Skip to contentSkip to conversation

x86/boot/64: Clear CR4.PGE to disable global 1:1 mappings

PATCH Superseded Ard Biesheuvel 3 participants 4 messages Apr 10, 2024 x86

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86/boot/64: Clear CR4.PGE to disable global 1:1 mappings

From: Ard Biesheuvel <[email protected]>

The early 64-bit boot code must be entered with a 1:1 mapping of the bootable image, but it cannot operate without a 1:1 mapping of all the assets in memory that it accesses, and therefore, it creates such mappings for all known assets upfront, and additional ones on demand when a page fault happens on a memory address.

These mappings are created with the global bit G set, as the flags used to create page table descriptors are based on __PAGE_KERNEL_LARGE_EXEC defined by the core kernel, even though the context where these mappings are used is very different.

This means that the TLB maintenance carried out by the decompressor is not sufficient if it is entered with CR4.PGE enabled, which has been observed to happen with the stage0 bootloader of project Oak. While this is a dubious practice if no global mappings are being used to begin with, the decompressor is clearly at fault here for creating global mappings and not performing the appropriate TLB maintenance.

Since commit

f97b67a773cd84b ("x86/decompressor: Only call the trampoline when changing paging levels")

CR4 is no longer modified by the decompressor if no change in the number of paging levels is needed. Before that, CR4 would always be set to a known value with PGE cleared.

So clear CR4.PGE explicitly in the decompressor before switching to the new 1:1 mapping which uses the G bit.

Cc Conrad Grobler <[email protected]>
Cc Kevin Loughlin <[email protected]>
Fixes f97b67a773cd84b ("x86/decompressor: Only call the trampoline when ...")
---
 arch/x86/boot/compressed/ident_map_64.c | 3 +++
 1 file changed, 3 insertions(+)
arch/x86/boot/compressed/ident_map_64.c +3 -0
index d040080d7edb..e746bf2efdf7 100644
@@ -179,6 +179,9 @@ void initialize_identity_maps(void *rmode)
 
 	sev_prep_identity_maps(top_level_pgt);
 
+	/* Disable global mappings */
+	asm("mov %0, %%cr4" :: "r"(native_read_cr4() & ~X86_CR4_PGE));
+
 	/* Load the new page-table. */
 	write_cr3(top_level_pgt);
 
Signature · 1 line
2.44.0.478.gd926399ef9-goog
#1 lore raw
Quoted 20 lines · Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

The early 64-bit boot code must be entered with a 1:1 mapping of the bootable image, but it cannot operate without a 1:1 mapping of all the assets in memory that it accesses, and therefore, it creates such mappings for all known assets upfront, and additional ones on demand when a page fault happens on a memory address.

These mappings are created with the global bit G set, as the flags used to create page table descriptors are based on __PAGE_KERNEL_LARGE_EXEC defined by the core kernel, even though the context where these mappings are used is very different.

This means that the TLB maintenance carried out by the decompressor is not sufficient if it is entered with CR4.PGE enabled, which has been observed to happen with the stage0 bootloader of project Oak. While this is a dubious practice if no global mappings are being used to begin with, the decompressor is clearly at fault here for creating global mappings and not performing the appropriate TLB maintenance.

Quoted 7 lines

Since commit

f97b67a773cd84b ("x86/decompressor: Only call the trampoline when changing paging levels")

CR4 is no longer modified by the decompressor if no change in the number of paging levels is needed. Before that, CR4 would always be set to a known value with PGE cleared.

So if we do this for robustness & historical pre-f97b67a773cd84b quirk-reliance's sake, I'd prefer if we loaded a known CR4 value again, instead of just turning off the PGE bit.

It's probably also a tiny bit faster, as no CR4 read has to be performed.

Thanks,

Ingo

#2 lore raw
Quoted 32 lines · Ingo Molnar
Quoted 20 lines · Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

The early 64-bit boot code must be entered with a 1:1 mapping of the bootable image, but it cannot operate without a 1:1 mapping of all the assets in memory that it accesses, and therefore, it creates such mappings for all known assets upfront, and additional ones on demand when a page fault happens on a memory address.

These mappings are created with the global bit G set, as the flags used to create page table descriptors are based on __PAGE_KERNEL_LARGE_EXEC defined by the core kernel, even though the context where these mappings are used is very different.

This means that the TLB maintenance carried out by the decompressor is not sufficient if it is entered with CR4.PGE enabled, which has been observed to happen with the stage0 bootloader of project Oak. While this is a dubious practice if no global mappings are being used to begin with, the decompressor is clearly at fault here for creating global mappings and not performing the appropriate TLB maintenance.

Quoted 7 lines

Since commit

f97b67a773cd84b ("x86/decompressor: Only call the trampoline when changing paging levels")

CR4 is no longer modified by the decompressor if no change in the number of paging levels is needed. Before that, CR4 would always be set to a known value with PGE cleared.

Quoted 6 lines

So if we do this for robustness & historical pre-f97b67a773cd84b quirk-reliance's sake, I'd prefer if we loaded a known CR4 value again, instead of just turning off the PGE bit.

It's probably also a tiny bit faster, as no CR4 read has to be performed.

Fair enough. I'll go and change that.

#3 lore raw
Quoted 33 lines · Ard Biesheuvel
Quoted 32 lines · Ingo Molnar
Quoted 27 lines · Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

The early 64-bit boot code must be entered with a 1:1 mapping of the bootable image, but it cannot operate without a 1:1 mapping of all the assets in memory that it accesses, and therefore, it creates such mappings for all known assets upfront, and additional ones on demand when a page fault happens on a memory address.

These mappings are created with the global bit G set, as the flags used to create page table descriptors are based on __PAGE_KERNEL_LARGE_EXEC defined by the core kernel, even though the context where these mappings are used is very different.

This means that the TLB maintenance carried out by the decompressor is not sufficient if it is entered with CR4.PGE enabled, which has been observed to happen with the stage0 bootloader of project Oak. While this is a dubious practice if no global mappings are being used to begin with, the decompressor is clearly at fault here for creating global mappings and not performing the appropriate TLB maintenance.

Since commit

f97b67a773cd84b ("x86/decompressor: Only call the trampoline when changing paging levels")

CR4 is no longer modified by the decompressor if no change in the number of paging levels is needed. Before that, CR4 would always be set to a known value with PGE cleared.

Quoted 8 lines
Quoted 6 lines

So if we do this for robustness & historical pre-f97b67a773cd84b quirk-reliance's sake, I'd prefer if we loaded a known CR4 value again, instead of just turning off the PGE bit.

It's probably also a tiny bit faster, as no CR4 read has to be performed.

Fair enough. I'll go and change that.

Thanks!

Ingo