public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: "Ard Biesheuvel" <ard.biesheuvel@linaro.org>
To: devel@edk2.groups.io
Cc: leif@nuviainc.com, Ard Biesheuvel <ard.biesheuvel@linaro.org>
Subject: [PATCH v2 4/9] ArmPkg/ArmMmuLib ARM: cache-invalidate initial page table entries
Date: Wed,  4 Mar 2020 19:12:41 +0100	[thread overview]
Message-ID: <20200304181246.23513-5-ard.biesheuvel@linaro.org> (raw)
In-Reply-To: <20200304181246.23513-1-ard.biesheuvel@linaro.org>

In the ARM version of ArmMmuLib, we are currently relying on set/way
invalidation to ensure that the caches are in a consistent state with
respect to main memory once we turn the MMU on. Even if set/way
operations were the appropriate method to achieve this, doing an
invalidate-all first and then populating the page table entries creates
a window where page table entries could be loaded speculatively into
the caches before we modify them, and shadow the new values that we
write there.

So let's get rid of the blanket clean/invalidate operations, and instead,
invalidate each section entry before and after it is updated (to address
all the little corner cases that the ARMv7 spec permits), and invalidate
sets of level 2 entries in blocks, using the generic invalidation routine
from CacheMaintenanceLib

On ARMv7, cache maintenance may be required also when the MMU is
enabled, in case the page table walker is not cache coherent. However,
the code being updated here is guaranteed to run only when the MMU is
still off, and so we can disregard the case when the MMU and caches
are on.

Since the MMU and D-cache are already off when we reach this point, we
can drop the MMU and D-cache disables as well. Maintenance of the I-cache
is unnecessary, since we are not modifying any code, and the installed
mapping is guaranteed to be 1:1. This means we can also leave it enabled
while the page table population code is running.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 ArmPkg/Library/ArmMmuLib/Arm/ArmMmuLibCore.c | 55 +++++++++++++++-----
 1 file changed, 41 insertions(+), 14 deletions(-)

diff --git a/ArmPkg/Library/ArmMmuLib/Arm/ArmMmuLibCore.c b/ArmPkg/Library/ArmMmuLib/Arm/ArmMmuLibCore.c
index aca7a37facac..7c7cad2c3d9d 100644
--- a/ArmPkg/Library/ArmMmuLib/Arm/ArmMmuLibCore.c
+++ b/ArmPkg/Library/ArmMmuLib/Arm/ArmMmuLibCore.c
@@ -178,11 +178,25 @@ PopulateLevel2PageTable (
 
   ASSERT (FirstPageOffset + Pages <= TRANSLATION_TABLE_PAGE_COUNT);
 
+  //
+  // Invalidate once to prevent page table updates to hit in the
+  // caches inadvertently.
+  //
+  InvalidateDataCacheRange ((UINT32 *)TranslationTable + FirstPageOffset,
+    RemainLength / TT_DESCRIPTOR_PAGE_SIZE * sizeof (*PageEntry));
+
   for (Index = 0; Index < Pages; Index++) {
     *PageEntry++     =  TT_DESCRIPTOR_PAGE_BASE_ADDRESS(PhysicalBase) | PageAttributes;
     PhysicalBase += TT_DESCRIPTOR_PAGE_SIZE;
   }
 
+  //
+  // Invalidate again to ensure that any line fetches that may have occurred
+  // [speculatively] since the previous invalidate are evicted again.
+  //
+  ArmDataMemoryBarrier ();
+  InvalidateDataCacheRange ((UINT32 *)TranslationTable + FirstPageOffset,
+    RemainLength / TT_DESCRIPTOR_PAGE_SIZE * sizeof (*PageEntry));
 }
 
 STATIC
@@ -253,11 +267,28 @@ FillTranslationTable (
   SectionEntry    = TRANSLATION_TABLE_ENTRY_FOR_VIRTUAL_ADDRESS(TranslationTable, MemoryRegion->VirtualBase);
 
