Skip to contentSkip to conversation

makedumpfile needs to remove the memory encryption

PATCH In review Lianbo Jiang 4 participants 13 messages Jan 22, 2019 – Jan 28, 2019

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover makedumpfile needs to remove the memory encryption

The patchset did two things:
[1] add a new variable 'sme_mask' to number_table

The variable will be used to store the sme mask for crashed kernel, the sme_mask denotes whether the old memory is encrypted or not.

[2] remove the memory encryption mask to obtain the true physical address

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

References:

x86/kdump: Export the SME mask to vmcoreinfo
https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?id=65f750e5457aef9a8085a99d613fea0430303e93
https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?id=f263245a0ce2c4e23b89a58fa5f7dfc048e11929

Lianbo Jiang (2):
Makedumpfile: add a new variable 'sme_mask' to number_table Remove the memory encryption mask to obtain the true physical address

 arch/x86_64.c  | 3 +++
 makedumpfile.c | 4 ++++
 makedumpfile.h | 1 +
 3 files changed, 8 insertions(+)
Signature · 7 lines
2.17.1


_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
#1 lore raw

PATCH 1/2 Makedumpfile: add a new variable 'sme_mask' to number_table

It will be used to store the sme mask for crashed kernel, the sme_mask denotes whether the old memory is encrypted or not.

---
 makedumpfile.c | 3 +++
 makedumpfile.h | 1 +
 2 files changed, 4 insertions(+)
makedumpfile.c +3 -0
index 8923538..a03aaa1 100644
@@ -1743,6 +1743,7 @@ get_structure_info(void)
 	ENUM_NUMBER_INIT(NR_FREE_PAGES, "NR_FREE_PAGES");
 	ENUM_NUMBER_INIT(N_ONLINE, "N_ONLINE");
 	ENUM_NUMBER_INIT(pgtable_l5_enabled, "pgtable_l5_enabled");
+	ENUM_NUMBER_INIT(sme_mask, "sme_mask");
 
 	ENUM_NUMBER_INIT(PG_lru, "PG_lru");
 	ENUM_NUMBER_INIT(PG_private, "PG_private");
@@ -2276,6 +2277,7 @@ write_vmcoreinfo_data(void)
 	WRITE_NUMBER("NR_FREE_PAGES", NR_FREE_PAGES);
 	WRITE_NUMBER("N_ONLINE", N_ONLINE);
 	WRITE_NUMBER("pgtable_l5_enabled", pgtable_l5_enabled);
+	WRITE_NUMBER("sme_mask", sme_mask);
 
 	WRITE_NUMBER("PG_lru", PG_lru);
 	WRITE_NUMBER("PG_private", PG_private);
@@ -2672,6 +2674,7 @@ read_vmcoreinfo(void)
 	READ_NUMBER("NR_FREE_PAGES", NR_FREE_PAGES);
 	READ_NUMBER("N_ONLINE", N_ONLINE);
 	READ_NUMBER("pgtable_l5_enabled", pgtable_l5_enabled);
+	READ_NUMBER("sme_mask", sme_mask);
 
 	READ_NUMBER("PG_lru", PG_lru);
 	READ_NUMBER("PG_private", PG_private);
makedumpfile.h +1 -0
index 73813ed..e97b2e7 100644
@@ -1912,6 +1912,7 @@ struct number_table {
 	long	NR_FREE_PAGES;
 	long	N_ONLINE;
 	long	pgtable_l5_enabled;
+	long	sme_mask;
 
 	/*
  	* Page flags
Signature · 7 lines
2.17.1


_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
#2 lore raw

RE: [PATCH 1/2] Makedumpfile: add a new variable 'sme_mask' to number_table

Hi Lianbo,

Quoted 14 lines · Lianbo Jiang

It will be used to store the sme mask for crashed kernel, the sme_mask denotes whether the old memory is encrypted or not.

