[Xen-devel] [PATCH v2] pci: clear {host/guest}_maskall field on assign

[Roger Pau Monne <roger.pau@xxxxxxxxxx>]

The current implementation of host_maskall makes it sticky across
assign and deassign calls, which means that once a guest forces Xen to
set host_maskall the maskall bit is not going to be cleared until a
call to PHYSDEVOP_prepare_msix is performed. Such call however
shouldn't be part of the normal flow when doing PCI passthrough, and
hence the flag needs to be cleared when assigning in order to prevent
host_maskall being carried over from previous assignations.

Note that the entry maskbit is reset when the msix capability is
initialized, and the guest_maskall field is also cleared so that the
hardware value matches Xen's internal state (hardware maskall =
host_maskall | guest_maskall).

Also note that doing the reset of host_maskall there would allow the
guest to reset such field by enabling and disabling MSIX, which is not
intended.

Signed-off-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>
---
Cc: Chao Gao <chao.gao@xxxxxxxxx>
Cc: "Spassov, Stanislav" <stanspas@xxxxxxxxx>
Cc: Pasi Kärkkäinen <pasik@xxxxxx>
---
Chao, Stanislav, can you please check if this patch fixes your
issues?
---
Changes since v1:
 - Also set guest_maskall.
 - Place the code in a helper function in x86/msi.c
 - Add a comment to describe the expected state.
 - Test that maskall is not set on hardware.
---
 xen/arch/x86/msi.c            | 22 ++++++++++++++++++++++
 xen/drivers/passthrough/pci.c |  5 +++++
 xen/include/asm-x86/msi.h     |  1 +
 3 files changed, 28 insertions(+)

diff --git a/xen/arch/x86/msi.c b/xen/arch/x86/msi.c
index 76d4034c4f..c239a00fb1 100644
--- a/xen/arch/x86/msi.c
+++ b/xen/arch/x86/msi.c
@@ -1249,6 +1249,28 @@ void pci_cleanup_msi(struct pci_dev *pdev)
     msi_free_irqs(pdev);
 }
 
+int pci_reset_msix_state(struct pci_dev *pdev)
+{
+    unsigned int pos = pci_find_cap_offset(pdev->seg, pdev->bus, 
pdev->sbdf.dev,
+                                           pdev->sbdf.fn, PCI_CAP_ID_MSIX);
+
+    ASSERT(pos);
+    /*
+     * Xen expects the device state to be the after reset one, and hence
+     * host_maskall = guest_maskall = false and all entries should have the
+     * mask bit set. Test that the maskall bit is not set, having it set could
+     * signal that the device hasn't been reset properly.
+     */
+    if ( pci_conf_read16(pdev->sbdf, msix_control_reg(pos)) &
+         PCI_MSIX_FLAGS_MASKALL )
+        return -EBUSY;
+
+    pdev->msix->host_maskall = false;
+    pdev->msix->guest_maskall = false;
+
+    return 0;
+}
+
 int pci_msi_conf_write_intercept(struct pci_dev *pdev, unsigned int reg,
                                  unsigned int size, uint32_t *data)
 {
diff --git a/xen/drivers/passthrough/pci.c b/xen/drivers/passthrough/pci.c
index 90ccb8370b..bdcc482d81 100644
--- a/xen/drivers/passthrough/pci.c
+++ b/xen/drivers/passthrough/pci.c
@@ -1505,7 +1505,12 @@ static int assign_device(struct domain *d, u16 seg, u8 
bus, u8 devfn, u32 flag)
     }
 
     if ( pdev->msix )
+    {
+        rc = pci_reset_msix_state(pdev);
+        if ( rc )
+            goto done;
         msixtbl_init(d);
+    }
 
     pdev->fault.count = 0;
 
diff --git a/xen/include/asm-x86/msi.h b/xen/include/asm-x86/msi.h
index d0b0045d0d..6e35713ec7 100644
--- a/xen/include/asm-x86/msi.h
+++ b/xen/include/asm-x86/msi.h
@@ -92,6 +92,7 @@ extern int __setup_msi_irq(struct irq_desc *, struct msi_desc 
*,
 extern void teardown_msi_irq(int irq);
 extern int msi_free_vector(struct msi_desc *entry);
 extern int pci_restore_msi_state(struct pci_dev *pdev);
+extern int pci_reset_msix_state(struct pci_dev *pdev);
 
 struct msi_desc {
        struct msi_attrib {
-- 
2.23.0


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel