Skip to contentSkip to conversation

Make PAT handling less brittle

PATCH Superseded Demi Marie Obenour 3 participants 28 messages Dec 20, 2022 – Dec 22, 2022

v5 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
#0 lore raw

v5 cover Make PAT handling less brittle

While working on Qubes OS Marek found out that there were some PAT hacks in the Linux i195 driver. I decided to make Xen use Linux’s PAT to see if it solved the graphics glitches that were observed; it did. This required a substantial amount of preliminary work that is useful even without using Linux’s PAT.

Patches 1 through 9 are the preliminary work and I would like them to be accepted into upstream Xen. Patch 9 does break ABI by rejecting the unused PAT entries, but this will only impact buggy PV guests and can be disabled with a Xen command-line option. Patch 10 actually switches to Linux’s PAT and is NOT intended to be merged (at least for now) as it would at a minimum break migration of PV guests from hosts that do not have the patch.

Only patches 9 and 10 actually change Xen’s observable behavior. Patches 1, 2, and 7 are prerequisites, and patches 3 through 6 are cleanups. Patch 8 makes changing the PAT much less error-prone, as problems with the PAT or with the associated _PAGE_* constants will be detected at compile time.

Demi Marie Obenour (10):
x86: Add memory type constants
x86/mm: Avoid hard-coding PAT in get_page_from_l1e()
x86: Replace PAT_* with X86_MT_*
x86: Replace MTRR_* constants with X86_MT_* constants
x86: Replace EPT_EMT_* constants with X86_MT_*
x86: Remove MEMORY_NUM_TYPES and NO_HARDCODE_MEM_TYPE
x86: Derive XEN_MSR_PAT from its individual entries x86/mm: make code robust to future PAT changes x86/mm: Reject invalid cacheability in PV guests by default
x86: Use Linux's PAT

 xen/arch/x86/cpu/mtrr/generic.c         |  10 +-
 xen/arch/x86/cpu/mtrr/main.c            |  26 +++---
 xen/arch/x86/e820.c                     |   4 +-
 xen/arch/x86/hvm/hvm.c                  |  12 +--
 xen/arch/x86/hvm/mtrr.c                 | 100 ++++++++++----------
 xen/arch/x86/hvm/vmx/vmcs.c             |   2 +-
 xen/arch/x86/hvm/vmx/vmx.c              |  18 ++--
 xen/arch/x86/include/asm/hvm/vmx/vmcs.h |   2 +-
 xen/arch/x86/include/asm/hvm/vmx/vmx.h  |   9 --
 xen/arch/x86/include/asm/mtrr.h         |  22 +----
 xen/arch/x86/include/asm/page.h         |   4 +-
 xen/arch/x86/include/asm/processor.h    |  11 ++-
 xen/arch/x86/include/asm/x86-defns.h    |  11 +++
 xen/arch/x86/mm.c                       | 118 ++++++++++++++++++++++--
 xen/arch/x86/mm/hap/nested_ept.c        |   4 +-
 xen/arch/x86/mm/p2m-ept.c               |  51 +++++-----
 xen/arch/x86/mm/shadow/multi.c          |   8 +-
 17 files changed, 253 insertions(+), 159 deletions(-)
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#1 lore raw

PATCH v5 1/10 x86: Add memory type constants

These are not currently used, so there is no functional change. Future patches will use these constants.

---
Changes since v4:
- Add Jan Beulich’s Acked-by

Changes since v3:
- Name the reserved values X86_MT_RSVD_2 and X86_MT_RSVD_3, to match the architectural values of 0x02 and 0x03.

Changes since v2:
- Avoid using _AC where not required.
- Define reserved memory types inline

---
 xen/arch/x86/include/asm/x86-defns.h | 11 +++++++++++
 1 file changed, 11 insertions(+)
xen/arch/x86/include/asm/x86-defns.h +11 -0
index 28628807cb9897cf6fa8e266f71f5f220813984d..42b5f382d438d21ac97b6438e8c810c7b964cf6d 100644
@@ -153,4 +153,15 @@
      (1u << X86_EXC_AC) | (1u << X86_EXC_CP) |                      \
      (1u << X86_EXC_VC) | (1u << X86_EXC_SX))
 
+/* Memory types */
+#define X86_MT_UC     0x00 /* uncachable */
+#define X86_MT_WC     0x01 /* write-combined */
+#define X86_MT_RSVD_2 0x02 /* reserved */
+#define X86_MT_RSVD_3 0x03 /* reserved */
+#define X86_MT_WT     0x04 /* write-through */
+#define X86_MT_WP     0x05 /* write-protect */
+#define X86_MT_WB     0x06 /* write-back */
+#define X86_MT_UCM    0x07 /* UC- */
+#define X86_NUM_MT    0x08
+
 #endif	/* __XEN_X86_DEFNS_H__ */
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#2 lore raw

PATCH v5 2/10 x86/mm: Avoid hard-coding PAT in get_page_from_l1e()

In reply to Demi Marie Obenour

get_page_from_l1e() relied on Xen's choice of PAT, which is brittle in the face of future PAT changes. Instead, compute the actual cacheability used by the CPU and switch on that, as this will work no matter what PAT Xen uses.

No functional change intended. This code is itself questionable and may be removed in the future, but removing it would be an observable behavior change and so is out of scope for this patch series.

---
Changes since v4:
- Do not add new pte_flags_to_cacheability() helper, as this code may be removed in the near future and so adding a new helper for it is a bad idea.
- Do not BUG() in the event of an unexpected cacheability. This cannot happen, but it is simpler to force such types to UC than to prove that the BUG() is not reachable.

Changes since v3:
- Compute and use the actual cacheability as seen by the processor.

Changes since v2:
- Improve commit message.

---
 xen/arch/x86/mm.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)
xen/arch/x86/mm.c +8 -6
index 78b1972e4170cacccc9c37c6e64e76e66a7da87f..dba6c77ef2f7ed7fcb7f7e526583ccadd35e62cc 100644
@@ -959,14 +959,16 @@ get_page_from_l1e(
             flip = _PAGE_RW;
         }
 
-        switch ( l1f & PAGE_CACHE_ATTRS )
+        switch ( 0xFF & (XEN_MSR_PAT >> (8 * pte_flags_to_cacheattr(l1f))) )
         {
-        case 0: /* WB */
-            flip |= _PAGE_PWT | _PAGE_PCD;
+        case X86_MT_UC:
+        case X86_MT_UCM:
+        case X86_MT_WC:
+            /* not cacheable */
             break;
-        case _PAGE_PWT: /* WT */
-        case _PAGE_PWT | _PAGE_PAT: /* WP */
-            flip |= _PAGE_PCD | (l1f & _PAGE_PAT);
+        default:
+            /* cacheable */
+            flip |= ((l1f & PAGE_CACHE_ATTRS) ^ _PAGE_UC);
             break;
         }
 
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
Reviewed-by on v5 2/10
#3 lore raw

Re: [PATCH v5 02/10] x86/mm: Avoid hard-coding PAT in get_page_from_l1e()

Quoted 43 lines · Demi Marie Obenour

get_page_from_l1e() relied on Xen's choice of PAT, which is brittle in the face of future PAT changes. Instead, compute the actual cacheability used by the CPU and switch on that, as this will work no matter what PAT Xen uses.

No functional change intended. This code is itself questionable and may be removed in the future, but removing it would be an observable behavior change and so is out of scope for this patch series.

---
Changes since v4:
- Do not add new pte_flags_to_cacheability() helper, as this code may be removed in the near future and so adding a new helper for it is a bad idea.
- Do not BUG() in the event of an unexpected cacheability. This cannot happen, but it is simpler to force such types to UC than to prove that the BUG() is not reachable.

Changes since v3:
- Compute and use the actual cacheability as seen by the processor.

Changes since v2:
- Improve commit message.

---
 xen/arch/x86/mm.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)
xen/arch/x86/mm.c +3 -3
index 78b1972e4170cacccc9c37c6e64e76e66a7da87f..dba6c77ef2f7ed7fcb7f7e526583ccadd35e62cc 100644
@@ -959,14 +959,16 @@ get_page_from_l1e(
             flip = _PAGE_RW;
         }
 
-        switch ( l1f & PAGE_CACHE_ATTRS )
+        switch ( 0xFF & (XEN_MSR_PAT >> (8 * pte_flags_to_cacheattr(l1f))) )
         {
-        case 0: /* WB */
-            flip |= _PAGE_PWT | _PAGE_PCD;
+        case X86_MT_UC:
+        case X86_MT_UCM:

xen/arch/x86/mm.c get_page_from_l1e( · line 972

Quoted 10 lines
xen/arch/x86/mm.c +5 -3
+        case X86_MT_WC:
+            /* not cacheable */
             break;
-        case _PAGE_PWT: /* WT */
-        case _PAGE_PWT | _PAGE_PAT: /* WP */
-            flip |= _PAGE_PCD | (l1f & _PAGE_PAT);
+        default:
+            /* cacheable */
+            flip |= ((l1f & PAGE_CACHE_ATTRS) ^ _PAGE_UC);
             break;

In v4 the comment here was "cacheable, force to UC". The latter aspect is quite relevant (and iirc also what Andrew had asked for to have as a comment). But with this now being the default case, the comment (in either this or the earlier form) would become stale. A forward compatible way of wording this would e.g. be "force any other type to UC". With an adjustment along these lines (which I think could be done while committing)

Jan

#4 lore raw

Re: [PATCH v5 02/10] x86/mm: Avoid hard-coding PAT in get_page_from_l1e()

Quoted 61 lines · Jan Beulich
Quoted 43 lines · Demi Marie Obenour

get_page_from_l1e() relied on Xen's choice of PAT, which is brittle in the face of future PAT changes. Instead, compute the actual cacheability used by the CPU and switch on that, as this will work no matter what PAT Xen uses.

No functional change intended. This code is itself questionable and may be removed in the future, but removing it would be an observable behavior change and so is out of scope for this patch series.

---
Changes since v4:
- Do not add new pte_flags_to_cacheability() helper, as this code may be removed in the near future and so adding a new helper for it is a bad idea.
- Do not BUG() in the event of an unexpected cacheability. This cannot happen, but it is simpler to force such types to UC than to prove that the BUG() is not reachable.

Changes since v3:
- Compute and use the actual cacheability as seen by the processor.

Changes since v2:
- Improve commit message.

---
 xen/arch/x86/mm.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)
xen/arch/x86/mm.c +3 -3
index 78b1972e4170cacccc9c37c6e64e76e66a7da87f..dba6c77ef2f7ed7fcb7f7e526583ccadd35e62cc 100644
@@ -959,14 +959,16 @@ get_page_from_l1e(
             flip = _PAGE_RW;
         }
 
-        switch ( l1f & PAGE_CACHE_ATTRS )
+        switch ( 0xFF & (XEN_MSR_PAT >> (8 * pte_flags_to_cacheattr(l1f))) )
         {
-        case 0: /* WB */
-            flip |= _PAGE_PWT | _PAGE_PCD;
+        case X86_MT_UC:
+        case X86_MT_UCM:

xen/arch/x86/mm.c get_page_from_l1e( · line 972

Quoted 10 lines
xen/arch/x86/mm.c +5 -3
+        case X86_MT_WC:
+            /* not cacheable */
             break;
-        case _PAGE_PWT: /* WT */
-        case _PAGE_PWT | _PAGE_PAT: /* WP */
-            flip |= _PAGE_PCD | (l1f & _PAGE_PAT);
+        default:
+            /* cacheable */
+            flip |= ((l1f & PAGE_CACHE_ATTRS) ^ _PAGE_UC);
             break;

In v4 the comment here was "cacheable, force to UC". The latter aspect is quite relevant (and iirc also what Andrew had asked for to have as a comment). But with this now being the default case, the comment (in either this or the earlier form) would become stale. A forward compatible way of wording this would e.g. be "force any other type to UC". With an adjustment along these lines (which I think could be done while committing)

Quoted 1 line

If you had replied signaling your consent (and perhaps the preferred by you wording), I would have committed this. Now it's going to be v6 afaic ...

Jan

#5 lore raw

Re: [PATCH v5 02/10] x86/mm: Avoid hard-coding PAT in get_page_from_l1e()

Quoted 64 lines · Jan Beulich
Quoted 61 lines · Jan Beulich
Quoted 43 lines · Demi Marie Obenour

get_page_from_l1e() relied on Xen's choice of PAT, which is brittle in the face of future PAT changes. Instead, compute the actual cacheability used by the CPU and switch on that, as this will work no matter what PAT Xen uses.

No functional change intended. This code is itself questionable and may be removed in the future, but removing it would be an observable behavior change and so is out of scope for this patch series.