---
 makedumpfile.c | 3 +++
 makedumpfile.h | 1 +
 2 files changed, 4 insertions(+)
makedumpfile.c +0 -0
index 8923538..a03aaa1 100644

makedumpfile.c get_structure_info(void) · line 1746

Quoted 4 lines
makedumpfile.c +1 -0
@@ -1743,6 +1743,7 @@ get_structure_info(void)
 	ENUM_NUMBER_INIT(NR_FREE_PAGES, "NR_FREE_PAGES");
 	ENUM_NUMBER_INIT(N_ONLINE, "N_ONLINE");
 	ENUM_NUMBER_INIT(pgtable_l5_enabled, "pgtable_l5_enabled");
+	ENUM_NUMBER_INIT(sme_mask, "sme_mask");

This is useless because the sme_mask is not enum number. Please remove it.

And, dividing this patchset into the two patches doesn't make sense to me in this case. Could you merge them into a patch?

Thanks,
Kazu

Quoted 32 lines
makedumpfile.c +2 -0
 	ENUM_NUMBER_INIT(PG_lru, "PG_lru");
 	ENUM_NUMBER_INIT(PG_private, "PG_private");
@@ -2276,6 +2277,7 @@ write_vmcoreinfo_data(void)
 	WRITE_NUMBER("NR_FREE_PAGES", NR_FREE_PAGES);
 	WRITE_NUMBER("N_ONLINE", N_ONLINE);
 	WRITE_NUMBER("pgtable_l5_enabled", pgtable_l5_enabled);
+	WRITE_NUMBER("sme_mask", sme_mask);
 
 	WRITE_NUMBER("PG_lru", PG_lru);
 	WRITE_NUMBER("PG_private", PG_private);
@@ -2672,6 +2674,7 @@ read_vmcoreinfo(void)
 	READ_NUMBER("NR_FREE_PAGES", NR_FREE_PAGES);
 	READ_NUMBER("N_ONLINE", N_ONLINE);
 	READ_NUMBER("pgtable_l5_enabled", pgtable_l5_enabled);
+	READ_NUMBER("sme_mask", sme_mask);
 
 	READ_NUMBER("PG_lru", PG_lru);
 	READ_NUMBER("PG_private", PG_private);
makedumpfile.h +1 -1
index 73813ed..e97b2e7 100644
@@ -1912,6 +1912,7 @@ struct number_table {
 	long	NR_FREE_PAGES;
 	long	N_ONLINE;
 	long	pgtable_l5_enabled;
+	long	sme_mask;
 
 	/*
  	* Page flags
--
Quoted 2 lines

2.17.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#3 lore raw

Re: [PATCH 1/2] Makedumpfile: add a new variable 'sme_mask' to number_table

Quoted 26 lines · Kazuhito Hagio

Hi Lianbo,

Quoted 14 lines · Lianbo Jiang

It will be used to store the sme mask for crashed kernel, the sme_mask denotes whether the old memory is encrypted or not.

---
 makedumpfile.c | 3 +++
 makedumpfile.h | 1 +
 2 files changed, 4 insertions(+)
makedumpfile.c +0 -0
index 8923538..a03aaa1 100644

makedumpfile.c get_structure_info(void) · line 1746

Quoted 4 lines
makedumpfile.c +1 -0
@@ -1743,6 +1743,7 @@ get_structure_info(void)
 	ENUM_NUMBER_INIT(NR_FREE_PAGES, "NR_FREE_PAGES");
 	ENUM_NUMBER_INIT(N_ONLINE, "N_ONLINE");
 	ENUM_NUMBER_INIT(pgtable_l5_enabled, "pgtable_l5_enabled");
+	ENUM_NUMBER_INIT(sme_mask, "sme_mask");

This is useless because the sme_mask is not enum number. Please remove it.
Ok. Thanks for your comment.

Quoted 3 lines

And, dividing this patchset into the two patches doesn't make sense to me in this case. Could you merge them into a patch?

Sure, they will be merged into a patch in next post.

