[Xen-devel] [PATCH] x86: avoid HPET use on certain Intel platforms

* [Xen-devel] [PATCH] x86: avoid HPET use on certain Intel platforms
@ 2019-11-22 12:46 Jan Beulich
  2019-11-22 12:50 ` Andrew Cooper
  0 siblings, 1 reply; 8+ messages in thread
From: Jan Beulich @ 2019-11-22 12:46 UTC (permalink / raw)
  To: xen-devel@lists.xenproject.org
  Cc: Juergen Gross, Andrew Cooper, Wei Liu, Roger Pau Monné

Linux commit fc5db58539b49351e76f19817ed1102bf7c712d0 says

"Some Coffee Lake platforms have a skewed HPET timer once the SoCs entered
 PC10, which in consequence marks TSC as unstable because HPET is used as
 watchdog clocksource for TSC."

Follow this for Xen as well. Looking at its patch context made me notice
they have a pre-existing quirk for Bay Trail as well. The comment there,
however, points at a Cherry Trail document. Looking at the datasheets of
both, there appear to be similar issues, so go beyond Linux'es coverage
and exclude both. Also key the disable on the PCI IDs of the actual
affected devices, rather than those of 00:00.0.

Apply the workarounds only when the use of HPET was not explicitly
requested on the command line and when use of (deep) C-states was not
disabled.

Adjust a few types in touched or nearby code at the same time.

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/time.c
+++ b/xen/arch/x86/time.c
@@ -18,6 +18,7 @@
 #include <xen/timer.h>
 #include <xen/smp.h>
 #include <xen/irq.h>
+#include <xen/pci_ids.h>
 #include <xen/softirq.h>
 #include <xen/efi.h>
 #include <xen/cpuidle.h>
@@ -367,12 +368,41 @@ static u64 read_hpet_count(void)
     return hpet_read32(HPET_COUNTER);
 }
 
-static s64 __init init_hpet(struct platform_timesource *pts)
+static int64_t __init init_hpet(struct platform_timesource *pts)
 {
-    u64 hpet_rate = hpet_setup(), start;
-    u32 count, target;
+    uint64_t hpet_rate, start;
+    uint32_t count, target;
 
-    if ( hpet_rate == 0 )
+    if ( hpet_address && strcmp(opt_clocksource, pts->id) &&
+         cpuidle_using_deep_cstate() )
+    {
+        if ( pci_conf_read16(PCI_SBDF(0, 0, 0x1f, 0),
+                             PCI_VENDOR_ID) == PCI_VENDOR_ID_INTEL )
+            switch ( pci_conf_read16(PCI_SBDF(0, 0, 0x1f, 0), PCI_DEVICE_ID) )
+            {
+            /* HPET on Bay Trail platforms will halt in deep C states. */
+            case 0x0f1c:
+            /* HPET on Cherry Trail platforms will halt in deep C states. */
+            case 0x229c:
+                hpet_address = 0;
+                break;
+            }
+
+        /*
+         * Some Coffee Lake platforms have a skewed HPET timer once the SoCs
+         * entered PC10.
+         */
+        if ( pci_conf_read16(PCI_SBDF(0, 0, 0, 0),
+                             PCI_VENDOR_ID) == PCI_VENDOR_ID_INTEL &&
+             pci_conf_read16(PCI_SBDF(0, 0, 0, 0),
+                             PCI_DEVICE_ID) == 0x3ec4 )
+            hpet_address = 0;
+
+        if ( !hpet_address )
+            printk("Disabling HPET for being unreliable\n");
+    }
+
+    if ( (hpet_rate = hpet_setup()) == 0 )
         return 0;
 
     pts->frequency = hpet_rate;

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

^ permalink raw reply	[flat|nested] 8+ messages in thread