---
Changes since v4:
- Do not add new pte_flags_to_cacheability() helper, as this code may be removed in the near future and so adding a new helper for it is a bad idea.
- Do not BUG() in the event of an unexpected cacheability. This cannot happen, but it is simpler to force such types to UC than to prove that the BUG() is not reachable.

Changes since v3:
- Compute and use the actual cacheability as seen by the processor.

Changes since v2:
- Improve commit message.

---
 xen/arch/x86/mm.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)
xen/arch/x86/mm.c +3 -3
index 78b1972e4170cacccc9c37c6e64e76e66a7da87f..dba6c77ef2f7ed7fcb7f7e526583ccadd35e62cc 100644
@@ -959,14 +959,16 @@ get_page_from_l1e(
             flip = _PAGE_RW;
         }
 
-        switch ( l1f & PAGE_CACHE_ATTRS )
+        switch ( 0xFF & (XEN_MSR_PAT >> (8 * pte_flags_to_cacheattr(l1f))) )
         {
-        case 0: /* WB */
-            flip |= _PAGE_PWT | _PAGE_PCD;
+        case X86_MT_UC:
+        case X86_MT_UCM:

xen/arch/x86/mm.c get_page_from_l1e( · line 972

Quoted 10 lines
xen/arch/x86/mm.c +5 -3
+        case X86_MT_WC:
+            /* not cacheable */
             break;
-        case _PAGE_PWT: /* WT */
-        case _PAGE_PWT | _PAGE_PAT: /* WP */
-            flip |= _PAGE_PCD | (l1f & _PAGE_PAT);
+        default:
+            /* cacheable */
+            flip |= ((l1f & PAGE_CACHE_ATTRS) ^ _PAGE_UC);
             break;

In v4 the comment here was "cacheable, force to UC". The latter aspect is quite relevant (and iirc also what Andrew had asked for to have as a comment). But with this now being the default case, the comment (in either this or the earlier form) would become stale. A forward compatible way of wording this would e.g. be "force any other type to UC". With an adjustment along these lines (which I think could be done while committing)

Quoted 1 line
Quoted 4 lines

If you had replied signaling your consent (and perhaps the preferred by you wording), I would have committed this. Now it's going to be v6 afaic ...

Jan

Sorry about that. "potentially cacheable, force to UC" is the wording I have planned for v6, along with "not cacheable, allow" in the other case. Feel free to go ahead and commit if you want.

Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#6 lore raw

PATCH v5 3/10 x86: Replace PAT_* with X86_MT_*

In reply to Demi Marie Obenour

This allows eliminating the former.

---
Changes since v2: Style adjustments

---
 xen/arch/x86/hvm/hvm.c          | 12 ++++----
 xen/arch/x86/hvm/mtrr.c         | 52 ++++++++++++++++-----------------
 xen/arch/x86/hvm/vmx/vmx.c      | 16 +++++-----
 xen/arch/x86/include/asm/mtrr.h | 12 +-------
 xen/arch/x86/mm/p2m-ept.c       |  4 +--
 xen/arch/x86/mm/shadow/multi.c  |  4 +--
 6 files changed, 45 insertions(+), 55 deletions(-)
xen/arch/x86/hvm/hvm.c +6 -6
index ae4368ec4b338cf8c6cb14d383f612c91c98e800..00b3fa56e25e2934e2870e11fd19b120daff2715 100644
@@ -307,12 +307,12 @@ int hvm_set_guest_pat(struct vcpu *v, uint64_t guest_pat)
     for ( i = 0, tmp = guest_pat; i < 8; i++, tmp >>= 8 )
         switch ( tmp & 0xff )
         {
-        case PAT_TYPE_UC_MINUS:
-        case PAT_TYPE_UNCACHABLE:
-        case PAT_TYPE_WRBACK:
-        case PAT_TYPE_WRCOMB:
-        case PAT_TYPE_WRPROT:
-        case PAT_TYPE_WRTHROUGH:
+        case X86_MT_UCM:
+        case X86_MT_UC:
+        case X86_MT_WB:
+        case X86_MT_WC:
+        case X86_MT_WP:
+        case X86_MT_WT:
             break;
         default:
             return 0;
xen/arch/x86/hvm/mtrr.c +26 -26

124 lines not on this page · raw message

xen/arch/x86/hvm/vmx/vmx.c +8 -8
index 7c81b80710f99e08fe8291d3e413c449322b777d..b543c3983d77ae807e8bd97330691a79d8d39bae 100644
@@ -1231,14 +1231,14 @@ static void cf_check vmx_handle_cd(struct vcpu *v, unsigned long value)
              * memory type are all UC.
              */
             u64 uc_pat =
-                ((uint64_t)PAT_TYPE_UNCACHABLE)       |       /* PAT0 */
-                ((uint64_t)PAT_TYPE_UNCACHABLE << 8)  |       /* PAT1 */
-                ((uint64_t)PAT_TYPE_UNCACHABLE << 16) |       /* PAT2 */
-                ((uint64_t)PAT_TYPE_UNCACHABLE << 24) |       /* PAT3 */
-                ((uint64_t)PAT_TYPE_UNCACHABLE << 32) |       /* PAT4 */
-                ((uint64_t)PAT_TYPE_UNCACHABLE << 40) |       /* PAT5 */
-                ((uint64_t)PAT_TYPE_UNCACHABLE << 48) |       /* PAT6 */
-                ((uint64_t)PAT_TYPE_UNCACHABLE << 56);        /* PAT7 */
+                ((uint64_t)X86_MT_UC)       |       /* PAT0 */
+                ((uint64_t)X86_MT_UC << 8)  |       /* PAT1 */
+                ((uint64_t)X86_MT_UC << 16) |       /* PAT2 */
+                ((uint64_t)X86_MT_UC << 24) |       /* PAT3 */
+                ((uint64_t)X86_MT_UC << 32) |       /* PAT4 */
+                ((uint64_t)X86_MT_UC << 40) |       /* PAT5 */
+                ((uint64_t)X86_MT_UC << 48) |       /* PAT6 */
+                ((uint64_t)X86_MT_UC << 56);        /* PAT7 */
 
             vmx_get_guest_pat(v, pat);
             vmx_set_guest_pat(v, uc_pat);
xen/arch/x86/include/asm/mtrr.h +1 -11
index 7733800b798fc2c72ba87e4ce6500e4183553d04..92fc930c692039b6c709d6a04f6553593f40aa55 100644
@@ -16,17 +16,7 @@
 #define NORMAL_CACHE_MODE          0
 #define NO_FILL_CACHE_MODE         2
 
-enum {
-    PAT_TYPE_UNCACHABLE=0,
-    PAT_TYPE_WRCOMB=1,
-    PAT_TYPE_WRTHROUGH=4,
-    PAT_TYPE_WRPROT=5,
-    PAT_TYPE_WRBACK=6,
-    PAT_TYPE_UC_MINUS=7,
-    PAT_TYPE_NUMS
-};
-
-#define INVALID_MEM_TYPE PAT_TYPE_NUMS
+#define INVALID_MEM_TYPE X86_NUM_MT
 
 /* In the Intel processor's MTRR interface, the MTRR type is always held in
    an 8 bit field: */
xen/arch/x86/mm/p2m-ept.c +2 -2
index d61d66c20e4180f8cbe21bcd97b568519e0b738e..126437285d8a9f222fca6a7b6ff4434b60637847 100644
@@ -573,8 +573,8 @@ int epte_get_entry_emt(struct domain *d, gfn_t gfn, mfn_t mfn,
     if ( gmtrr_mtype >= 0 )
     {
         *ipat = true;
-        return gmtrr_mtype != PAT_TYPE_UC_MINUS ? gmtrr_mtype
-                                                : MTRR_TYPE_UNCACHABLE;
+        return gmtrr_mtype != X86_MT_UCM ? gmtrr_mtype
+                                         : MTRR_TYPE_UNCACHABLE;
     }
     if ( gmtrr_mtype == -EADDRNOTAVAIL )
         return -1;
xen/arch/x86/mm/shadow/multi.c +2 -2
index 6bb564b0145285afc93b72a60b7797fcfe8696dc..b64bba70fc17906236872a017ad48ce91fd30803 100644
@@ -561,7 +561,7 @@ _sh_propagate(struct vcpu *v,
              (type = hvm_get_mem_pinned_cacheattr(d, target_gfn, 0)) >= 0 )
             sflags |= pat_type_2_pte_flags(type);
         else if ( d->arch.hvm.is_in_uc_mode )
-            sflags |= pat_type_2_pte_flags(PAT_TYPE_UNCACHABLE);
+            sflags |= pat_type_2_pte_flags(X86_MT_UC);
         else
             if ( iomem_access_permitted(d, mfn_x(target_mfn), mfn_x(target_mfn)) )
             {
@@ -572,7 +572,7 @@ _sh_propagate(struct vcpu *v,
                             mfn_to_maddr(target_mfn),
                             MTRR_TYPE_UNCACHABLE);
                 else if ( iommu_snoop )
-                    sflags |= pat_type_2_pte_flags(PAT_TYPE_WRBACK);
+                    sflags |= pat_type_2_pte_flags(X86_MT_WB);
                 else
                     sflags |= get_pat_flags(v,
                             gflags,
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#7 lore raw

PATCH v5 4/10 x86: Replace MTRR_* constants with X86_MT_* constants

In reply to Demi Marie Obenour

This allows eliminating of the former, with the exception of MTRR_NUM_TYPES. MTRR_NUM_TYPES is kept, as due to a quirk of the x86 architecture X86_MT_UCM (7) is not valid in an MTRR.

---
Changes since v2:

- Improve commit message
- Do not replace MTRR_NUM_TYPES with X86_MT_UCM
- State explicitly that MTRR_NUM_TYPES is kept

---
 xen/arch/x86/cpu/mtrr/generic.c         | 10 ++---
 xen/arch/x86/cpu/mtrr/main.c            | 26 ++++++-------
 xen/arch/x86/e820.c                     |  4 +-
 xen/arch/x86/hvm/mtrr.c                 | 30 +++++++--------
 xen/arch/x86/hvm/vmx/vmcs.c             |  2 +-
 xen/arch/x86/hvm/vmx/vmx.c              |  2 +-
 xen/arch/x86/include/asm/hvm/vmx/vmcs.h |  2 +-
 xen/arch/x86/include/asm/mtrr.h         | 10 +----
 xen/arch/x86/mm/p2m-ept.c               | 51 ++++++++++++-------------
 xen/arch/x86/mm/shadow/multi.c          |  2 +-
 10 files changed, 66 insertions(+), 73 deletions(-)
xen/arch/x86/cpu/mtrr/generic.c +5 -5
index 47aaf76226e0a8a0712b7211ed339a4a032ab3f3..660ae26c2350b3436a471155fc0426699ba8ac1d 100644
@@ -127,11 +127,11 @@ static const char *__init mtrr_attrib_to_str(mtrr_type x)
 {
 	static const char __initconst strings[MTRR_NUM_TYPES][16] =
 	{
-		[MTRR_TYPE_UNCACHABLE]     = "uncachable",
-		[MTRR_TYPE_WRCOMB]         = "write-combining",
-		[MTRR_TYPE_WRTHROUGH]      = "write-through",
-		[MTRR_TYPE_WRPROT]         = "write-protect",
-		[MTRR_TYPE_WRBACK]         = "write-back",
+		[X86_MT_UC] = "uncachable",
+		[X86_MT_WC] = "write-combining",
+		[X86_MT_WT] = "write-through",
+		[X86_MT_WP] = "write-protect",
+		[X86_MT_WB] = "write-back",
 	};
 
 	return (x < ARRAY_SIZE(strings) && strings[x][0]) ? strings[x] : "?";
xen/arch/x86/cpu/mtrr/main.c +13 -13
index 4e01c8d6f9df6562b94438f265d79a0a6fca8de6..2946003b84938f3b83c98b62dfaa3ace90822983 100644
@@ -163,10 +163,10 @@ static void cf_check ipi_handler(void *info)
 }
 
 static inline int types_compatible(mtrr_type type1, mtrr_type type2) {
-	return type1 == MTRR_TYPE_UNCACHABLE ||
-	       type2 == MTRR_TYPE_UNCACHABLE ||
-	       (type1 == MTRR_TYPE_WRTHROUGH && type2 == MTRR_TYPE_WRBACK) ||
-	       (type1 == MTRR_TYPE_WRBACK && type2 == MTRR_TYPE_WRTHROUGH);
+	return type1 == X86_MT_UC ||
+	       type2 == X86_MT_UC ||
+	       (type1 == X86_MT_WT && type2 == X86_MT_WB) ||
+	       (type1 == X86_MT_WB && type2 == X86_MT_WT);
 }
 
