[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[PATCH for-4.17 5/6] pci: do not disable memory decoding for devices


  • To: xen-devel@xxxxxxxxxxxxxxxxxxxx
  • From: Roger Pau Monne <roger.pau@xxxxxxxxxx>
  • Date: Thu, 20 Oct 2022 11:46:48 +0200
  • Arc-authentication-results: i=1; mx.microsoft.com 1; spf=pass smtp.mailfrom=citrix.com; dmarc=pass action=none header.from=citrix.com; dkim=pass header.d=citrix.com; arc=none
  • Arc-message-signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector9901; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=9VZPXsMkn9fam+VCRpitzkT4VDEDmOwmP7P5aKFDPN0=; b=hWOziJ2QIKMOBTjxZ2TIF3ewzHIDw2VxS5P3+wNL9G597BHst7lPh6bNsvjHtWBsfxgpl3AiPXzsR4hrJJTmvbOCr34ea0hco5Aywxqro3vdfsaEMhuW54sIqTlQzK/G7PXnsX/qk/2THLqn14QqCM2nns2KlemaujWf/4Nsyej766wTSkFmU6txBesclrZV9ZvjpVapA/PjuLWVU6H2oEYXb8tUk/E0/MZ8t2blBbkfFIast8Bd9JScrD1rU2+KPQT+bmokOFm80yco+IdnOUI7DGz/2RwycpXHgeh5X4SnRNJ4ozVpBemqrIR+MzSlPzBw2QICJVUl350cDv9dDA==
  • Arc-seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=ewCm6bVX+99uBKN8XM7piXtfqoikl5qLfZIyF2++Uo6deKJWtzGTOmigtD0Poz5Q4tmp8R1jXfXazv+YVTI6Bvvtv688SIZQAiBxns2tMJ8ONFi37rKw7eN+tUwzxHDzTUzf+CiB8NVCrehLs1tBjhZQyUt1rVVE/CeF4b4NTIYTTeZa11VyVsfPFv4AjaHunXQmlTuWLFR4/HReaNNaaVdGPGHQEIQPG9P55bRgPfgMywR9a5IeRstha2S29oy48/OTBqmo0iI6ldo1FPfx20DJTys6u+7OO0qwC4ZLN4kN3qxmZw6h+FRzPzbaVZd8DxdJxQp0uoxY+P8Njxdq2g==
  • Authentication-results: dkim=none (message not signed) header.d=none;dmarc=none action=none header.from=citrix.com;
  • Cc: Roger Pau Monne <roger.pau@xxxxxxxxxx>, Jan Beulich <jbeulich@xxxxxxxx>, Paul Durrant <paul@xxxxxxx>
  • Delivery-date: Thu, 20 Oct 2022 09:47:44 +0000
  • Ironport-data: A9a23:6Qrc/q+z+0uZkUohKUhuDrUDpX+TJUtcMsCJ2f8bNWPcYEJGY0x3z TAWWW7VbqzfN2P2Kdtwa4rnoEpTup/QyINrHFFt+yA8E34SpcT7XtnIdU2Y0wF+jCHgZBk+s 5hBMImowOQcFCK0SsKFa+C5xZVE/fjUAOC6UIYoAwgpLSd8UiAtlBl/rOAwh49skLCRDhiE/ Nj/uKUzAnf8s9JPGj9Suv/rRC9H5qyo4mpA5ARmPJingXeF/5UrJMNHTU2OByOQrrl8RoaSW +vFxbelyWLVlz9F5gSNy+uTnuUiG9Y+DCDW4pZkc/HKbitq/0Te5p0TJvsEAXq7vh3S9zxHJ HehgrTrIeshFvWkdO3wyHC0GQkmVUFN0OevzXRSLaV/ZqAJGpfh66wGMa04AWEX0sxbOzpPr N86EyEuNEuMn+iR5+34WOY506zPLOGzVG8ekldJ6GmDSM0AGNXESaiM4sJE1jAtgMwIBezZe 8cSdTtoalLHfgFLPVAUTpk5mY9EhFGmK2Ee9A3T+PVxvza7IA9ZidABNPLPfdOHX4NNl1uwr WPa5WXpRBodMbRzzBLVryzy17OSzUsXXqoWGYPmz+M33Ge0nEtUSz85bwO5jb6Q3xvWt9V3b hZ8FjAVhbg/8gmnQ8fwWzW8oWWYpVgMVtxICeo45QqRjK3O7G6xBGIJUzpAY9wOr9ItSHoh0 Vrht8vgGDhHoLCTD3WH+d+pQSiaPCEUKSoZY3YNSwIAu4PnuNtr0kyJSct/GqmoiNGzASv33 z2BsCk5gfMUkNIP0KK4u1vAhlpAu6T0c+L83S2PNkrN0++zTNXNi1CAgbQD0ct9EQ==
  • Ironport-hdrordr: A9a23:NsT4SqDTjDKO3KTlHeg3sceALOsnbusQ8zAXPh9KJCC9I/bzqy nxpp8mPH/P5wr5lktQ++xoX5PwO080lKQFmrX5WI3PYOCIghrNEGgP1+vfKnjbalTDH41mpN hdmtZFebrN5DFB5K6VgTVQUexQuOVvmJrY+ds2pE0dKD2CBZsQjDuQXW2gYzBLrUR9dOwEPa vZwvACiyureHwRYMj+Ln4ZX9Lbr9mOsJ79exYJCzMu9QHL1FqTmfbHOind+i1bfyJEwL8k/2 SAuwvl5p+7u/X+7hPHzWfc47lfhdOk4NpeA86njNQTN1zX+0+VTbUkf4fHkCE+oemp5lpvuN 7Qoy04N8A20H/VdnHdm2qZ5yDQlBIVr1Pyw16RhnXu5ebjQighNsZHjYVFNjPE9ksJprhHoe 529lPck6ASIQLLnSz76dSNfQptjFCIrX0rlvNWp2BDULEZdKRaoeUkjQ5o+a87bWzHAb0cYa hT5Jm23ocXTbraVQGSgoBX+q3iYpxpdS32AXTruaSuokprdT5CvgklLfck7wc9HaIGOud5Dt v/Q9VVfcl1P6krhIJGdZM8qJiMexvwaCOJFl6uCnLaM4xCE07xivfMkcYIDaeRCdc18Kc=
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