Regards,
Lianbo

Quoted 32 lines

Thanks,
Kazu

Quoted 29 lines
makedumpfile.c +2 -0
 	ENUM_NUMBER_INIT(PG_lru, "PG_lru");
 	ENUM_NUMBER_INIT(PG_private, "PG_private");
@@ -2276,6 +2277,7 @@ write_vmcoreinfo_data(void)
 	WRITE_NUMBER("NR_FREE_PAGES", NR_FREE_PAGES);
 	WRITE_NUMBER("N_ONLINE", N_ONLINE);
 	WRITE_NUMBER("pgtable_l5_enabled", pgtable_l5_enabled);
+	WRITE_NUMBER("sme_mask", sme_mask);
 
 	WRITE_NUMBER("PG_lru", PG_lru);
 	WRITE_NUMBER("PG_private", PG_private);
@@ -2672,6 +2674,7 @@ read_vmcoreinfo(void)
 	READ_NUMBER("NR_FREE_PAGES", NR_FREE_PAGES);
 	READ_NUMBER("N_ONLINE", N_ONLINE);
 	READ_NUMBER("pgtable_l5_enabled", pgtable_l5_enabled);
+	READ_NUMBER("sme_mask", sme_mask);
 
 	READ_NUMBER("PG_lru", PG_lru);
 	READ_NUMBER("PG_private", PG_private);
makedumpfile.h +1 -0
index 73813ed..e97b2e7 100644
@@ -1912,6 +1912,7 @@ struct number_table {
 	long	NR_FREE_PAGES;
 	long	N_ONLINE;
 	long	pgtable_l5_enabled;
+	long	sme_mask;
 
Quoted 7 lines
Quoted 5 lines
makedumpfile.h +0 -1
 	/*
  	* Page flags
--

2.17.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#4 lore raw

PATCH 2/2 Remove the memory encryption mask to obtain the true physical address

In reply to Lianbo Jiang

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

---
 arch/x86_64.c  | 3 +++
 makedumpfile.c | 1 +
 2 files changed, 4 insertions(+)
arch/x86_64.c +3 -0
index 537fb78..7651d36 100644
@@ -346,6 +346,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 			return NOT_PADDR;
 		}
 		pud_paddr  = pgd & ENTRY_MASK;
+		pud_paddr = pud_paddr & ~(NUMBER(sme_mask));
 	}
 
 	/*
@@ -371,6 +372,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PMD.
 	 */
 	pmd_paddr  = pud_pte & ENTRY_MASK;
+	pmd_paddr = pmd_paddr & ~(NUMBER(sme_mask));
 	pmd_paddr += pmd_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pmd_paddr, &pmd_pte, sizeof pmd_pte)) {
 		ERRMSG("Can't get pmd_pte (pmd_paddr:%lx).\n", pmd_paddr);
@@ -391,6 +393,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PTE.
 	 */
 	pte_paddr  = pmd_pte & ENTRY_MASK;
+	pte_paddr = pte_paddr & ~(NUMBER(sme_mask));
 	pte_paddr += pte_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pte_paddr, &pte, sizeof pte)) {
 		ERRMSG("Can't get pte (pte_paddr:%lx).\n", pte_paddr);
makedumpfile.c +1 -0
index a03aaa1..81c7bb4 100644
@@ -977,6 +977,7 @@ next_page:
 	read_size = MIN(info->page_size - PAGEOFFSET(paddr), size);
 
 	pgaddr = PAGEBASE(paddr);
+	pgaddr = pgaddr & ~(NUMBER(sme_mask));
 	pgbuf = cache_search(pgaddr, read_size);
 	if (!pgbuf) {
 		++cache_miss;
Signature · 7 lines
2.17.1


_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
#5 lore raw

RE: [PATCH 2/2] Remove the memory encryption mask to obtain the true physical address

Quoted 44 lines · Lianbo Jiang

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

---
 arch/x86_64.c  | 3 +++
 makedumpfile.c | 1 +
 2 files changed, 4 insertions(+)
arch/x86_64.c +3 -0
index 537fb78..7651d36 100644
@@ -346,6 +346,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 			return NOT_PADDR;
 		}
 		pud_paddr  = pgd & ENTRY_MASK;
+		pud_paddr = pud_paddr & ~(NUMBER(sme_mask));
 	}
 