 /**
@@ -297,13 +297,13 @@ static void set_mtrr(unsigned int reg, unsigned long base,
  *
  *	The available types are
  *
- *	%MTRR_TYPE_UNCACHABLE	-	No caching
+ *	%X86_MT_UC	-	No caching
  *
- *	%MTRR_TYPE_WRBACK	-	Write data back in bursts whenever
+ *	%X86_MT_WB	-	Write data back in bursts whenever
  *
- *	%MTRR_TYPE_WRCOMB	-	Write data back soon but allow bursts
+ *	%X86_MT_WC	-	Write data back soon but allow bursts
  *
- *	%MTRR_TYPE_WRTHROUGH	-	Cache reads but not writes
+ *	%X86_MT_WT	-	Cache reads but not writes
  *
  *	BUGS: Needs a quiet flag for the cases where drivers do not mind
  *	failures and do not wish system log messages to be sent.
@@ -328,7 +328,7 @@ int mtrr_add_page(unsigned long base, unsigned long size,
 	}
 
 	/*  If the type is WC, check that this processor supports it  */
-	if ((type == MTRR_TYPE_WRCOMB) && !have_wrcomb()) {
+	if ((type == X86_MT_WC) && !have_wrcomb()) {
 		printk(KERN_WARNING
 		       "mtrr: your processor doesn't support write-combining\n");
 		return -EOPNOTSUPP;
@@ -442,13 +442,13 @@ static int mtrr_check(unsigned long base, unsigned long size)
  *
  *	The available types are
  *
- *	%MTRR_TYPE_UNCACHABLE	-	No caching
+ *	%X86_MT_UC	-	No caching
  *
- *	%MTRR_TYPE_WRBACK	-	Write data back in bursts whenever
+ *	%X86_MT_WB	-	Write data back in bursts whenever
  *
- *	%MTRR_TYPE_WRCOMB	-	Write data back soon but allow bursts
+ *	%X86_MT_WC	-	Write data back soon but allow bursts
  *
- *	%MTRR_TYPE_WRTHROUGH	-	Cache reads but not writes
+ *	%X86_MT_WT	-	Cache reads but not writes
  *
  *	BUGS: Needs a quiet flag for the cases where drivers do not mind
  *	failures and do not wish system log messages to be sent.
xen/arch/x86/e820.c +2 -2
index b653a19c93afb98c2d64330384cb4fa7b4d2e1ec..c5911cf48dc4a281c03ddef35f23b19bc7af42eb 100644
@@ -459,7 +459,7 @@ static uint64_t __init mtrr_top_of_ram(void)
         printk(" MTRR cap: %"PRIx64" type: %"PRIx64"\n", mtrr_cap, mtrr_def);
 
     /* MTRRs enabled, and default memory type is not writeback? */
-    if ( !test_bit(11, &mtrr_def) || ((uint8_t)mtrr_def == MTRR_TYPE_WRBACK) )
+    if ( !test_bit(11, &mtrr_def) || ((uint8_t)mtrr_def == X86_MT_WB) )
         return 0;
 
     /*
@@ -476,7 +476,7 @@ static uint64_t __init mtrr_top_of_ram(void)
             printk(" MTRR[%d]: base %"PRIx64" mask %"PRIx64"\n",
                    i, base, mask);
 
-        if ( !test_bit(11, &mask) || ((uint8_t)base != MTRR_TYPE_WRBACK) )
+        if ( !test_bit(11, &mask) || ((uint8_t)base != X86_MT_WB) )
             continue;
         base &= addr_mask;
         mask &= addr_mask;
xen/arch/x86/hvm/mtrr.c +15 -15
index 242623f3c239ee18a44f882ecb3910a00c615825..093103f6c768cf64f880d1b20e1c14f5918c1250 100644
@@ -39,11 +39,11 @@ static const uint8_t pat_entry_2_pte_flags[8] = {
 /* Effective mm type lookup table, according to MTRR and PAT. */
 static const uint8_t mm_type_tbl[MTRR_NUM_TYPES][X86_NUM_MT] = {
 #define RS MEMORY_NUM_TYPES
-#define UC MTRR_TYPE_UNCACHABLE
-#define WB MTRR_TYPE_WRBACK
-#define WC MTRR_TYPE_WRCOMB
-#define WP MTRR_TYPE_WRPROT
-#define WT MTRR_TYPE_WRTHROUGH
+#define UC X86_MT_UC
+#define WB X86_MT_WB
+#define WC X86_MT_WC
+#define WP X86_MT_WP
+#define WT X86_MT_WT
 
 /*          PAT(UC, WC, RS, RS, WT, WP, WB, UC-) */
 /* MTRR(UC) */ {UC, WC, RS, RS, UC, UC, UC, UC},
@@ -202,7 +202,7 @@ int mtrr_get_type(const struct mtrr_state *m, paddr_t pa, unsigned int order)
    unsigned int seg, num_var_ranges = MASK_EXTR(m->mtrr_cap, MTRRcap_VCNT);
 
    if ( unlikely(!m->enabled) )
-       return MTRR_TYPE_UNCACHABLE;
+       return X86_MT_UC;
 
    pa &= mask;
    if ( (pa < 0x100000) && m->fixed_enabled )
@@ -277,13 +277,13 @@ int mtrr_get_type(const struct mtrr_state *m, paddr_t pa, unsigned int order)
        return -1;
 
    /* Two or more matches, one being UC? */
-   if ( overlap_mtrr & (1 << MTRR_TYPE_UNCACHABLE) )
-       return MTRR_TYPE_UNCACHABLE;
+   if ( overlap_mtrr & (1 << X86_MT_UC) )
+       return X86_MT_UC;
 
    /* Two or more matches, all of them WT and WB? */
    if ( overlap_mtrr ==
-        ((1 << MTRR_TYPE_WRTHROUGH) | (1 << MTRR_TYPE_WRBACK)) )
-       return MTRR_TYPE_WRTHROUGH;
+        ((1 << X86_MT_WT) | (1 << X86_MT_WB)) )
+       return X86_MT_WT;
 
    /* Behaviour is undefined, but return the last overlapped type. */
    return overlap_mtrr_pos;
@@ -381,11 +381,11 @@ static inline bool_t valid_mtrr_type(uint8_t type)
 {
     switch ( type )
     {
-    case MTRR_TYPE_UNCACHABLE:
-    case MTRR_TYPE_WRBACK:
-    case MTRR_TYPE_WRCOMB:
-    case MTRR_TYPE_WRPROT:
-    case MTRR_TYPE_WRTHROUGH:
+    case X86_MT_UC:
+    case X86_MT_WB:
+    case X86_MT_WC:
+    case X86_MT_WP:
+    case X86_MT_WT:
         return 1;
     }
     return 0;
xen/arch/x86/hvm/vmx/vmcs.c +1 -1
index 84dbb88d33b76111833a37339186199f8bc03b5e..f0825216d722d978f221bb34a797d8de5505cb80 100644
@@ -555,7 +555,7 @@ static int vmx_init_vmcs_config(bool bsp)
     /* Require Write-Back (WB) memory type for VMCS accesses. */
     opt = (vmx_basic_msr_high & (VMX_BASIC_MEMORY_TYPE_MASK >> 32)) /
           ((VMX_BASIC_MEMORY_TYPE_MASK & -VMX_BASIC_MEMORY_TYPE_MASK) >> 32);
-    if ( opt != MTRR_TYPE_WRBACK )
+    if ( opt != X86_MT_WB )
     {
         printk("VMX: CPU%d has unexpected VMCS access type %u\n",
                smp_processor_id(), opt);
xen/arch/x86/hvm/vmx/vmx.c +1 -1
index b543c3983d77ae807e8bd97330691a79d8d39bae..4ae7dd56c9981d32ac545d6e7b7c126b15f68969 100644
@@ -434,7 +434,7 @@ static void cf_check domain_creation_finished(struct domain *d)
         return;
 
     ASSERT(epte_get_entry_emt(d, gfn, apic_access_mfn, 0, &ipat,
-                              p2m_mmio_direct) == MTRR_TYPE_WRBACK);
+                              p2m_mmio_direct) == X86_MT_WB);
     ASSERT(ipat);
 
     if ( set_mmio_p2m_entry(d, gfn, apic_access_mfn, PAGE_ORDER_4K) )
xen/arch/x86/include/asm/hvm/vmx/vmcs.h +1 -1
index 75f9928abfad28e3895fe3dd4058b2b0a6e145c3..65e9e27b5437adff59abc46976f73a9f2cc587da 100644
@@ -38,7 +38,7 @@ struct vmx_msr_entry {
     u64 data;
 };
 
-#define EPT_DEFAULT_MT      MTRR_TYPE_WRBACK
+#define EPT_DEFAULT_MT      X86_MT_WB
 
 struct ept_data {
     union {
xen/arch/x86/include/asm/mtrr.h +2 -8
index 92fc930c692039b6c709d6a04f6553593f40aa55..e4f6ca6048334b2094a1836cc2f298453641232f 100644
@@ -3,15 +3,9 @@
 
 #include <xen/mm.h>
 
-/* These are the region types. They match the architectural specification. */
-#define MTRR_TYPE_UNCACHABLE 0
-#define MTRR_TYPE_WRCOMB     1
-#define MTRR_TYPE_WRTHROUGH  4
-#define MTRR_TYPE_WRPROT     5
-#define MTRR_TYPE_WRBACK     6
-#define MTRR_NUM_TYPES       7
+#define MTRR_NUM_TYPES       X86_MT_UCM
 #define MEMORY_NUM_TYPES     MTRR_NUM_TYPES
-#define NO_HARDCODE_MEM_TYPE    MTRR_NUM_TYPES
+#define NO_HARDCODE_MEM_TYPE MTRR_NUM_TYPES
 
 #define NORMAL_CACHE_MODE          0
 #define NO_FILL_CACHE_MODE         2
xen/arch/x86/mm/p2m-ept.c +25 -26

120 lines not on this page · raw message

xen/arch/x86/mm/shadow/multi.c +1 -1
index b64bba70fc17906236872a017ad48ce91fd30803..f5f7ff021bd9e057c5b6f6329de7acb5ef05d58f 100644
@@ -570,7 +570,7 @@ _sh_propagate(struct vcpu *v,
                             gflags,
                             gfn_to_paddr(target_gfn),
                             mfn_to_maddr(target_mfn),
-                            MTRR_TYPE_UNCACHABLE);
+                            X86_MT_UC);
                 else if ( iommu_snoop )
                     sflags |= pat_type_2_pte_flags(X86_MT_WB);
                 else
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#8 lore raw

PATCH v5 5/10 x86: Replace EPT_EMT_* constants with X86_MT_*

In reply to Demi Marie Obenour

This allows eliminating the former. No functional change intended.

---
 xen/arch/x86/include/asm/hvm/vmx/vmx.h | 9 ---------
 xen/arch/x86/mm/hap/nested_ept.c       | 4 ++--
 2 files changed, 2 insertions(+), 11 deletions(-)
xen/arch/x86/include/asm/hvm/vmx/vmx.h +0 -9
index 8eedf59155e01ec1ca84dcc6b30961f9c884cb3b..49fe9822fac5eae15b67f0cfd3d0cb96347dc7ed 100644
@@ -80,15 +80,6 @@ typedef enum {
 #define EPTE_RWX_MASK           0x7
 #define EPTE_FLAG_MASK          0x7f
 
-#define EPT_EMT_UC              0
-#define EPT_EMT_WC              1
-#define EPT_EMT_RSV0            2
-#define EPT_EMT_RSV1            3
-#define EPT_EMT_WT              4
-#define EPT_EMT_WP              5
-#define EPT_EMT_WB              6
-#define EPT_EMT_RSV2            7
-
 #define PI_xAPIC_NDST_MASK      0xFF00
 
 void vmx_asm_vmexit_handler(struct cpu_user_regs);
xen/arch/x86/mm/hap/nested_ept.c +2 -2
index 1cb7fefc37091bf7d92a277203e652add5611871..23fb3889b7605be62805731218c314819d5027b5 100644
@@ -84,8 +84,8 @@ static bool_t nept_emt_bits_check(ept_entry_t e, uint32_t level)
 {
     if ( e.sp || level == 1 )
     {
-        if ( e.emt == EPT_EMT_RSV0 || e.emt == EPT_EMT_RSV1 ||
-             e.emt == EPT_EMT_RSV2 )
+        if ( e.emt == X86_MT_RSVD_2 || e.emt == X86_MT_RSVD_3 ||
+             e.emt == X86_MT_UCM )
             return 1;
     }
     return 0;
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#10 lore raw

PATCH v5 6/10 x86: Remove MEMORY_NUM_TYPES and NO_HARDCODE_MEM_TYPE

In reply to Demi Marie Obenour

No functional change intended.

---
This patch is optional. Subsequent patches should not depend on it.

Changes since v2:

- Keep MTRR_NUM_TYPES and adjust commit message accordingly

---
 xen/arch/x86/hvm/mtrr.c         | 18 +++++++++---------
 xen/arch/x86/include/asm/mtrr.h |  2 --
 xen/arch/x86/mm/shadow/multi.c  |  2 +-
 3 files changed, 10 insertions(+), 12 deletions(-)
xen/arch/x86/hvm/mtrr.c +9 -9
index 093103f6c768cf64f880d1b20e1c14f5918c1250..05e978041d62fd0d559462de181a04bef8a5bca9 100644
@@ -38,7 +38,7 @@ static const uint8_t pat_entry_2_pte_flags[8] = {
 
 /* Effective mm type lookup table, according to MTRR and PAT. */
 static const uint8_t mm_type_tbl[MTRR_NUM_TYPES][X86_NUM_MT] = {
-#define RS MEMORY_NUM_TYPES
+#define RS MTRR_NUM_TYPES
 #define UC X86_MT_UC
 #define WB X86_MT_WB
 #define WC X86_MT_WC
@@ -66,9 +66,9 @@ static const uint8_t mm_type_tbl[MTRR_NUM_TYPES][X86_NUM_MT] = {
  * Reverse lookup table, to find a pat type according to MTRR and effective
  * memory type. This table is dynamically generated.
  */
-static uint8_t __read_mostly mtrr_epat_tbl[MTRR_NUM_TYPES][MEMORY_NUM_TYPES] =
-    { [0 ... MTRR_NUM_TYPES-1] =
-        { [0 ... MEMORY_NUM_TYPES-1] = INVALID_MEM_TYPE }
+static uint8_t __read_mostly mtrr_epat_tbl[MTRR_NUM_TYPES][MTRR_NUM_TYPES] =
+    { [0 ... MTRR_NUM_TYPES - 1] =
+        { [0 ... MTRR_NUM_TYPES - 1] = INVALID_MEM_TYPE }
     };
 
 /* Lookup table for PAT entry of a given PAT value in host PAT. */
@@ -85,7 +85,7 @@ static int __init cf_check hvm_mtrr_pat_init(void)
         {
             unsigned int tmp = mm_type_tbl[i][j];
 
-            if ( tmp < MEMORY_NUM_TYPES )
+            if ( tmp < MTRR_NUM_TYPES )
                 mtrr_epat_tbl[i][tmp] = j;
         }
     }
@@ -317,11 +317,11 @@ static uint8_t effective_mm_type(struct mtrr_state *m,
                                  uint8_t gmtrr_mtype)
 {
     uint8_t mtrr_mtype, pat_value;
-   
+
     /* if get_pat_flags() gives a dedicated MTRR type,
      * just use it
-     */ 
-    if ( gmtrr_mtype == NO_HARDCODE_MEM_TYPE )
+     */
+    if ( gmtrr_mtype == MTRR_NUM_TYPES )
         mtrr_mtype = mtrr_get_type(m, gpa, 0);
     else
         mtrr_mtype = gmtrr_mtype;
@@ -346,7 +346,7 @@ uint32_t get_pat_flags(struct vcpu *v,
     /* 1. Get the effective memory type of guest physical address,
      * with the pair of guest MTRR and PAT
      */
-    guest_eff_mm_type = effective_mm_type(g, pat, gpaddr, 
+    guest_eff_mm_type = effective_mm_type(g, pat, gpaddr,
                                           gl1e_flags, gmtrr_mtype);
     /* 2. Get the memory type of host physical address, with MTRR */
     shadow_mtrr_type = mtrr_get_type(&mtrr_state, spaddr, 0);
xen/arch/x86/include/asm/mtrr.h +0 -2
index e4f6ca6048334b2094a1836cc2f298453641232f..4b7f840a965954cc4b59698327a37e47026893a4 100644
@@ -4,8 +4,6 @@
 #include <xen/mm.h>
 
 #define MTRR_NUM_TYPES       X86_MT_UCM
-#define MEMORY_NUM_TYPES     MTRR_NUM_TYPES
-#define NO_HARDCODE_MEM_TYPE MTRR_NUM_TYPES
 
 #define NORMAL_CACHE_MODE          0
 #define NO_FILL_CACHE_MODE         2
xen/arch/x86/mm/shadow/multi.c +1 -1
index f5f7ff021bd9e057c5b6f6329de7acb5ef05d58f..1faf9940db6b0afefc5977c00c00fb6a39cd27d2 100644
@@ -578,7 +578,7 @@ _sh_propagate(struct vcpu *v,
                             gflags,
                             gfn_to_paddr(target_gfn),
                             mfn_to_maddr(target_mfn),
-                            NO_HARDCODE_MEM_TYPE);
+                            MTRR_NUM_TYPES);
             }
     }
 
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#11 lore raw

PATCH v5 7/10 x86: Derive XEN_MSR_PAT from its individual entries

In reply to Demi Marie Obenour

This avoids it being a magic constant that is difficult for humans to decode. Use BUILD_BUG_ON to check that the old and new values are identical.

---
Changes since v4:
- Explain that changing XEN_MSR_PAT breaks guests that rely on the API in xen.h instead of reading the PAT from Xen.

---
 xen/arch/x86/include/asm/processor.h |  9 ++++++++-
 xen/arch/x86/mm.c                    | 11 +++++++++++
 2 files changed, 19 insertions(+), 1 deletion(-)
xen/arch/x86/include/asm/processor.h +8 -1
index 8e2816fae9b97bd4e153a30cc3802971fe0355af..60b902060914584957db8afa5c7c1e6abdad4d13 100644
@@ -96,7 +96,14 @@
  * Host IA32_CR_PAT value to cover all memory types.  This is not the default
  * MSR_PAT value, and is an ABI with PV guests.
  */
-#define XEN_MSR_PAT _AC(0x050100070406, ULL)
+#define XEN_MSR_PAT ((_AC(X86_MT_WB,  ULL) << 0x00) | \
+                     (_AC(X86_MT_WT,  ULL) << 0x08) | \
+                     (_AC(X86_MT_UCM, ULL) << 0x10) | \
+                     (_AC(X86_MT_UC,  ULL) << 0x18) | \
+                     (_AC(X86_MT_WC,  ULL) << 0x20) | \
+                     (_AC(X86_MT_WP,  ULL) << 0x28) | \
+                     (_AC(X86_MT_UC,  ULL) << 0x30) | \
+                     (_AC(X86_MT_UC,  ULL) << 0x38))
 
 #ifndef __ASSEMBLY__
 
xen/arch/x86/mm.c +11 -0
index dba6c77ef2f7ed7fcb7f7e526583ccadd35e62cc..b40a575b61418ea1137299e68b64f7efd9efeced 100644
@@ -6352,6 +6352,17 @@ unsigned long get_upper_mfn_bound(void)
     return min(max_mfn, 1UL << (paddr_bits - PAGE_SHIFT)) - 1;
 }
 
+static void __init __maybe_unused build_assertions(void)
+{
+    /*
+     * If this trips, any guest that blindly rely on the public API in xen.h
+     * (instead of reading the PAT from Xen, as Linux 3.19+ does) will be
+     * broken.  Furthermore, live migration of PV guests between Xen versions
+     * using different PATs will not work.
+     */
+    BUILD_BUG_ON(XEN_MSR_PAT != 0x050100070406ULL);
+}
+
 /*
  * Local variables:
  * mode: C
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#12 lore raw

PATCH v5 8/10 x86/mm: make code robust to future PAT changes

In reply to Demi Marie Obenour

It may be desirable to change Xen's PAT for various reasons. This requires changes to several _PAGE_* macros as well. Add static assertions to check that XEN_MSR_PAT is consistent with the _PAGE_* macros, and that _PAGE_WB is 0 as required by Linux.

---
Changes since v4:
- Add lots of comments explaining what the various BUILD_BUG_ON()s mean.

Changes since v3:
- Refactor some macros
- Avoid including a string literal in BUILD_BUG_ON

---
 xen/arch/x86/mm.c | 71 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 71 insertions(+)
xen/arch/x86/mm.c +71 -0
index b40a575b61418ea1137299e68b64f7efd9efeced..a72556668633ee57b77c9a57d3a13dd5a12d9bbf 100644
@@ -6352,6 +6352,11 @@ unsigned long get_upper_mfn_bound(void)
     return min(max_mfn, 1UL << (paddr_bits - PAGE_SHIFT)) - 1;
 }
 
+
+/*
+ * A bunch of static assertions to check that the XEN_MSR_PAT is valid
+ * and consistent with the _PAGE_* macros, and that _PAGE_WB is zero.
+ */
 static void __init __maybe_unused build_assertions(void)
 {
     /*
@@ -6361,6 +6366,72 @@ static void __init __maybe_unused build_assertions(void)
      * using different PATs will not work.
      */
     BUILD_BUG_ON(XEN_MSR_PAT != 0x050100070406ULL);
+
+    /*
+     * _PAGE_WB must be zero for several reasons, not least because Linux
+     * assumes it.
+     */
+    BUILD_BUG_ON(_PAGE_WB);
+
+    /* A macro to convert from cache attributes to actual cacheability */
+#define PAT_ENTRY(v) (0xFF & (XEN_MSR_PAT >> (8 * (v))))
+
+    /* Validate at compile-time that v is a valid value for a PAT entry */
+#define CHECK_PAT_ENTRY_VALUE(v)                                               \
+    BUILD_BUG_ON((v) < 0 || (v) > 7 ||                                         \
+                 (v) == X86_MT_RSVD_2 || (v) == X86_MT_RSVD_3)
+
+    /* Validate at compile-time that PAT entry v is valid */
+#define CHECK_PAT_ENTRY(v) do {                                                \
+    BUILD_BUG_ON((v) < 0 || (v) > 7);                                          \
+    CHECK_PAT_ENTRY_VALUE(PAT_ENTRY(v));                                       \
+} while (0);
+
+    /*
+     * If one of these trips, the corresponding entry in XEN_MSR_PAT is invalid.
+     * This would cause Xen to crash (with #GP) at startup.
+     */
+    CHECK_PAT_ENTRY(0);
+    CHECK_PAT_ENTRY(1);
+    CHECK_PAT_ENTRY(2);
+    CHECK_PAT_ENTRY(3);
+    CHECK_PAT_ENTRY(4);
+    CHECK_PAT_ENTRY(5);
+    CHECK_PAT_ENTRY(6);
+    CHECK_PAT_ENTRY(7);
+
+#undef CHECK_PAT_ENTRY
+#undef CHECK_PAT_ENTRY_VALUE
+
+    /* Macro version of page_flags_to_cacheattr(), for use in BUILD_BUG_ON()s */
+#define PAGE_FLAGS_TO_CACHEATTR(page_value)                                    \
+    ((((page_value) >> 5) & 4) | (((page_value) >> 3) & 3))
+
+    /* Check that a PAT-related _PAGE_* macro is correct */
+#define CHECK_PAGE_VALUE(page_value) do {                                      \
+    /* Check that the _PAGE_* macros only use bits from PAGE_CACHE_ATTRS */    \
+    BUILD_BUG_ON(((_PAGE_##page_value) & PAGE_CACHE_ATTRS) !=                  \
+                  (_PAGE_##page_value));                                       \
+    /* Check that the _PAGE_* are consistent with XEN_MSR_PAT */               \
+    BUILD_BUG_ON(PAT_ENTRY(PAGE_FLAGS_TO_CACHEATTR(_PAGE_##page_value)) !=     \
+                 (X86_MT_##page_value));                                       \
+} while (0)
+
+    /*
+     * If one of these trips, the corresponding _PAGE_* macro is inconsistent
+     * with XEN_MSR_PAT.  This would cause Xen to use incorrect cacheability
+     * flags, with results that are undefined and probably harmful.
+     */
+    CHECK_PAGE_VALUE(WT);
+    CHECK_PAGE_VALUE(WB);
+    CHECK_PAGE_VALUE(WC);
+    CHECK_PAGE_VALUE(UC);
+    CHECK_PAGE_VALUE(UCM);
+    CHECK_PAGE_VALUE(WP);
+
+#undef CHECK_PAGE_VALUE
+#undef PAGE_FLAGS_TO_CACHEATTR
+#undef PAT_ENTRY
 }
 
 /*
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#13 lore raw

Re: [PATCH v5 08/10] x86/mm: make code robust to future PAT changes

Quoted 4 lines · Demi Marie Obenour

It may be desirable to change Xen's PAT for various reasons. This requires changes to several _PAGE_* macros as well. Add static assertions to check that XEN_MSR_PAT is consistent with the _PAGE_* macros, and that _PAGE_WB is 0 as required by Linux.