Commit 75cc460a1b added checks to ensure the position of the BARs from
PCI devices don't overlap with regions defined on the memory map.
When there's a collision memory decoding is left disabled for the
device, assuming that dom0 will reposition the BAR if necessary and
enable memory decoding.

While this would be the case for devices being used by dom0, devices
being used by the firmware itself that have no driver would usually be
left with memory decoding disabled by dom0 if that's the state dom0
found them in, and thus firmware trying to make use of them will not
function correctly.

The initial intent of 75cc460a1b was to prevent vPCI from creating
MMIO mappings on the dom0 p2m over regions that would otherwise
already have mappings established.  It's my view now that we likely
went too far with 75cc460a1b, and Xen disabling memory decoding of
devices (as buggy as they might be) is harmful, and reduces the set of
hardware on which Xen works.

This commits reverts most of 75cc460a1b, and instead adds checks to
vPCI in order to prevent misplaced BARs from being added to the
hardware domain p2m.  Signaling on whether BARs are mapped is tracked
in the vpci structure, so that misplaced BARs are not mapped, and thus
Xen won't attempt to unmap them when memory decoding is disabled.

This restores the behavior of Xen for PV dom0 to the state it was
previous to 75cc460a1b, while also introducing a more contained fix
for the vPCI BAR mapping issues.

Fixes: 75cc460a1b ('xen/pci: detect when BARs are not suitably positioned')
Signed-off-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>
---
AT Citrix we have a system with a device with the following BARs:

BAR [0xfe010, 0xfe010] -> in a EfiMemoryMappedIO region
BAR [0, 0x1fff] -> not positioned, outside host bridge window

And memory decoding enabled by the firmware.  With the current code
(or any of the previous fix proposals), Xen would still disable memory
decoding for the device, and the system will freeze when attempting to
set EFI vars.

I'm afraid the best solution to avoid regressions caused by 75cc460a1b
is the proposal here.
---
 xen/drivers/passthrough/pci.c | 69 -----------------------------------
 xen/drivers/vpci/header.c     | 22 ++++++++++-
 2 files changed, 20 insertions(+), 71 deletions(-)

diff --git a/xen/drivers/passthrough/pci.c b/xen/drivers/passthrough/pci.c
index 149f68bb6e..b42acb8d7c 100644
--- a/xen/drivers/passthrough/pci.c
+++ b/xen/drivers/passthrough/pci.c
@@ -233,9 +233,6 @@ static void check_pdev(const struct pci_dev *pdev)
      PCI_STATUS_REC_TARGET_ABORT | PCI_STATUS_REC_MASTER_ABORT | \
      PCI_STATUS_SIG_SYSTEM_ERROR | PCI_STATUS_DETECTED_PARITY)
     u16 val;
-    unsigned int nbars = 0, rom_pos = 0, i;
-    static const char warn[] = XENLOG_WARNING
-        "%pp disabled: %sBAR [%#lx, %#lx] overlaps with memory map\n";
 
     if ( command_mask )
     {
@@ -254,8 +251,6 @@ static void check_pdev(const struct pci_dev *pdev)
     switch ( pci_conf_read8(pdev->sbdf, PCI_HEADER_TYPE) & 0x7f )
     {
     case PCI_HEADER_TYPE_BRIDGE:
-        nbars = PCI_HEADER_BRIDGE_NR_BARS;
-        rom_pos = PCI_ROM_ADDRESS1;
         if ( !bridge_ctl_mask )
             break;
         val = pci_conf_read16(pdev->sbdf, PCI_BRIDGE_CONTROL);
@@ -272,75 +267,11 @@ static void check_pdev(const struct pci_dev *pdev)
         }
         break;
 
