL;
+
rcu_read_unlock();
return iommu;
@@ -2479,7 +2522,7 @@ struct dmar_domain *find_domain(struct device *dev)
{
struct device_domain_info *info;
- if (unlikely(attach_deferred(dev) || iommu_dummy(dev)))
+ if (unlikely(attach_deferred(dev)))
return NULL;
/* No lock here, assumes no domain exit in normal case */
@@ -4021,35 +4064,6 @@ static void __init iommu_exit_mempool(void)
iova_cache_put();
}
-static void quirk_ioat_snb_local_iommu(struct pci_dev *pdev)
-{
- struct dmar_drhd_unit *drhd;
- u32 vtbar;
- int rc;
-
- /* We know that this device on this chipset has its own IOMMU.
- * If we find it under a different IOMMU, then the BIOS is lying
- * to us. Hope that the IOMMU for this device is actually
- * disabled, and it needs no translation...
- */
- rc = pci_bus_read_config_dword(pdev->bus, PCI_DEVFN(0, 0), 0xb0, &vtbar);
- if (rc) {
- /* "can't" happen */
- dev_info(&pdev->dev, "failed to run vt-d quirk\n");
- return;
- }
- vtbar &= 0xffff0000;
-
- /* we know that the this iommu should be at offset 0xa000 from vtbar */
- drhd = dmar_find_matched_drhd_unit(pdev);
- if (!drhd || drhd->reg_base_addr - vtbar != 0xa000) {
- pr_warn_once(FW_BUG "BIOS assigned incorrect VT-d unit for Intel(R) QuickData Technology device\n");
- add_taint(TAINT_FIRMWARE_WORKAROUND, LOCKDEP_STILL_OK);
- dev_iommu_priv_set(&pdev->dev, DUMMY_DEVICE_DOMAIN_INFO);
- }
-}
-DECLARE_PCI_FIXUP_ENABLE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB, quirk_ioat_snb_local_iommu);
-
static void __init init_no_remapping_devices(void)
{
struct dmar_drhd_unit *drhd;
@@ -4081,12 +4095,8 @@ static void __init init_no_remapping_devices(void)
/* This IOMMU has *only* gfx devices. Either bypass it or
set the gfx_mapped flag, as appropriate */
drhd->gfx_dedicated = 1;
- if (!dmar_map_gfx) {
+ if (!dmar_map_gfx)
drhd->ignored = 1;
- for_each_active_dev_scope(drhd->devices,
- drhd->devices_cnt, i, dev)
- dev_iommu_priv_set(dev, DUMMY_DEVICE_DOMAIN_INFO);
- }
}
}
--
2.17.1
On 03.09.20 08:51, Lu Baolu wrote:
> The dev_iommu_priv_set() must be called after probe_device(). This fixes
> a NULL pointer deference bug when booting a system with kernel cmdline
> "intel_iommu=on,igfx_off", where the dev_iommu_priv_set() is abused.
[...]
>
> Fixes: 01b9d4e21148c ("iommu/vt-d: Use dev_iommu_priv_get/set()")
> Reported-by: Torsten Hilbrich <[email protected]>
> Reported-by: Wendy Wang <[email protected]>
> Link: https://lore.kernel.org/linux-iommu/[email protected]/
> Signed-off-by: Lu Baolu <[email protected]>
Tested-by: Torsten Hilbrich <[email protected]>
Regards,
Torsten
On Thu, Sep 03, 2020 at 02:51:32PM +0800, Lu Baolu wrote:
> ---
> drivers/iommu/intel/iommu.c | 100 ++++++++++++++++++++----------------
> 1 file changed, 55 insertions(+), 45 deletions(-)
Applied for v5.9, thanks.