In line with the code comment, perhaps add (not just)"?

Quoted 15 lines
xen/arch/x86/mm.c +5 -0
@@ -6352,6 +6352,11 @@ unsigned long get_upper_mfn_bound(void)
     return min(max_mfn, 1UL << (paddr_bits - PAGE_SHIFT)) - 1;
 }
 
+
+/*
+ * A bunch of static assertions to check that the XEN_MSR_PAT is valid
+ * and consistent with the _PAGE_* macros, and that _PAGE_WB is zero.
+ */
 static void __init __maybe_unused build_assertions(void)
 {
     /*

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6374

Quoted 9 lines
xen/arch/x86/mm.c +6 -0
@@ -6361,6 +6366,72 @@ static void __init __maybe_unused build_assertions(void)
      * using different PATs will not work.
      */
     BUILD_BUG_ON(XEN_MSR_PAT != 0x050100070406ULL);
+
+    /*
+     * _PAGE_WB must be zero for several reasons, not least because Linux
+     * assumes it.
+     */
+    BUILD_BUG_ON(_PAGE_WB);

Strictly speaking this is a requirement only for PV guests. We may not want to go as far as putting "#ifdef CONFIG_PV" around it, but at least the code comment (and then also the part of the description referring to this) will imo want to say so.

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6377

Quoted 2 lines
xen/arch/x86/mm.c +2 -0
+    /* A macro to convert from cache attributes to actual cacheability */
+#define PAT_ENTRY(v) (0xFF & (XEN_MSR_PAT >> (8 * (v))))

I don't think the comment is appropriate here. All you do is extract a slot from the hard-coded PAT value we use.

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6381

Quoted 3 lines
xen/arch/x86/mm.c +3 -0
+    /* Validate at compile-time that v is a valid value for a PAT entry */
+#define CHECK_PAT_ENTRY_VALUE(v)                                               \
+    BUILD_BUG_ON((v) < 0 || (v) > 7 ||                                         \

PAT_ENTRY() won't produce negative values, so I don't think "(v) < 0" is really needed here. And the "(v) > 7" will imo want replacing by "(v) >= X86_NUM_MT".

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6386

Quoted 5 lines
xen/arch/x86/mm.c +5 -0
+                 (v) == X86_MT_RSVD_2 || (v) == X86_MT_RSVD_3)
+
+    /* Validate at compile-time that PAT entry v is valid */
+#define CHECK_PAT_ENTRY(v) do {                                                \
+    BUILD_BUG_ON((v) < 0 || (v) > 7);                                          \

I think this would better be part of PAT_ENTRY(), as the validity of the shift there depends on it. If this check is needed in the first place, seeing that the macro is #undef-ed right after use below, and hence the checks only cover the eight invocations of the macro right here.

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6388

Quoted 2 lines
xen/arch/x86/mm.c +2 -0
+    CHECK_PAT_ENTRY_VALUE(PAT_ENTRY(v));                                       \
+} while (0);

Nit (style): Missing blanks around 0 and perhaps better nowadays to use "false" in such constructs. (Because of other comments this may go away here anyway, but there's another similar instance below).

Quoted 7 lines
xen/arch/x86/mm.c +7 -0
+    /*
+     * If one of these trips, the corresponding entry in XEN_MSR_PAT is invalid.
+     * This would cause Xen to crash (with #GP) at startup.
+     */
+    CHECK_PAT_ENTRY(0);
+    CHECK_PAT_ENTRY(1);
+    CHECK_PAT_ENTRY(2);

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6406

Quoted 10 lines
xen/arch/x86/mm.c +10 -0
+    CHECK_PAT_ENTRY(3);
+    CHECK_PAT_ENTRY(4);
+    CHECK_PAT_ENTRY(5);
+    CHECK_PAT_ENTRY(6);
+    CHECK_PAT_ENTRY(7);
+
+#undef CHECK_PAT_ENTRY
+#undef CHECK_PAT_ENTRY_VALUE
+
+    /* Macro version of page_flags_to_cacheattr(), for use in BUILD_BUG_ON()s */

DYM pte_flags_to_cacheattr()? At which point the macro name wants to match and its parameter may also better be named "pte_value"?

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6408

Quoted 2 lines
xen/arch/x86/mm.c +2 -0
+#define PAGE_FLAGS_TO_CACHEATTR(page_value)                                    \
+    ((((page_value) >> 5) & 4) | (((page_value) >> 3) & 3))

Hmm, yet more uses of magic literal numbers.

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6414

Quoted 5 lines
xen/arch/x86/mm.c +5 -0
+    /* Check that a PAT-related _PAGE_* macro is correct */
+#define CHECK_PAGE_VALUE(page_value) do {                                      \
+    /* Check that the _PAGE_* macros only use bits from PAGE_CACHE_ATTRS */    \
+    BUILD_BUG_ON(((_PAGE_##page_value) & PAGE_CACHE_ATTRS) !=                  \
+                  (_PAGE_##page_value));                                       \

Nit (style): One too many blanks used for indentation.

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6417

Quoted 3 lines
xen/arch/x86/mm.c +3 -0
+    /* Check that the _PAGE_* are consistent with XEN_MSR_PAT */               \
+    BUILD_BUG_ON(PAT_ENTRY(PAGE_FLAGS_TO_CACHEATTR(_PAGE_##page_value)) !=     \
+                 (X86_MT_##page_value));                                       \

Nit (style): Nowadays we tend to consider ## a binary operator just like e.g. +, and hence it wants to be surrounded by blanks.

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6423

Quoted 6 lines
xen/arch/x86/mm.c +6 -0
+} while (0)
+
+    /*
+     * If one of these trips, the corresponding _PAGE_* macro is inconsistent
+     * with XEN_MSR_PAT.  This would cause Xen to use incorrect cacheability
+     * flags, with results that are undefined and probably harmful.

Why "undefined"? They may be wrong / broken, but the result is still well- defined afaict.

Jan

#14 lore raw

Re: [PATCH v5 08/10] x86/mm: make code robust to future PAT changes

Quoted 7 lines · Jan Beulich
Quoted 4 lines · Demi Marie Obenour

It may be desirable to change Xen's PAT for various reasons. This requires changes to several _PAGE_* macros as well. Add static assertions to check that XEN_MSR_PAT is consistent with the _PAGE_* macros, and that _PAGE_WB is 0 as required by Linux.

In line with the code comment, perhaps add (not just)"?

Will reword in v6.

Quoted 21 lines
Quoted 21 lines
xen/arch/x86/mm.c +8 -0
@@ -6352,6 +6352,11 @@ unsigned long get_upper_mfn_bound(void)
     return min(max_mfn, 1UL << (paddr_bits - PAGE_SHIFT)) - 1;
 }
 
+
+/*
+ * A bunch of static assertions to check that the XEN_MSR_PAT is valid
+ * and consistent with the _PAGE_* macros, and that _PAGE_WB is zero.
+ */
 static void __init __maybe_unused build_assertions(void)
 {
     /*
@@ -6361,6 +6366,72 @@ static void __init __maybe_unused build_assertions(void)
      * using different PATs will not work.
      */
     BUILD_BUG_ON(XEN_MSR_PAT != 0x050100070406ULL);
+
+    /*
+     * _PAGE_WB must be zero for several reasons, not least because Linux
Quoted 8 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6374

Quoted 3 lines
xen/arch/x86/mm.c +3 -0
+     * assumes it.
+     */
+    BUILD_BUG_ON(_PAGE_WB);

Strictly speaking this is a requirement only for PV guests. We may not want to go as far as putting "#ifdef CONFIG_PV" around it, but at least the code comment (and then also the part of the description referring to this) will imo want to say so.

Does Xen itself depend on this?

Quoted 5 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6377

Quoted 2 lines
xen/arch/x86/mm.c +2 -0
+    /* A macro to convert from cache attributes to actual cacheability */
+#define PAT_ENTRY(v) (0xFF & (XEN_MSR_PAT >> (8 * (v))))

I don't think the comment is appropriate here. All you do is extract a slot from the hard-coded PAT value we use.

Will drop in v6.

Quoted 7 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6381

Quoted 3 lines
xen/arch/x86/mm.c +3 -0
+    /* Validate at compile-time that v is a valid value for a PAT entry */
+#define CHECK_PAT_ENTRY_VALUE(v)                                               \
+    BUILD_BUG_ON((v) < 0 || (v) > 7 ||                                         \

PAT_ENTRY() won't produce negative values, so I don't think "(v) < 0" is really needed here. And the "(v) > 7" will imo want replacing by "(v) >= X86_NUM_MT".

Will fix in v6.

Quoted 11 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6386

Quoted 5 lines
xen/arch/x86/mm.c +5 -0
+                 (v) == X86_MT_RSVD_2 || (v) == X86_MT_RSVD_3)
+
+    /* Validate at compile-time that PAT entry v is valid */
+#define CHECK_PAT_ENTRY(v) do {                                                \
+    BUILD_BUG_ON((v) < 0 || (v) > 7);                                          \

I think this would better be part of PAT_ENTRY(), as the validity of the shift there depends on it. If this check is needed in the first place, seeing that the macro is #undef-ed right after use below, and hence the checks only cover the eight invocations of the macro right here.

Quoted 6 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6388

Quoted 2 lines
xen/arch/x86/mm.c +2 -0
+    CHECK_PAT_ENTRY_VALUE(PAT_ENTRY(v));                                       \
+} while (0);

Nit (style): Missing blanks around 0 and perhaps better nowadays to use "false" in such constructs. (Because of other comments this may go away here anyway, but there's another similar instance below).

Will fix in v6.

Quoted 12 lines
Quoted 12 lines
xen/arch/x86/mm.c +12 -0
+    /*
+     * If one of these trips, the corresponding entry in XEN_MSR_PAT is invalid.
+     * This would cause Xen to crash (with #GP) at startup.
+     */
+    CHECK_PAT_ENTRY(0);
+    CHECK_PAT_ENTRY(1);
+    CHECK_PAT_ENTRY(2);
+    CHECK_PAT_ENTRY(3);
+    CHECK_PAT_ENTRY(4);
+    CHECK_PAT_ENTRY(5);
+    CHECK_PAT_ENTRY(6);
+    CHECK_PAT_ENTRY(7);
Quoted 8 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6406

Quoted 5 lines
xen/arch/x86/mm.c +5 -0
+
+#undef CHECK_PAT_ENTRY
+#undef CHECK_PAT_ENTRY_VALUE
+
+    /* Macro version of page_flags_to_cacheattr(), for use in BUILD_BUG_ON()s */

DYM pte_flags_to_cacheattr()? At which point the macro name wants to match and its parameter may also better be named "pte_value"?

Indeed so.

Quoted 4 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6408

Quoted 2 lines
xen/arch/x86/mm.c +2 -0
+#define PAGE_FLAGS_TO_CACHEATTR(page_value)                                    \
+    ((((page_value) >> 5) & 4) | (((page_value) >> 3) & 3))

Hmm, yet more uses of magic literal numbers.

I wanted to keep the definition as close to that of pte_flags_to_cacheattr() as possible.

Quoted 7 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6414

Quoted 5 lines
xen/arch/x86/mm.c +5 -0
+    /* Check that a PAT-related _PAGE_* macro is correct */
+#define CHECK_PAGE_VALUE(page_value) do {                                      \
+    /* Check that the _PAGE_* macros only use bits from PAGE_CACHE_ATTRS */    \
+    BUILD_BUG_ON(((_PAGE_##page_value) & PAGE_CACHE_ATTRS) !=                  \
+                  (_PAGE_##page_value));                                       \

Nit (style): One too many blanks used for indentation.

Will fix in v6.

Quoted 6 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6417

Quoted 3 lines
xen/arch/x86/mm.c +3 -0
+    /* Check that the _PAGE_* are consistent with XEN_MSR_PAT */               \
+    BUILD_BUG_ON(PAT_ENTRY(PAGE_FLAGS_TO_CACHEATTR(_PAGE_##page_value)) !=     \
+                 (X86_MT_##page_value));                                       \

Nit (style): Nowadays we tend to consider ## a binary operator just like e.g. +, and hence it wants to be surrounded by blanks.

Will fix in v6.

Quoted 9 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6423

Quoted 6 lines
xen/arch/x86/mm.c +6 -0
+} while (0)
+
+    /*
+     * If one of these trips, the corresponding _PAGE_* macro is inconsistent
+     * with XEN_MSR_PAT.  This would cause Xen to use incorrect cacheability
+     * flags, with results that are undefined and probably harmful.

Why "undefined"? They may be wrong / broken, but the result is still well- defined afaict.