   while (RemainLength != 0) {
+    //
+    // Ensure that the assignment of the page table entry will not hit
+    // in the cache. Whether this could occur is IMPLEMENTATION DEFINED
+    // and thus permitted by the ARMv7 architecture.
+    //
+    ArmInvalidateDataCacheEntryByMVA ((UINTN)SectionEntry);
+    ArmDataSynchronizationBarrier ();
+
     if (PhysicalBase % TT_DESCRIPTOR_SECTION_SIZE == 0 &&
         RemainLength >= TT_DESCRIPTOR_SECTION_SIZE) {
       // Case: Physical address aligned on the Section Size (1MB) && the length
       // is greater than the Section Size
-      *SectionEntry++ = TT_DESCRIPTOR_SECTION_BASE_ADDRESS(PhysicalBase) | Attributes;
+      *SectionEntry = TT_DESCRIPTOR_SECTION_BASE_ADDRESS(PhysicalBase) | Attributes;
+
+      //
+      // Issue a DMB to ensure that the page table entry update made it to
+      // memory before we issue the invalidate, otherwise, a subsequent
+      // speculative fetch could observe the old value.
+      //
+      ArmDataMemoryBarrier ();
+      ArmInvalidateDataCacheEntryByMVA ((UINTN)SectionEntry++);
+
       PhysicalBase += TT_DESCRIPTOR_SECTION_SIZE;
       RemainLength -= TT_DESCRIPTOR_SECTION_SIZE;
     } else {
@@ -267,9 +298,17 @@ FillTranslationTable (
       // Case: Physical address aligned on the Section Size (1MB) && the length
       //       does not fill a section
       // Case: Physical address NOT aligned on the Section Size (1MB)
-      PopulateLevel2PageTable (SectionEntry++, PhysicalBase, PageMapLength,
+      PopulateLevel2PageTable (SectionEntry, PhysicalBase, PageMapLength,
         MemoryRegion->Attributes);
 
+      //
+      // Issue a DMB to ensure that the page table entry update made it to
+      // memory before we issue the invalidate, otherwise, a subsequent
+      // speculative fetch could observe the old value.
+      //
+      ArmDataMemoryBarrier ();
+      ArmInvalidateDataCacheEntryByMVA ((UINTN)SectionEntry++);
+
       // If it is the last entry
       if (RemainLength < TT_DESCRIPTOR_SECTION_SIZE) {
         break;
@@ -349,18 +388,6 @@ ArmConfigureMmu (
     }
   }
 
-  ArmCleanInvalidateDataCache ();
-  ArmInvalidateInstructionCache ();
-
-  ArmDisableDataCache ();
-  ArmDisableInstructionCache();
-  // TLBs are also invalidated when calling ArmDisableMmu()
-  ArmDisableMmu ();
-
-  // Make sure nothing sneaked into the cache
-  ArmCleanInvalidateDataCache ();
-  ArmInvalidateInstructionCache ();
-
   ArmSetTTBR0 ((VOID *)(UINTN)(((UINTN)TranslationTable & ~TRANSLATION_TABLE_SECTION_ALIGNMENT_MASK) | (TTBRAttributes & 0x7F)));
 
   //
-- 
2.17.1


  parent reply	other threads:[~2020-03-04 18:13 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-04 18:12 [PATCH v2 0/9] ArmPkg: eradicate and deprecate by set/way cache ops Ard Biesheuvel
2020-03-04 18:12 ` [PATCH v2 1/9] ArmPlatformPkg/PrePi: replace set/way cache ops with by-VA ones Ard Biesheuvel
2020-03-05 16:23   ` Leif Lindholm
2020-03-04 18:12 ` [PATCH v2 2/9] ArmPkg/ArmMmuLib ARM: remove dummy constructor Ard Biesheuvel
2020-03-04 18:12 ` [PATCH v2 3/9] ArmPkg/ArmMmuLib ARM: split ArmMmuLibCore.c into core and update code Ard Biesheuvel
2020-03-04 18:12 ` Ard Biesheuvel [this message]
2020-03-04 18:12 ` [PATCH v2 5/9] ArmPkg/ArmMmuLib AARCH64: cache-invalidate initial page table entries Ard Biesheuvel
2020-03-04 18:12 ` [PATCH v2 6/9] ArmPkg/ArmLib: move set/way helper functions into private header Ard Biesheuvel
2020-03-04 18:12 ` [PATCH v2 7/9] ArmPkg/ArmLib: clean up library includes Ard Biesheuvel
2020-03-04 18:12 ` [PATCH v2 8/9] ArmPkg/ArmLib: remove bogus protocol declaration Ard Biesheuvel
2020-03-04 18:12 ` [PATCH v2 9/9] ArmPkg/ArmLib: ASSERT on set/way cache ops being used with MMU on Ard Biesheuvel
2020-03-05 16:29 ` [PATCH v2 0/9] ArmPkg: eradicate and deprecate by set/way cache ops Leif Lindholm
2020-03-05 21:40   ` Ard Biesheuvel

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-list from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200304181246.23513-5-ard.biesheuvel@linaro.org \
    --to=devel@edk2.groups.io \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox