Skip to contentSkip to conversation

x86: Rid .head.text of all abs references

RFC PATCH Superseded Ard Biesheuvel 2 participants 6 messages Mar 7, 2024

Subsystems (2)

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

cover x86: Rid .head.text of all abs references

From: Ard Biesheuvel <[email protected]>

Questions below!

This series removes the last remaining absolute symbol references from .head.text. Doing so is necessary because code in this section may be called from a 1:1 mapping of memory, which deviates from the mapping this code was linked and/or relocated to run at. This is not something that the toolchains support: even PIC/PIE code is still assumed to execute from the same mapping that it was relocated to run from by the startup code or dynamic loader. This means we are basically on our own here, and need to add measures to ensure the code works as expected in this manner. (This work was inspired by boot problems on Clang-built SEV-SNP guest kernels, where the confusion between RIP-relative and absolute references was causing variable accesses to fault)

Given that the startup code needs to create the kernel virtual mapping in the page tables, early references to some kernel virtual addresses are valid even if they cannot be dereferenced yet. To avoid having to make this distinction at build time, patches #3 and #4 replace such valid references with RIP-relative references with an offset applied.

Patches #1 and #2 remove some absolute references from .head.text that don't need to be there in the first place.

Questions:
- How can we police this at build time? Could we teach objtool to check for absolute ELF relocations in .head.text, or does this belong in modpost perhaps?

- Checking for absolute symbol references is not a complete solution, as .head.text code could call into other code as well. Do we need rigid checks for that too? Or could we have a soft rule that says you should only call __head code from __head code?

Cc Kevin Loughlin <[email protected]>
Cc Tom Lendacky <[email protected]>
Cc Dionna Glaze <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Dave Hansen <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Kees Cook <[email protected]>
Cc Brian Gerst <[email protected]>

Ard Biesheuvel (4):
x86/sev: Avoid WARN()s in early boot code x86/xen/pvh: Move startup code into .ref.text x86/boot/64: Determine VA/PA offset before entering C code x86/boot/64: Avoid intentional absolute symbol references in
.head.text

 arch/x86/include/asm/setup.h |  3 +-
 arch/x86/kernel/head64.c     | 38 ++++++++++++--------
 arch/x86/kernel/head_64.S    |  2 ++
 arch/x86/kernel/sev.c        | 15 +++-----
 arch/x86/platform/pvh/head.S |  2 +-
 5 files changed, 33 insertions(+), 27 deletions(-)

base-commit: 428080c9b19bfda37c478cd626dbd3851db1aff9

Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
#1 lore raw

RFC PATCH 1/4 x86/sev: Avoid WARN()s in early boot code

From: Ard Biesheuvel <[email protected]>

Using WARN() before the kernel is even mapped is unlikely to do anything
useful: the string literals are passed using their kernel virtual addresses which are not even mapped yet. But even if they were, calling into the printk machinery from the early 1:1 mapped code is not going to get very far.

So drop the WARN()s entirely.

---
 arch/x86/kernel/sev.c | 15 +++++----------
 1 file changed, 5 insertions(+), 10 deletions(-)
arch/x86/kernel/sev.c +5 -10
index 33c14aa1f06c..fc1b7b331815 100644
@@ -700,7 +700,7 @@ early_set_pages_state(unsigned long vaddr, unsigned long paddr,
 		if (op == SNP_PAGE_STATE_SHARED) {
 			/* Page validation must be rescinded before changing to shared */
 			ret = pvalidate(vaddr, RMP_PG_SIZE_4K, false);
-			if (WARN(ret, "Failed to validate address 0x%lx ret %d", paddr, ret))
+			if (ret)
 				goto e_term;
 		}
 
@@ -713,21 +713,16 @@ early_set_pages_state(unsigned long vaddr, unsigned long paddr,
 
 		val = sev_es_rd_ghcb_msr();
 
-		if (WARN(GHCB_RESP_CODE(val) != GHCB_MSR_PSC_RESP,
-			 "Wrong PSC response code: 0x%x\n",
-			 (unsigned int)GHCB_RESP_CODE(val)))
+		if (GHCB_RESP_CODE(val) != GHCB_MSR_PSC_RESP)
 			goto e_term;
 
-		if (WARN(GHCB_MSR_PSC_RESP_VAL(val),
-			 "Failed to change page state to '%s' paddr 0x%lx error 0x%llx\n",
-			 op == SNP_PAGE_STATE_PRIVATE ? "private" : "shared",
-			 paddr, GHCB_MSR_PSC_RESP_VAL(val)))
+		if (GHCB_MSR_PSC_RESP_VAL(val))
 			goto e_term;
 
 		if (op == SNP_PAGE_STATE_PRIVATE) {
 			/* Page validation must be performed after changing to private */
 			ret = pvalidate(vaddr, RMP_PG_SIZE_4K, true);
-			if (WARN(ret, "Failed to validate address 0x%lx ret %d", paddr, ret))
+			if (ret)
 				goto e_term;
 		}
 
@@ -760,7 +755,7 @@ void __head early_snp_set_memory_private(unsigned long vaddr, unsigned long padd
 	early_set_pages_state(vaddr, paddr, npages, SNP_PAGE_STATE_PRIVATE);
 }
 
-void __init early_snp_set_memory_shared(unsigned long vaddr, unsigned long paddr,
+void __head early_snp_set_memory_shared(unsigned long vaddr, unsigned long paddr,
 					unsigned long npages)
 {
 	/*
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
#2 lore raw

RFC PATCH 2/4 x86/xen/pvh: Move startup code into .ref.text

In reply to Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

The Xen PVH startup code does not need to live in .head.text, given that its entry point is not at a fixed offset, and is communicated to the host/VMM via an ELF note.

So move it out of .head.text into another code section. To avoid spurious warnings about references to .init code, move it into .ref.text rather than .text. (Note that the ELF note itself is not .init and so moving this code into .init.text would result in warnings as well)

---
 arch/x86/platform/pvh/head.S | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/platform/pvh/head.S +1 -1
index f7235ef87bc3..0cf6008e834b 100644
@@ -20,7 +20,7 @@
 #include <asm/nospec-branch.h>
 #include <xen/interface/elfnote.h>
 
-	__HEAD
+	__REF
 
 /*
  * Entry point for PVH guests.
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
#3 lore raw

RFC PATCH 3/4 x86/boot/64: Determine VA/PA offset before entering C code

In reply to Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

We will start using an explicit virtual-to-physical offset in the early 1:1 mapped C code to derive the kernel virtual addresses of _text and _end without having to rely on absolute symbol references, which should be avoided in such code.

Currently, phys_base is used for this purpose, which is derived from the kernel virtual address of _text, and this would lead to a circular dependency. So instead, derive virtual-to-physical offset in asm code, using the kernel VA of common_startup_64, which we already keep in a global variable for other reasons.

---
 arch/x86/include/asm/setup.h | 3 ++-
 arch/x86/kernel/head64.c     | 8 +++++---
 arch/x86/kernel/head_64.S    | 2 ++
 3 files changed, 9 insertions(+), 4 deletions(-)
arch/x86/include/asm/setup.h +2 -1
index e61e68d71cba..cc1994516af2 100644
@@ -47,7 +47,8 @@ extern unsigned long saved_video_mode;
 
 extern void reserve_standard_io_resources(void);
 extern void i386_reserve_resources(void);
-extern unsigned long __startup_64(unsigned long physaddr, struct boot_params *bp);
+extern unsigned long __startup_64(unsigned long physaddr, struct boot_params *bp,
+				  unsigned long va_offset);
 extern void startup_64_setup_gdt_idt(void);
 extern void early_setup_idt(void);
 extern void __init do_early_exception(struct pt_regs *regs, int trapnr);
arch/x86/kernel/head64.c +5 -3
index 212e8e06aeba..8fd80cf07691 100644
@@ -131,10 +131,12 @@ static unsigned long __head sme_postprocess_startup(struct boot_params *bp, pmdv
  * doesn't have to generate PC-relative relocations when accessing globals from
  * that function. Clang actually does not generate them, which leads to
  * boot-time crashes. To work around this problem, every global pointer must
- * be accessed using RIP_REL_REF().
+ * be accessed using RIP_REL_REF(). Kernel virtual addresses can be determined
+ * by subtracting va_offset from the RIP-relative address.
  */
 unsigned long __head __startup_64(unsigned long physaddr,
-				  struct boot_params *bp)
+				  struct boot_params *bp,
+				  unsigned long va_offset)
 {
 	pmd_t (*early_pgts)[PTRS_PER_PMD] = RIP_REL_REF(early_dynamic_pgts);
 	unsigned long pgtable_flags;
@@ -156,7 +158,7 @@ unsigned long __head __startup_64(unsigned long physaddr,
 	 * Compute the delta between the address I am compiled to run at
 	 * and the address I am actually running at.
 	 */
-	load_delta = physaddr - (unsigned long)(_text - __START_KERNEL_map);
+	load_delta = __START_KERNEL_map + va_offset;
 	RIP_REL_REF(phys_base) = load_delta;
 
