File 65263470-AMD-IOMMU-flush-TLB-when-flushing-DTE.patch of Package xen.32846
# Commit 5fc98b97084a46884acef9320e643faf40d42212
# Date 2023-10-11 06:36:49 +0100
# Author Roger Pau Monne <roger.pau@citrix.com>
# Committer Andrew Cooper <andrew.cooper3@citrix.com>
iommu/amd-vi: flush IOMMU TLB when flushing the DTE
The caching invalidation guidelines from the AMD-Vi specification (48882—Rev
3.07-PUB—Oct 2022) seem to be misleading on some hardware, as devices will
malfunction (see stale DMA mappings) if some fields of the DTE are updated but
the IOMMU TLB is not flushed. This has been observed in practice on AMD
systems. Due to the lack of guidance from the currently published
specification this patch aims to increase the flushing done in order to prevent
device malfunction.
In order to fix, issue an INVALIDATE_IOMMU_PAGES command from
amd_iommu_flush_device(), flushing all the address space. Note this requires
callers to be adjusted in order to pass the DomID on the DTE previous to the
modification.
Some call sites don't provide a valid DomID to amd_iommu_flush_device() in
order to avoid the flush. That's because the device had address translations
disabled and hence the previous DomID on the DTE is not valid. Note the
current logic relies on the entity disabling address translations to also flush
the TLB of the in use DomID.
Device I/O TLB flushing when ATS are enabled is not covered by the current
change, as ATS usage is not security supported.
This is XSA-442 / CVE-2023-34326
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
--- a/xen/drivers/passthrough/amd/iommu.h
+++ b/xen/drivers/passthrough/amd/iommu.h
@@ -280,7 +280,8 @@ void amd_iommu_flush_pages(struct domain
unsigned int order);
void amd_iommu_flush_iotlb(u8 devfn, const struct pci_dev *pdev,
uint64_t gaddr, unsigned int order);
-void amd_iommu_flush_device(struct amd_iommu *iommu, uint16_t bdf);
+void amd_iommu_flush_device(struct amd_iommu *iommu, uint16_t bdf,
+ domid_t domid);
void amd_iommu_flush_intremap(struct amd_iommu *iommu, uint16_t bdf);
void amd_iommu_flush_all_caches(struct amd_iommu *iommu);
--- a/xen/drivers/passthrough/amd/iommu_cmd.c
+++ b/xen/drivers/passthrough/amd/iommu_cmd.c
@@ -362,10 +362,18 @@ void amd_iommu_flush_pages(struct domain
_amd_iommu_flush_pages(d, __dfn_to_daddr(dfn), order);
}
-void amd_iommu_flush_device(struct amd_iommu *iommu, uint16_t bdf)
+void amd_iommu_flush_device(struct amd_iommu *iommu, uint16_t bdf,
+ domid_t domid)
{
invalidate_dev_table_entry(iommu, bdf);
flush_command_buffer(iommu, 0);
+
+ /* Also invalidate IOMMU TLB entries when flushing the DTE. */
+ if ( domid != DOMID_INVALID )
+ {
+ invalidate_iommu_pages(iommu, INV_IOMMU_ALL_PAGES_ADDRESS, domid, 0);
+ flush_command_buffer(iommu, 0);
+ }
}
void amd_iommu_flush_intremap(struct amd_iommu *iommu, uint16_t bdf)
--- a/xen/drivers/passthrough/amd/iommu_guest.c
+++ b/xen/drivers/passthrough/amd/iommu_guest.c
@@ -385,7 +385,7 @@ static int do_completion_wait(struct dom
static int do_invalidate_dte(struct domain *d, cmd_entry_t *cmd)
{
- uint16_t gbdf, mbdf, req_id, gdom_id, hdom_id;
+ uint16_t gbdf, mbdf, req_id, gdom_id, hdom_id, prev_domid;
struct amd_iommu_dte *gdte, *mdte, *dte_base;
struct amd_iommu *iommu = NULL;
struct guest_iommu *g_iommu;
@@ -445,13 +445,14 @@ static int do_invalidate_dte(struct doma
req_id = get_dma_requestor_id(iommu->seg, mbdf);
dte_base = iommu->dev_table.buffer;
mdte = &dte_base[req_id];
+ prev_domid = mdte->domain_id;
spin_lock_irqsave(&iommu->lock, flags);
dte_set_gcr3_table(mdte, hdom_id, gcr3_mfn << PAGE_SHIFT, gv, glx);
spin_unlock_irqrestore(&iommu->lock, flags);
- amd_iommu_flush_device(iommu, req_id);
+ amd_iommu_flush_device(iommu, req_id, prev_domid);
return 0;
}
--- a/xen/drivers/passthrough/amd/iommu_init.c
+++ b/xen/drivers/passthrough/amd/iommu_init.c
@@ -1556,7 +1556,11 @@ static int _invalidate_all_devices(
req_id = ivrs_mappings[bdf].dte_requestor_id;
if ( iommu )
{
- amd_iommu_flush_device(iommu, req_id);
+ /*
+ * IOMMU TLB flush performed separately (see
+ * invalidate_all_domain_pages()).
+ */
+ amd_iommu_flush_device(iommu, req_id, DOMID_INVALID);
amd_iommu_flush_intremap(iommu, req_id);
}
}
--- a/xen/drivers/passthrough/amd/pci_amd_iommu.c
+++ b/xen/drivers/passthrough/amd/pci_amd_iommu.c
@@ -192,10 +192,13 @@ static int __must_check amd_iommu_setup_
spin_unlock_irqrestore(&iommu->lock, flags);
- amd_iommu_flush_device(iommu, req_id);
+ /* DTE didn't have DMA translations enabled, do not flush the TLB. */
+ amd_iommu_flush_device(iommu, req_id, DOMID_INVALID);
}
else if ( dte->pt_root != mfn_x(page_to_mfn(root_pg)) )
{
+ domid_t prev_domid = dte->domain_id;
+
/*
* Strictly speaking if the device is the only one with this requestor
* ID, it could be allowed to be re-assigned regardless of unity map
@@ -252,7 +255,7 @@ static int __must_check amd_iommu_setup_
spin_unlock_irqrestore(&iommu->lock, flags);
- amd_iommu_flush_device(iommu, req_id);
+ amd_iommu_flush_device(iommu, req_id, prev_domid);
}
else
spin_unlock_irqrestore(&iommu->lock, flags);
@@ -421,6 +424,8 @@ static void amd_iommu_disable_domain_dev
spin_lock_irqsave(&iommu->lock, flags);
if ( dte->tv || dte->v )
{
+ domid_t prev_domid = dte->domain_id;
+
/* See the comment in amd_iommu_setup_device_table(). */
dte->int_ctl = IOMMU_DEV_TABLE_INT_CONTROL_ABORTED;
smp_wmb();
@@ -439,7 +444,7 @@ static void amd_iommu_disable_domain_dev
spin_unlock_irqrestore(&iommu->lock, flags);
- amd_iommu_flush_device(iommu, req_id);
+ amd_iommu_flush_device(iommu, req_id, prev_domid);
AMD_IOMMU_DEBUG("Disable: device id = %#x, "
"domain = %d, paging mode = %d\n",
@@ -611,7 +616,8 @@ static int amd_iommu_add_device(u8 devfn
spin_unlock_irqrestore(&iommu->lock, flags);
- amd_iommu_flush_device(iommu, bdf);
+ /* DTE didn't have DMA translations enabled, do not flush the TLB. */
+ amd_iommu_flush_device(iommu, bdf, DOMID_INVALID);
}
if ( amd_iommu_reserve_domain_unity_map(