“undefined” is meant as “one has violated a core assumption that higher-level stuff depends on, so things can go arbitrarily wrong, including e.g. corrupting memory or data”. Is this accurate? Should I drop the dependent clause, or do you have a suggestion for something better?

Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#15 lore raw

Re: [PATCH v5 08/10] x86/mm: make code robust to future PAT changes

Quoted 32 lines · Demi Marie Obenour
Quoted 22 lines · Jan Beulich
Quoted 21 lines · Demi Marie Obenour
xen/arch/x86/mm.c +8 -0
@@ -6352,6 +6352,11 @@ unsigned long get_upper_mfn_bound(void)
     return min(max_mfn, 1UL << (paddr_bits - PAGE_SHIFT)) - 1;
 }
 
+
+/*
+ * A bunch of static assertions to check that the XEN_MSR_PAT is valid
+ * and consistent with the _PAGE_* macros, and that _PAGE_WB is zero.
+ */
 static void __init __maybe_unused build_assertions(void)
 {
     /*
@@ -6361,6 +6366,72 @@ static void __init __maybe_unused build_assertions(void)
      * using different PATs will not work.
      */
     BUILD_BUG_ON(XEN_MSR_PAT != 0x050100070406ULL);
+
+    /*
+     * _PAGE_WB must be zero for several reasons, not least because Linux
Quoted 8 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6374

Quoted 3 lines
xen/arch/x86/mm.c +3 -0
+     * assumes it.
+     */
+    BUILD_BUG_ON(_PAGE_WB);

Strictly speaking this is a requirement only for PV guests. We may not want to go as far as putting "#ifdef CONFIG_PV" around it, but at least the code comment (and then also the part of the description referring to this) will imo want to say so.

Quoted 1 line

Does Xen itself depend on this?

With the wording in the description I was going from the assumption that you have audited code and found that we properly use _PAGE_* constants everywhere.

Quoted 10 lines
Quoted 9 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6423

Quoted 6 lines
xen/arch/x86/mm.c +6 -0
+} while (0)
+
+    /*
+     * If one of these trips, the corresponding _PAGE_* macro is inconsistent
+     * with XEN_MSR_PAT.  This would cause Xen to use incorrect cacheability
+     * flags, with results that are undefined and probably harmful.

Why "undefined"? They may be wrong / broken, but the result is still well- defined afaict.

Quoted 5 lines

“undefined” is meant as “one has violated a core assumption that higher-level stuff depends on, so things can go arbitrarily wrong, including e.g. corrupting memory or data”. Is this accurate? Should I drop the dependent clause, or do you have a suggestion for something better?

s/undefined/unknown/ ?

Jan

#16 lore raw

Re: [PATCH v5 08/10] x86/mm: make code robust to future PAT changes

Quoted 33 lines · Jan Beulich
Quoted 32 lines · Demi Marie Obenour
Quoted 30 lines · Jan Beulich
Quoted 15 lines · Demi Marie Obenour
xen/arch/x86/mm.c +5 -0
@@ -6352,6 +6352,11 @@ unsigned long get_upper_mfn_bound(void)
     return min(max_mfn, 1UL << (paddr_bits - PAGE_SHIFT)) - 1;
 }
 
+
+/*
+ * A bunch of static assertions to check that the XEN_MSR_PAT is valid
+ * and consistent with the _PAGE_* macros, and that _PAGE_WB is zero.
+ */
 static void __init __maybe_unused build_assertions(void)
 {
     /*

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6374

Quoted 9 lines
xen/arch/x86/mm.c +6 -0
@@ -6361,6 +6366,72 @@ static void __init __maybe_unused build_assertions(void)
      * using different PATs will not work.
      */
     BUILD_BUG_ON(XEN_MSR_PAT != 0x050100070406ULL);
+
+    /*
+     * _PAGE_WB must be zero for several reasons, not least because Linux
+     * assumes it.
+     */
+    BUILD_BUG_ON(_PAGE_WB);

Strictly speaking this is a requirement only for PV guests. We may not want to go as far as putting "#ifdef CONFIG_PV" around it, but at least the code comment (and then also the part of the description referring to this) will imo want to say so.

Quoted 5 lines
Quoted 1 line

Does Xen itself depend on this?

With the wording in the description I was going from the assumption that you have audited code and found that we properly use _PAGE_* constants everywhere.

There could be other hard-coded uses of magic numbers I haven’t found, and _PAGE_WB is currently zero so I would be quite surpised if no code in Xen omits it. Linux also has a BUILD_BUG_ON() to check the same thing.

Quoted 10 lines
Quoted 10 lines
Quoted 9 lines

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6423

Quoted 6 lines
xen/arch/x86/mm.c +6 -0
+} while (0)
+
+    /*
+     * If one of these trips, the corresponding _PAGE_* macro is inconsistent
+     * with XEN_MSR_PAT.  This would cause Xen to use incorrect cacheability
+     * flags, with results that are undefined and probably harmful.

Why "undefined"? They may be wrong / broken, but the result is still well- defined afaict.

Quoted 7 lines
Quoted 5 lines

“undefined” is meant as “one has violated a core assumption that higher-level stuff depends on, so things can go arbitrarily wrong, including e.g. corrupting memory or data”. Is this accurate? Should I drop the dependent clause, or do you have a suggestion for something better?

s/undefined/unknown/ ?

Will fix in v6.

Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#17 lore raw

Re: [PATCH v5 08/10] x86/mm: make code robust to future PAT changes

Quoted 22 lines · Demi Marie Obenour
Quoted 21 lines · Jan Beulich
Quoted 19 lines · Demi Marie Obenour
Quoted 16 lines · Jan Beulich

xen/arch/x86/mm.c static void __init __maybe_unused build_assertions(void) · line 6374

Quoted 10 lines · Demi Marie Obenour
xen/arch/x86/mm.c +6 -0
@@ -6361,6 +6366,72 @@ static void __init __maybe_unused build_assertions(void)
      * using different PATs will not work.
      */
     BUILD_BUG_ON(XEN_MSR_PAT != 0x050100070406ULL);
+
+    /*
+     * _PAGE_WB must be zero for several reasons, not least because Linux
+     * assumes it.
+     */
+    BUILD_BUG_ON(_PAGE_WB);

Strictly speaking this is a requirement only for PV guests. We may not want to go as far as putting "#ifdef CONFIG_PV" around it, but at least the code comment (and then also the part of the description referring to this) will imo want to say so.

Does Xen itself depend on this?

Quoted 8 lines
Quoted 3 lines

With the wording in the description I was going from the assumption that you have audited code and found that we properly use _PAGE_* constants everywhere.

There could be other hard-coded uses of magic numbers I haven’t found, and _PAGE_WB is currently zero so I would be quite surpised if no code in Xen omits it. Linux also has a BUILD_BUG_ON() to check the same thing.

Fair enough - please adjust description and comment text accordingly then.

Jan

#18 lore raw

PATCH v5 9/10 x86/mm: Reject invalid cacheability in PV guests by default

In reply to Demi Marie Obenour

Setting cacheability flags that are not ones specified by Xen is a bug in the guest. By default, inject #GP into any guest that does this. allow_invalid_cacheability can be used on the Xen command line to disable this check.

---
Changes since v4:
- Remove pointless BUILD_BUG_ON().
- Add comment explaining why an exception is being injected.

Changes since v3:
- Add Andrew Cooper’s Suggested-by

---
 xen/arch/x86/mm.c | 30 +++++++++++++++++++++++++++++-
 1 file changed, 29 insertions(+), 1 deletion(-)
xen/arch/x86/mm.c +29 -1
index a72556668633ee57b77c9a57d3a13dd5a12d9bbf..69ce597c7cd5283ae4b5f3bc0a6dfa0bb3228d3d 100644
@@ -145,6 +145,8 @@
 
 #ifdef CONFIG_PV
 #include "pv/mm.h"
+bool allow_invalid_cacheability;
+boolean_param("allow_invalid_cacheability", allow_invalid_cacheability);
 #endif
 
 /* Override macros from asm/page.h to make them work with mfn_t */
@@ -1343,7 +1345,33 @@ static int promote_l1_table(struct page_info *page)
         }
         else
         {
-            switch ( ret = get_page_from_l1e(pl1e[i], d, d) )
+            l1_pgentry_t l1e = pl1e[i];
+
+            if ( !allow_invalid_cacheability )
+            {
+                switch ( l1e.l1 & PAGE_CACHE_ATTRS )
+                {
+                case _PAGE_WB:
+                case _PAGE_UC:
+                case _PAGE_UCM:
+                case _PAGE_WC:
+                case _PAGE_WT:
+                case _PAGE_WP:
+                    break;
+                default:
+                    /*
+                     * If we get here, a PV guest tried to use one of the
+                     * reserved values in Xen's PAT.  This indicates a bug in
+                     * the guest, so inject #GP to cause the guest to log a
+                     * stack trace.
+                     */
+                    pv_inject_hw_exception(TRAP_gp_fault, 0);
+                    ret = -EINVAL;
+                    goto fail;
+                }
+            }
+
+            switch ( ret = get_page_from_l1e(l1e, d, d) )
             {
             default:
                 goto fail;
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#19 lore raw

Re: [PATCH v5 09/10] x86/mm: Reject invalid cacheability in PV guests by default

xen/arch/x86/mm.c line 148

Quoted 7 lines · Demi Marie Obenour
xen/arch/x86/mm.c +1 -0
@@ -145,6 +145,8 @@
 
 #ifdef CONFIG_PV
 #include "pv/mm.h"
+bool allow_invalid_cacheability;

static and __ro_after_init please (the former not the least with Misra in mind).

xen/arch/x86/mm.c line 150

Quoted 2 lines
xen/arch/x86/mm.c +1 -0
+boolean_param("allow_invalid_cacheability", allow_invalid_cacheability);
 #endif

Any new command line option will need to come with an entry in docs/misc/xen-command-line.pandoc. Furthermore we're trying to avoid underscores in command line option names, when dashes serve the purpose quite fine.

Also please add a blank line at least between #include and your addition. Personally I would find it more natural if such a single-use control was defined next to the place it is used, i.e.

xen/arch/x86/mm.c static int promote_l1_table(struct page_info *page) · line 1347

Quoted 4 lines
xen/arch/x86/mm.c +0 -0
@@ -1343,7 +1345,33 @@ static int promote_l1_table(struct page_info *page)
         }
         else
         {

... here.

Quoted 5 lines
xen/arch/x86/mm.c +4 -1
-            switch ( ret = get_page_from_l1e(pl1e[i], d, d) )
+            l1_pgentry_t l1e = pl1e[i];
+
+            if ( !allow_invalid_cacheability )
+            {

xen/arch/x86/mm.c static int promote_l1_table(struct page_info *page) · line 1361

Quoted 10 lines
xen/arch/x86/mm.c +10 -0
+                switch ( l1e.l1 & PAGE_CACHE_ATTRS )
+                {
+                case _PAGE_WB:
+                case _PAGE_UC:
+                case _PAGE_UCM:
+                case _PAGE_WC:
+                case _PAGE_WT:
+                case _PAGE_WP:
+                    break;
+                default:

Nit (style): Blank line please between non-fall-through case blocks.

xen/arch/x86/mm.c static int promote_l1_table(struct page_info *page) · line 1367

Quoted 6 lines
xen/arch/x86/mm.c +6 -0
+                    /*
+                     * If we get here, a PV guest tried to use one of the
+                     * reserved values in Xen's PAT.  This indicates a bug in
+                     * the guest, so inject #GP to cause the guest to log a
+                     * stack trace.
+                     */

And likely make it die. Which, yes, ...

xen/arch/x86/mm.c static int promote_l1_table(struct page_info *page) · line 1369

Quoted 2 lines
xen/arch/x86/mm.c +2 -0
+                    pv_inject_hw_exception(TRAP_gp_fault, 0);
+                    ret = -EINVAL;

... also may or may not be the result of returning failure here (if the guest "checks" for errors by using a BUG()-like construct), but that's then more of its own fault than not coping with a non-architectural #GP.

Also wasn't there talk of having this behavior in debug hypervisors only? Without that restriction I'm yet less happy with the change ...

Jan

#20 lore raw

Re: [PATCH v5 09/10] x86/mm: Reject invalid cacheability in PV guests by default

Quoted 11 lines · Jan Beulich

xen/arch/x86/mm.c line 148

Quoted 7 lines · Demi Marie Obenour
xen/arch/x86/mm.c +1 -0
@@ -145,6 +145,8 @@
 
 #ifdef CONFIG_PV
 #include "pv/mm.h"
+bool allow_invalid_cacheability;

static and __ro_after_init please (the former not the least with Misra in mind).

Will fix in v6.