 	/* Is the address not 2M aligned? */
arch/x86/kernel/head_64.S +2 -0
index 79f7c342e3da..3622744349d1 100644
@@ -107,6 +107,8 @@ SYM_CODE_START_NOALIGN(startup_64)
 	 */
 	leaq	_text(%rip), %rdi
 	movq	%r15, %rsi
+	leaq	common_startup_64(%rip), %rdx
+	subq	0f(%rip), %rdx
 	call	__startup_64
 
 	/* Form the CR3 value being sure to include the CR3 modifier */
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
#4 lore raw

RFC PATCH 4/4 x86/boot/64: Avoid intentional absolute symbol references in .head.text

In reply to Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

The code in .head.text executes from a 1:1 mapping and cannot generally refer to global variables using their kernel virtual addresses. However, there are some occurrences of such references that are valid: the kernel virtual addresses of _text and _end are needed to populate the page tables correctly, and some other section markers are used in a similar way.

To avoid the need for making exceptions to the rule that .head.text must not contain any absolute symbol references, derive these addresses from the RIP-relative 1:1 mapped physical addresses, which can be safely determined using RIP_REL_REF().

---
 arch/x86/kernel/head64.c | 30 ++++++++++++--------
 1 file changed, 18 insertions(+), 12 deletions(-)
arch/x86/kernel/head64.c +18 -12
index 8fd80cf07691..ce1a77e26ce3 100644
@@ -84,9 +84,11 @@ static inline bool check_la57_support(void)
 	return true;
 }
 
-static unsigned long __head sme_postprocess_startup(struct boot_params *bp, pmdval_t *pmd)
+static unsigned long __head sme_postprocess_startup(struct boot_params *bp,
+						    pmdval_t *pmd,
+						    unsigned long va_offset)
 {
-	unsigned long vaddr, vaddr_end;
+	unsigned long paddr, paddr_end;
 	int i;
 
 	/* Encrypt the kernel and related (if SME is active) */
@@ -99,10 +101,10 @@ static unsigned long __head sme_postprocess_startup(struct boot_params *bp, pmdv
 	 * attribute.
 	 */
 	if (sme_get_me_mask()) {
-		vaddr = (unsigned long)__start_bss_decrypted;
-		vaddr_end = (unsigned long)__end_bss_decrypted;
+		paddr = (unsigned long)&RIP_REL_REF(__start_bss_decrypted);
+		paddr_end = (unsigned long)&RIP_REL_REF(__end_bss_decrypted);
 
-		for (; vaddr < vaddr_end; vaddr += PMD_SIZE) {
+		for (; paddr < paddr_end; paddr += PMD_SIZE) {
 			/*
 			 * On SNP, transition the page to shared in the RMP table so that
 			 * it is consistent with the page table attribute change.
@@ -111,11 +113,11 @@ static unsigned long __head sme_postprocess_startup(struct boot_params *bp, pmdv
 			 * mapping (kernel .text). PVALIDATE, by way of
 			 * early_snp_set_memory_shared(), requires a valid virtual
 			 * address but the kernel is currently running off of the identity
-			 * mapping so use __pa() to get a *currently* valid virtual address.
+			 * mapping so use the PA to get a *currently* valid virtual address.
 			 */
-			early_snp_set_memory_shared(__pa(vaddr), __pa(vaddr), PTRS_PER_PMD);
+			early_snp_set_memory_shared(paddr, paddr, PTRS_PER_PMD);
 
-			i = pmd_index(vaddr);
+			i = pmd_index(paddr - va_offset);
 			pmd[i] -= sme_get_me_mask();
 		}
 	}
@@ -139,6 +141,7 @@ unsigned long __head __startup_64(unsigned long physaddr,
 				  unsigned long va_offset)
 {
 	pmd_t (*early_pgts)[PTRS_PER_PMD] = RIP_REL_REF(early_dynamic_pgts);
+	unsigned long va_text, va_end;
 	unsigned long pgtable_flags;
 	unsigned long load_delta;
 	pgdval_t *pgd;
@@ -165,6 +168,9 @@ unsigned long __head __startup_64(unsigned long physaddr,
 	if (load_delta & ~PMD_MASK)
 		for (;;);
 