-    case PCI_HEADER_TYPE_NORMAL:
-        nbars = PCI_HEADER_NORMAL_NR_BARS;
-        rom_pos = PCI_ROM_ADDRESS;
-        break;
-
     case PCI_HEADER_TYPE_CARDBUS:
         /* TODO */
         break;
     }
 #undef PCI_STATUS_CHECK
-
-    /* Check if BARs overlap with other memory regions. */
-    val = pci_conf_read16(pdev->sbdf, PCI_COMMAND);
-    if ( !(val & PCI_COMMAND_MEMORY) || pdev->ignore_bars )
-        return;
-
-    pci_conf_write16(pdev->sbdf, PCI_COMMAND, val & ~PCI_COMMAND_MEMORY);
-    for ( i = 0; i < nbars; )
-    {
-        uint64_t addr, size;
-        unsigned int reg = PCI_BASE_ADDRESS_0 + i * 4;
-        int rc = 1;
-
-        if ( (pci_conf_read32(pdev->sbdf, reg) & PCI_BASE_ADDRESS_SPACE) !=
-             PCI_BASE_ADDRESS_SPACE_MEMORY )
-            goto next;
-
-        rc = pci_size_mem_bar(pdev->sbdf, reg, &addr, &size,
-                              (i == nbars - 1) ? PCI_BAR_LAST : 0);
-        if ( rc < 0 )
-            /* Unable to size, better leave memory decoding disabled. */
-            return;
-        if ( size && !pci_check_bar(pdev, maddr_to_mfn(addr),
-                                    maddr_to_mfn(addr + size - 1)) )
-        {
-            /*
-             * Return without enabling memory decoding if BAR position is not
-             * in IO suitable memory. Let the hardware domain re-position the
-             * BAR.
-             */
-            printk(warn,
-                   &pdev->sbdf, "", PFN_DOWN(addr), PFN_DOWN(addr + size - 1));
-            return;
-        }
-
- next:
-        ASSERT(rc > 0);
-        i += rc;
-    }
-
-    if ( rom_pos &&
-         (pci_conf_read32(pdev->sbdf, rom_pos) & PCI_ROM_ADDRESS_ENABLE) )
-    {
-        uint64_t addr, size;
-        int rc = pci_size_mem_bar(pdev->sbdf, rom_pos, &addr, &size,
-                                  PCI_BAR_ROM);
-
-        if ( rc < 0 )
-            return;
-        if ( size && !pci_check_bar(pdev, maddr_to_mfn(addr),
-                                    maddr_to_mfn(addr + size - 1)) )
-        {
-            printk(warn, &pdev->sbdf, "ROM ", PFN_DOWN(addr),
-                   PFN_DOWN(addr + size - 1));
-            return;
-        }
-    }
-
-    pci_conf_write16(pdev->sbdf, PCI_COMMAND, val);
 }
 
 static void apply_quirks(struct pci_dev *pdev)
diff --git a/xen/drivers/vpci/header.c b/xen/drivers/vpci/header.c
index eb9219a49a..4d7f8f4a30 100644
--- a/xen/drivers/vpci/header.c
+++ b/xen/drivers/vpci/header.c
@@ -121,7 +121,9 @@ static void modify_decoding(const struct pci_dev *pdev, 
uint16_t cmd,
         }
 
         if ( !rom_only &&
-             (bar->type != VPCI_BAR_ROM || header->rom_enabled) )
+             (bar->type != VPCI_BAR_ROM || header->rom_enabled) &&
+             pci_check_bar(pdev, _mfn(PFN_DOWN(bar->addr)),
+                           _mfn(PFN_DOWN(bar->addr + bar->size - 1))) )
             bar->enabled = map;
     }
 
@@ -234,9 +236,25 @@ static int modify_bars(const struct pci_dev *pdev, 
uint16_t cmd, bool rom_only)
 
         if ( !MAPPABLE_BAR(bar) ||
              (rom_only ? bar->type != VPCI_BAR_ROM
-                       : (bar->type == VPCI_BAR_ROM && !header->rom_enabled)) )
+                       : (bar->type == VPCI_BAR_ROM && !header->rom_enabled)) 
||
+             /* Skip BARs already in the requested state. */
+             bar->enabled == !!(cmd & PCI_COMMAND_MEMORY) )
             continue;
 
+        /*
+         * Only do BAR position checks for the hardware domain, for guests it's
+         * assumed that the hardware domain has changed the position of any
+         * problematic BARs.
+         */
+        if ( is_hardware_domain(pdev->domain) &&
+             !pci_check_bar(pdev, _mfn(start), _mfn(end)) )
+        {
+            printk(XENLOG_G_WARNING
+                   "%pp: not mapping BAR [%lx, %lx] invalid position\n",
+                   &pdev->sbdf, start, end);
+            continue;
+        }
+
         rc = rangeset_add_range(mem, start, end);
         if ( rc )
         {
-- 
2.37.3




 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.