Quoted 7 lines

xen/arch/x86/mm.c line 150

Quoted 2 lines
xen/arch/x86/mm.c +1 -0
+boolean_param("allow_invalid_cacheability", allow_invalid_cacheability);
 #endif

Any new command line option will need to come with an entry in docs/misc/xen-command-line.pandoc. Furthermore we're trying to avoid underscores in command line option names, when dashes serve the purpose quite fine.

Will fix in v6.

Quoted 3 lines

Also please add a blank line at least between #include and your addition. Personally I would find it more natural if such a single-use control was defined next to the place it is used, i.e.

Will fix in v6.

Quoted 6 lines

xen/arch/x86/mm.c static int promote_l1_table(struct page_info *page) · line 1347

Quoted 4 lines
xen/arch/x86/mm.c +0 -0
@@ -1343,7 +1345,33 @@ static int promote_l1_table(struct page_info *page)
         }
         else
         {

... here.

Will move in v6.

Quoted 16 lines
Quoted 5 lines
xen/arch/x86/mm.c +4 -1
-            switch ( ret = get_page_from_l1e(pl1e[i], d, d) )
+            l1_pgentry_t l1e = pl1e[i];
+
+            if ( !allow_invalid_cacheability )
+            {

xen/arch/x86/mm.c static int promote_l1_table(struct page_info *page) · line 1361

Quoted 10 lines
xen/arch/x86/mm.c +10 -0
+                switch ( l1e.l1 & PAGE_CACHE_ATTRS )
+                {
+                case _PAGE_WB:
+                case _PAGE_UC:
+                case _PAGE_UCM:
+                case _PAGE_WC:
+                case _PAGE_WT:
+                case _PAGE_WP:
+                    break;
+                default:
Quoted 1 line

Nit (style): Blank line please between non-fall-through case blocks.

Will fix in v6.

Quoted 7 lines

xen/arch/x86/mm.c static int promote_l1_table(struct page_info *page) · line 1367

Quoted 6 lines
xen/arch/x86/mm.c +6 -0
+                    /*
+                     * If we get here, a PV guest tried to use one of the
+                     * reserved values in Xen's PAT.  This indicates a bug in
+                     * the guest, so inject #GP to cause the guest to log a
+                     * stack trace.
+                     */
Quoted 8 lines

And likely make it die. Which, yes, ...

xen/arch/x86/mm.c static int promote_l1_table(struct page_info *page) · line 1369

Quoted 2 lines
xen/arch/x86/mm.c +2 -0
+                    pv_inject_hw_exception(TRAP_gp_fault, 0);
+                    ret = -EINVAL;

... also may or may not be the result of returning failure here (if the guest "checks" for errors by using a BUG()-like construct), but that's then more of its own fault than not coping with a non-architectural #GP.

Is there really any architectural behavior here?

Quoted 2 lines

Also wasn't there talk of having this behavior in debug hypervisors only? Without that restriction I'm yet less happy with the change ...

My understanding was that Andrew preferred the behavior to be turned on in release hypervisors too. I am inclined to agree with Andrew, if for no other reason than that those working on guest OSs do not generally use debug hypervisors if they are not also working on Xen itself.

Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#21 lore raw

Re: [PATCH v5 09/10] x86/mm: Reject invalid cacheability in PV guests by default

Quoted 8 lines · Demi Marie Obenour
Quoted 2 lines · Jan Beulich

Also wasn't there talk of having this behavior in debug hypervisors only? Without that restriction I'm yet less happy with the change ...

My understanding was that Andrew preferred the behavior to be turned on in release hypervisors too. I am inclined to agree with Andrew, if for no other reason than that those working on guest OSs do not generally use debug hypervisors if they are not also working on Xen itself.

That's likely a mistake then, at least for work on PV guest OSes. In particular PV memory management code is sprinkled with gdprintk(), in an attempt to help those people understand why some operation has failed.

Jan

#22 lore raw

PATCH v5 10/10 x86: Use Linux's PAT

In reply to Demi Marie Obenour

This is purely for testing, to see if it works around a bug in i915. It is not intended to be merged.

NOT-signed-off-by: DO NOT MERGE

---
 xen/arch/x86/include/asm/page.h      |  4 ++--
 xen/arch/x86/include/asm/processor.h | 10 +++++-----
 xen/arch/x86/mm.c                    |  8 --------
 3 files changed, 7 insertions(+), 15 deletions(-)
xen/arch/x86/include/asm/page.h +2 -2
index b585235d064a567082582c8e92a4e8283fd949ca..ab9b46f1d0901e50a83fd035ff28d1bda0b781a2 100644
@@ -333,11 +333,11 @@ void efi_update_l4_pgtable(unsigned int l4idx, l4_pgentry_t);
 
 /* Memory types, encoded under Xen's choice of MSR_PAT. */
 #define _PAGE_WB         (                                0)
-#define _PAGE_WT         (                        _PAGE_PWT)
+#define _PAGE_WC         (                        _PAGE_PWT)
 #define _PAGE_UCM        (            _PAGE_PCD            )
 #define _PAGE_UC         (            _PAGE_PCD | _PAGE_PWT)
-#define _PAGE_WC         (_PAGE_PAT                        )
 #define _PAGE_WP         (_PAGE_PAT |             _PAGE_PWT)
+#define _PAGE_WT         (_PAGE_PAT | _PAGE_PCD | _PAGE_PWT)
 
 /*
  * Debug option: Ensure that granted mappings are not implicitly unmapped.
xen/arch/x86/include/asm/processor.h +5 -5
index 60b902060914584957db8afa5c7c1e6abdad4d13..3993d5638626f0948bb7ac8192d2eda187eb1bdb 100644
@@ -94,16 +94,16 @@
 
 /*
  * Host IA32_CR_PAT value to cover all memory types.  This is not the default
- * MSR_PAT value, and is an ABI with PV guests.
+ * MSR_PAT value, and is needed by the Linux i915 driver.
  */
 #define XEN_MSR_PAT ((_AC(X86_MT_WB,  ULL) << 0x00) | \
-                     (_AC(X86_MT_WT,  ULL) << 0x08) | \
+                     (_AC(X86_MT_WC,  ULL) << 0x08) | \
                      (_AC(X86_MT_UCM, ULL) << 0x10) | \
                      (_AC(X86_MT_UC,  ULL) << 0x18) | \
-                     (_AC(X86_MT_WC,  ULL) << 0x20) | \
+                     (_AC(X86_MT_WB,  ULL) << 0x20) | \
                      (_AC(X86_MT_WP,  ULL) << 0x28) | \
-                     (_AC(X86_MT_UC,  ULL) << 0x30) | \
-                     (_AC(X86_MT_UC,  ULL) << 0x38))
+                     (_AC(X86_MT_UCM, ULL) << 0x30) | \
+                     (_AC(X86_MT_WT,  ULL) << 0x38))
 
 #ifndef __ASSEMBLY__
 
xen/arch/x86/mm.c +0 -8
index 69ce597c7cd5283ae4b5f3bc0a6dfa0bb3228d3d..c536f7807a418c160366c22b6c4f937a5023f14b 100644
@@ -6387,14 +6387,6 @@ unsigned long get_upper_mfn_bound(void)
  */
 static void __init __maybe_unused build_assertions(void)
 {
-    /*
-     * If this trips, any guest that blindly rely on the public API in xen.h
-     * (instead of reading the PAT from Xen, as Linux 3.19+ does) will be
-     * broken.  Furthermore, live migration of PV guests between Xen versions
-     * using different PATs will not work.
-     */
-    BUILD_BUG_ON(XEN_MSR_PAT != 0x050100070406ULL);
-
     /*
      * _PAGE_WB must be zero for several reasons, not least because Linux
      * assumes it.
Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#23 lore raw

[PATCH v5 11/10] hvmloader: use memory type constants

In reply to Demi Marie Obenour

Now that we have them available in a header which is okay to use from hvmloader sources, do away with respective literal numbers and silent assumptions.

tools/firmware/hvmloader/cacheattr.c +17 -6
@@ -22,6 +22,8 @@
 #include "util.h"
 #include "config.h"
 
+#include <xen/asm/x86-defns.h>
+
 #define MSR_MTRRphysBase(reg) (0x200 + 2 * (reg))
 #define MSR_MTRRphysMask(reg) (0x200 + 2 * (reg) + 1)
 #define MSR_MTRRcap          0x00fe
@@ -71,23 +73,32 @@ void cacheattr_init(void)
 
     addr_mask = ((1ull << phys_bits) - 1) & ~((1ull << 12) - 1);
     mtrr_cap = rdmsr(MSR_MTRRcap);
-    mtrr_def = (1u << 11) | 6; /* E, default type WB */
+    mtrr_def = (1u << 11) | X86_MT_WB; /* E, default type WB */
 
     /* Fixed-range MTRRs supported? */
     if ( mtrr_cap & (1u << 8) )
     {
+#define BCST2(mt) ((mt) | ((mt) << 8))
+#define BCST4(mt) (BCST2(mt) | (BCST2(mt) << 16))
+#define BCST8(mt) (BCST4(mt) | ((uint64_t)BCST4(mt) << 32))
         /* 0x00000-0x9ffff: Write Back (WB) */
-        content = 0x0606060606060606ull;
+        content = BCST8(X86_MT_WB);
         wrmsr(MSR_MTRRfix64K_00000, content);
         wrmsr(MSR_MTRRfix16K_80000, content);
+
         /* 0xa0000-0xbffff: Write Combining (WC) */
         if ( mtrr_cap & (1u << 10) ) /* WC supported? */
-            content = 0x0101010101010101ull;
+            content = BCST8(X86_MT_WC);
         wrmsr(MSR_MTRRfix16K_A0000, content);
+
         /* 0xc0000-0xfffff: Write Back (WB) */
-        content = 0x0606060606060606ull;
+        content = BCST8(X86_MT_WB);
         for ( i = 0; i < 8; i++ )
             wrmsr(MSR_MTRRfix4K_C0000 + i, content);
+#undef BCST8
+#undef BCST4
+#undef BCST2
+
         mtrr_def |= 1u << 10; /* FE */
         printf("fixed MTRRs ... ");
     }
@@ -106,7 +117,7 @@ void cacheattr_init(void)
             while ( ((base + size) < base) || ((base + size) > pci_mem_end) )
                 size >>= 1;
 
-            wrmsr(MSR_MTRRphysBase(i), base);
+            wrmsr(MSR_MTRRphysBase(i), base | X86_MT_UC);
             wrmsr(MSR_MTRRphysMask(i), (~(size - 1) & addr_mask) | (1u << 11));
 
             base += size;
@@ -121,7 +132,7 @@ void cacheattr_init(void)
             while ( (base + size < base) || (base + size > pci_hi_mem_end) )
                 size >>= 1;
 
-            wrmsr(MSR_MTRRphysBase(i), base);
+            wrmsr(MSR_MTRRphysBase(i), base | X86_MT_UC);
             wrmsr(MSR_MTRRphysMask(i), (~(size - 1) & addr_mask) | (1u << 11));
 
             base += size;
#24 lore raw

Re: [PATCH v5 11/10] hvmloader: use memory type constants

Quoted 18 lines · Jan Beulich

Now that we have them available in a header which is okay to use from hvmloader sources, do away with respective literal numbers and silent assumptions.

tools/firmware/hvmloader/cacheattr.c +2 -0
@@ -22,6 +22,8 @@
 #include "util.h"
 #include "config.h"
 
+#include <xen/asm/x86-defns.h>
+
 #define MSR_MTRRphysBase(reg) (0x200 + 2 * (reg))
 #define MSR_MTRRphysMask(reg) (0x200 + 2 * (reg) + 1)
 #define MSR_MTRRcap          0x00fe

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 74

Quoted 2 lines
tools/firmware/hvmloader/cacheattr.c +0 -0
@@ -71,23 +73,32 @@ void cacheattr_init(void)
 
     addr_mask = ((1ull << phys_bits) - 1) & ~((1ull << 12) - 1);

Does this want to be PAGE_SHIFT ?

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 83

Quoted 10 lines
tools/firmware/hvmloader/cacheattr.c +4 -1
     mtrr_cap = rdmsr(MSR_MTRRcap);
-    mtrr_def = (1u << 11) | 6; /* E, default type WB */
+    mtrr_def = (1u << 11) | X86_MT_WB; /* E, default type WB */
 
     /* Fixed-range MTRRs supported? */
     if ( mtrr_cap & (1u << 8) )
     {
+#define BCST2(mt) ((mt) | ((mt) << 8))
+#define BCST4(mt) (BCST2(mt) | (BCST2(mt) << 16))
+#define BCST8(mt) (BCST4(mt) | ((uint64_t)BCST4(mt) << 32))

IMO this is clearer as

#define BCST8(mt) ((mt) * 0x0101010101010101ULL)

which saves several macros.