+	va_text = physaddr - va_offset;
+	va_end  = (unsigned long)&RIP_REL_REF(_end) - va_offset;
+
 	/* Include the SME encryption mask in the fixup value */
 	load_delta += sme_get_me_mask();
 
@@ -225,7 +231,7 @@ unsigned long __head __startup_64(unsigned long physaddr,
 	pmd_entry += sme_get_me_mask();
 	pmd_entry +=  physaddr;
 
-	for (i = 0; i < DIV_ROUND_UP(_end - _text, PMD_SIZE); i++) {
+	for (i = 0; i < DIV_ROUND_UP(va_end - va_text, PMD_SIZE); i++) {
 		int idx = i + (physaddr >> PMD_SHIFT);
 
 		pmd[idx % PTRS_PER_PMD] = pmd_entry + i * PMD_SIZE;
@@ -250,11 +256,11 @@ unsigned long __head __startup_64(unsigned long physaddr,
 	pmd = &RIP_REL_REF(level2_kernel_pgt)->pmd;
 
 	/* invalidate pages before the kernel image */
-	for (i = 0; i < pmd_index((unsigned long)_text); i++)
+	for (i = 0; i < pmd_index(va_text); i++)
 		pmd[i] &= ~_PAGE_PRESENT;
 
 	/* fixup pages that are part of the kernel image */
-	for (; i <= pmd_index((unsigned long)_end); i++)
+	for (; i <= pmd_index(va_end); i++)
 		if (pmd[i] & _PAGE_PRESENT)
 			pmd[i] += load_delta;
 
@@ -262,7 +268,7 @@ unsigned long __head __startup_64(unsigned long physaddr,
 	for (; i < PTRS_PER_PMD; i++)
 		pmd[i] &= ~_PAGE_PRESENT;
 
-	return sme_postprocess_startup(bp, pmd);
+	return sme_postprocess_startup(bp, pmd, va_offset);
 }
 
 /* Wipe all early page tables except for the kernel symbol map */
Signature · 1 line
2.44.0.278.ge034bb2e1d-goog
#5 lore raw
In reply to Ard Biesheuvel

(remove bogus '[email protected]' from the To: line)

Quoted 68 lines · Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

Questions below!

This series removes the last remaining absolute symbol references from .head.text. Doing so is necessary because code in this section may be called from a 1:1 mapping of memory, which deviates from the mapping this code was linked and/or relocated to run at. This is not something that the toolchains support: even PIC/PIE code is still assumed to execute from the same mapping that it was relocated to run from by the startup code or dynamic loader. This means we are basically on our own here, and need to add measures to ensure the code works as expected in this manner. (This work was inspired by boot problems on Clang-built SEV-SNP guest kernels, where the confusion between RIP-relative and absolute references was causing variable accesses to fault)

Given that the startup code needs to create the kernel virtual mapping in the page tables, early references to some kernel virtual addresses are valid even if they cannot be dereferenced yet. To avoid having to make this distinction at build time, patches #3 and #4 replace such valid references with RIP-relative references with an offset applied.

Patches #1 and #2 remove some absolute references from .head.text that don't need to be there in the first place.

Questions:
- How can we police this at build time? Could we teach objtool to check for absolute ELF relocations in .head.text, or does this belong in modpost perhaps?

- Checking for absolute symbol references is not a complete solution, as .head.text code could call into other code as well. Do we need rigid checks for that too? Or could we have a soft rule that says you should only call __head code from __head code?

Cc Kevin Loughlin <[email protected]>
Cc Tom Lendacky <[email protected]>
Cc Dionna Glaze <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Dave Hansen <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Kees Cook <[email protected]>
Cc Brian Gerst <[email protected]>

Ard Biesheuvel (4):
x86/sev: Avoid WARN()s in early boot code x86/xen/pvh: Move startup code into .ref.text x86/boot/64: Determine VA/PA offset before entering C code x86/boot/64: Avoid intentional absolute symbol references in
.head.text

 arch/x86/include/asm/setup.h |  3 +-
 arch/x86/kernel/head64.c     | 38 ++++++++++++--------
 arch/x86/kernel/head_64.S    |  2 ++
 arch/x86/kernel/sev.c        | 15 +++-----
 arch/x86/platform/pvh/head.S |  2 +-
 5 files changed, 33 insertions(+), 27 deletions(-)

base-commit: 428080c9b19bfda37c478cd626dbd3851db1aff9 --
2.44.0.278.ge034bb2e1d-goog