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 92

Quoted 11 lines
tools/firmware/hvmloader/cacheattr.c +3 -2
         /* 0x00000-0x9ffff: Write Back (WB) */
-        content = 0x0606060606060606ull;
+        content = BCST8(X86_MT_WB);
         wrmsr(MSR_MTRRfix64K_00000, content);
         wrmsr(MSR_MTRRfix16K_80000, content);
+
         /* 0xa0000-0xbffff: Write Combining (WC) */
         if ( mtrr_cap & (1u << 10) ) /* WC supported? */
-            content = 0x0101010101010101ull;
+            content = BCST8(X86_MT_WC);
         wrmsr(MSR_MTRRfix16K_A0000, content);

This looks like it's latently buggy.  We'll end up with WB if WC isn't supported, when it ought to be UC.

I suppose it doesn't actually matter because if there is a VGA region there, it will actually be holes in the P2M and trap for emulation.

Also, Xen being 64-bit, WC is always available.

Quoted 14 lines
tools/firmware/hvmloader/cacheattr.c +6 -1
+
         /* 0xc0000-0xfffff: Write Back (WB) */
-        content = 0x0606060606060606ull;
+        content = BCST8(X86_MT_WB);
         for ( i = 0; i < 8; i++ )
             wrmsr(MSR_MTRRfix4K_C0000 + i, content);
+#undef BCST8
+#undef BCST4
+#undef BCST2
+
         mtrr_def |= 1u << 10; /* FE */
         printf("fixed MTRRs ... ");
     }

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 121

Quoted 6 lines
tools/firmware/hvmloader/cacheattr.c +1 -1
@@ -106,7 +117,7 @@ void cacheattr_init(void)
             while ( ((base + size) < base) || ((base + size) > pci_mem_end) )
                 size >>= 1;
 
-            wrmsr(MSR_MTRRphysBase(i), base);
+            wrmsr(MSR_MTRRphysBase(i), base | X86_MT_UC);
             wrmsr(MSR_MTRRphysMask(i), (~(size - 1) & addr_mask) | (1u << 11));

If you're re-spinning for page_size or the macros, any chance of also gaining some constants for E/FE to avoid these naked shifts?

~Andrew

Quoted 13 lines
tools/firmware/hvmloader/cacheattr.c +1 -1
             base += size;
@@ -121,7 +132,7 @@ void cacheattr_init(void)
             while ( (base + size < base) || (base + size > pci_hi_mem_end) )
                 size >>= 1;
 
-            wrmsr(MSR_MTRRphysBase(i), base);
+            wrmsr(MSR_MTRRphysBase(i), base | X86_MT_UC);
             wrmsr(MSR_MTRRphysMask(i), (~(size - 1) & addr_mask) | (1u << 11));
 
             base += size;
#25 lore raw

Re: [PATCH v5 11/10] hvmloader: use memory type constants

Quoted 14 lines · Andrew Cooper
Quoted 13 lines · Jan Beulich
tools/firmware/hvmloader/cacheattr.c +2 -0
@@ -22,6 +22,8 @@
 #include "util.h"
 #include "config.h"
 
+#include <xen/asm/x86-defns.h>
+
 #define MSR_MTRRphysBase(reg) (0x200 + 2 * (reg))
 #define MSR_MTRRphysMask(reg) (0x200 + 2 * (reg) + 1)
 #define MSR_MTRRcap          0x00fe
@@ -71,23 +73,32 @@ void cacheattr_init(void)
 
Quoted 3 lines

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 74

Quoted 1 line
tools/firmware/hvmloader/cacheattr.c +0 -0
     addr_mask = ((1ull << phys_bits) - 1) & ~((1ull << 12) - 1);

Does this want to be PAGE_SHIFT ?

Not really (it's rather an MTRR layout related constant which we ought to use here, which only happens to match PAGE_{SIZE,SHIFT}) and not in this patch (see the title).

Quoted 13 lines

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 83

Quoted 10 lines
tools/firmware/hvmloader/cacheattr.c +4 -1
     mtrr_cap = rdmsr(MSR_MTRRcap);
-    mtrr_def = (1u << 11) | 6; /* E, default type WB */
+    mtrr_def = (1u << 11) | X86_MT_WB; /* E, default type WB */
 
     /* Fixed-range MTRRs supported? */
     if ( mtrr_cap & (1u << 8) )
     {
+#define BCST2(mt) ((mt) | ((mt) << 8))
+#define BCST4(mt) (BCST2(mt) | (BCST2(mt) << 16))
+#define BCST8(mt) (BCST4(mt) | ((uint64_t)BCST4(mt) << 32))

IMO this is clearer as

Quoted 3 lines

#define BCST8(mt) ((mt) * 0x0101010101010101ULL)

which saves several macros.

Ah, yes, will switch (and then name the thing just BCST()).

Quoted 12 lines

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 92

Quoted 11 lines
tools/firmware/hvmloader/cacheattr.c +3 -2
         /* 0x00000-0x9ffff: Write Back (WB) */
-        content = 0x0606060606060606ull;
+        content = BCST8(X86_MT_WB);
         wrmsr(MSR_MTRRfix64K_00000, content);
         wrmsr(MSR_MTRRfix16K_80000, content);
+
         /* 0xa0000-0xbffff: Write Combining (WC) */
         if ( mtrr_cap & (1u << 10) ) /* WC supported? */
-            content = 0x0101010101010101ull;
+            content = BCST8(X86_MT_WC);
         wrmsr(MSR_MTRRfix16K_A0000, content);
Quoted 7 lines

This looks like it's latently buggy.  We'll end up with WB if WC isn't supported, when it ought to be UC.

I suppose it doesn't actually matter because if there is a VGA region there, it will actually be holes in the P2M and trap for emulation.

Also, Xen being 64-bit, WC is always available.

Right, but we (or the admin) may elect to not surface availability to the guest.

Quoted 10 lines

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 121

Quoted 7 lines
tools/firmware/hvmloader/cacheattr.c +1 -1
@@ -106,7 +117,7 @@ void cacheattr_init(void)
             while ( ((base + size) < base) || ((base + size) > pci_mem_end) )
                 size >>= 1;
 
-            wrmsr(MSR_MTRRphysBase(i), base);
+            wrmsr(MSR_MTRRphysBase(i), base | X86_MT_UC);
             wrmsr(MSR_MTRRphysMask(i), (~(size - 1) & addr_mask) | (1u << 11));

If you're re-spinning for page_size or the macros, any chance of also gaining some constants for E/FE to avoid these naked shifts?

Again, yes in principle, but not in this patch. This requires shuffling more stuff into x86-defns.h first.

Jan

author Acked-by
#26 lore raw

Re: [PATCH v5 11/10] hvmloader: use memory type constants

In reply to Jan Beulich
Quoted 19 lines · Jan Beulich

Now that we have them available in a header which is okay to use from hvmloader sources, do away with respective literal numbers and silent assumptions.

tools/firmware/hvmloader/cacheattr.c +2 -0
@@ -22,6 +22,8 @@
 #include "util.h"
 #include "config.h"
 
+#include <xen/asm/x86-defns.h>
+
 #define MSR_MTRRphysBase(reg) (0x200 + 2 * (reg))
 #define MSR_MTRRphysMask(reg) (0x200 + 2 * (reg) + 1)
 #define MSR_MTRRcap          0x00fe
@@ -71,23 +73,32 @@ void cacheattr_init(void)
 

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 82

Quoted 10 lines
tools/firmware/hvmloader/cacheattr.c +3 -1
     addr_mask = ((1ull << phys_bits) - 1) & ~((1ull << 12) - 1);
     mtrr_cap = rdmsr(MSR_MTRRcap);
-    mtrr_def = (1u << 11) | 6; /* E, default type WB */
+    mtrr_def = (1u << 11) | X86_MT_WB; /* E, default type WB */
 
     /* Fixed-range MTRRs supported? */
     if ( mtrr_cap & (1u << 8) )
     {
+#define BCST2(mt) ((mt) | ((mt) << 8))
+#define BCST4(mt) (BCST2(mt) | (BCST2(mt) << 16))

This should include a cast to uint32_t, just like BCST8 includes a cast to uint64_t. It doesn’t have any functional impact since none of the memory types have the high bit set, but it makes the correctness of the code much more obvious to readers.

Quoted 36 lines
tools/firmware/hvmloader/cacheattr.c +11 -4
+#define BCST8(mt) (BCST4(mt) | ((uint64_t)BCST4(mt) << 32))
         /* 0x00000-0x9ffff: Write Back (WB) */
-        content = 0x0606060606060606ull;
+        content = BCST8(X86_MT_WB);
         wrmsr(MSR_MTRRfix64K_00000, content);
         wrmsr(MSR_MTRRfix16K_80000, content);
+
         /* 0xa0000-0xbffff: Write Combining (WC) */
         if ( mtrr_cap & (1u << 10) ) /* WC supported? */
-            content = 0x0101010101010101ull;
+            content = BCST8(X86_MT_WC);
         wrmsr(MSR_MTRRfix16K_A0000, content);
+
         /* 0xc0000-0xfffff: Write Back (WB) */
-        content = 0x0606060606060606ull;
+        content = BCST8(X86_MT_WB);
         for ( i = 0; i < 8; i++ )
             wrmsr(MSR_MTRRfix4K_C0000 + i, content);
+#undef BCST8
+#undef BCST4
+#undef BCST2
+
         mtrr_def |= 1u << 10; /* FE */
         printf("fixed MTRRs ... ");
     }
@@ -106,7 +117,7 @@ void cacheattr_init(void)
             while ( ((base + size) < base) || ((base + size) > pci_mem_end) )
                 size >>= 1;
 
-            wrmsr(MSR_MTRRphysBase(i), base);
+            wrmsr(MSR_MTRRphysBase(i), base | X86_MT_UC);
             wrmsr(MSR_MTRRphysMask(i), (~(size - 1) & addr_mask) | (1u << 11));
 
             base += size;

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 138

Quoted 8 lines
tools/firmware/hvmloader/cacheattr.c +1 -1
@@ -121,7 +132,7 @@ void cacheattr_init(void)
             while ( (base + size < base) || (base + size > pci_hi_mem_end) )
                 size >>= 1;
 
-            wrmsr(MSR_MTRRphysBase(i), base);
+            wrmsr(MSR_MTRRphysBase(i), base | X86_MT_UC);
             wrmsr(MSR_MTRRphysMask(i), (~(size - 1) & addr_mask) | (1u << 11));
 
             base += size;

With the suggested change:

Signature · 3 lines
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab
#27 lore raw

Re: [PATCH v5 11/10] hvmloader: use memory type constants

Quoted 19 lines · Demi Marie Obenour
Quoted 18 lines · Jan Beulich
tools/firmware/hvmloader/cacheattr.c +3 -1
@@ -22,6 +22,8 @@
 #include "util.h"
 #include "config.h"
 
+#include <xen/asm/x86-defns.h>
+
 #define MSR_MTRRphysBase(reg) (0x200 + 2 * (reg))
 #define MSR_MTRRphysMask(reg) (0x200 + 2 * (reg) + 1)
 #define MSR_MTRRcap          0x00fe
@@ -71,23 +73,32 @@ void cacheattr_init(void)
 
     addr_mask = ((1ull << phys_bits) - 1) & ~((1ull << 12) - 1);
     mtrr_cap = rdmsr(MSR_MTRRcap);
-    mtrr_def = (1u << 11) | 6; /* E, default type WB */
+    mtrr_def = (1u << 11) | X86_MT_WB; /* E, default type WB */
 
Quoted 10 lines

tools/firmware/hvmloader/cacheattr.c void cacheattr_init(void) · line 82

Quoted 5 lines
tools/firmware/hvmloader/cacheattr.c +2 -0
     /* Fixed-range MTRRs supported? */
     if ( mtrr_cap & (1u << 8) )
     {
+#define BCST2(mt) ((mt) | ((mt) << 8))
+#define BCST4(mt) (BCST2(mt) | (BCST2(mt) << 16))

This should include a cast to uint32_t, just like BCST8 includes a cast to uint64_t. It doesn’t have any functional impact since none of the memory types have the high bit set, but it makes the correctness of the code much more obvious to readers.

I've already followed Andrew's suggestion.

Quoted 3 lines

With the suggested change:

Thanks. Since I've addressed this differently, I'll hold back applying of this.

From a formal perspective I'd also like to point out that an Acked-by:
from a person not being maintainer of any code being changed by a patch has no meaning at all. Only Reviewed-by: has a meaning (but of course implies more thorough looking at the change than Acked-by: does).

Jan