 	/*
@@ -371,6 +372,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PMD.
 	 */
 	pmd_paddr  = pud_pte & ENTRY_MASK;
+	pmd_paddr = pmd_paddr & ~(NUMBER(sme_mask));
 	pmd_paddr += pmd_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pmd_paddr, &pmd_pte, sizeof pmd_pte)) {
 		ERRMSG("Can't get pmd_pte (pmd_paddr:%lx).\n", pmd_paddr);
@@ -391,6 +393,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PTE.
 	 */
 	pte_paddr  = pmd_pte & ENTRY_MASK;
+	pte_paddr = pte_paddr & ~(NUMBER(sme_mask));
 	pte_paddr += pte_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pte_paddr, &pte, sizeof pte)) {
 		ERRMSG("Can't get pte (pte_paddr:%lx).\n", pte_paddr);
makedumpfile.c +0 -0
index a03aaa1..81c7bb4 100644

makedumpfile.c next_page: · line 980

Quoted 4 lines
makedumpfile.c +1 -0
@@ -977,6 +977,7 @@ next_page:
 	read_size = MIN(info->page_size - PAGEOFFSET(paddr), size);
 
 	pgaddr = PAGEBASE(paddr);
+	pgaddr = pgaddr & ~(NUMBER(sme_mask));

Since NUMBER(sme_mask) is initialized with -1 (NOT_FOUND_NUMBER), if the sme_mask is not in vmcoreinfo, ~(NUMBER(sme_mask)) will be 0. So the four lines added above need

if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)

    ...

and, what I'm wondering is whether it doesn't need to take hugepages into account such as this

392 if (pmd_pte & _PAGE_PSE) /* 2MB pages */
393 return (pmd_pte & ENTRY_MASK & PMD_MASK) +
394 (vaddr & ~PMD_MASK);
"arch/x86_64.c"

Thanks,
Kazu

Quoted 6 lines
makedumpfile.c +0 -1
 	pgbuf = cache_search(pgaddr, read_size);
 	if (!pgbuf) {
 		++cache_miss;
--

2.17.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#6 lore raw

RE: [PATCH 2/2] Remove the memory encryption mask to obtain the true physical address

Quoted 54 lines · Kazuhito Hagio
Quoted 44 lines · Lianbo Jiang

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

---
 arch/x86_64.c  | 3 +++
 makedumpfile.c | 1 +
 2 files changed, 4 insertions(+)
arch/x86_64.c +3 -0
index 537fb78..7651d36 100644
@@ -346,6 +346,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 			return NOT_PADDR;
 		}
 		pud_paddr  = pgd & ENTRY_MASK;
+		pud_paddr = pud_paddr & ~(NUMBER(sme_mask));
 	}
 
 	/*
@@ -371,6 +372,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PMD.
 	 */
 	pmd_paddr  = pud_pte & ENTRY_MASK;
+	pmd_paddr = pmd_paddr & ~(NUMBER(sme_mask));
 	pmd_paddr += pmd_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pmd_paddr, &pmd_pte, sizeof pmd_pte)) {
 		ERRMSG("Can't get pmd_pte (pmd_paddr:%lx).\n", pmd_paddr);
@@ -391,6 +393,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PTE.
 	 */
 	pte_paddr  = pmd_pte & ENTRY_MASK;
+	pte_paddr = pte_paddr & ~(NUMBER(sme_mask));
 	pte_paddr += pte_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pte_paddr, &pte, sizeof pte)) {
 		ERRMSG("Can't get pte (pte_paddr:%lx).\n", pte_paddr);
makedumpfile.c +0 -0
index a03aaa1..81c7bb4 100644

makedumpfile.c next_page: · line 980

Quoted 4 lines
makedumpfile.c +1 -0
@@ -977,6 +977,7 @@ next_page:
 	read_size = MIN(info->page_size - PAGEOFFSET(paddr), size);
 
 	pgaddr = PAGEBASE(paddr);
+	pgaddr = pgaddr & ~(NUMBER(sme_mask));

Since NUMBER(sme_mask) is initialized with -1 (NOT_FOUND_NUMBER), if the sme_mask is not in vmcoreinfo, ~(NUMBER(sme_mask)) will be 0. So the four lines added above need

Quoted 2 lines

if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)

    ...

Considering hugepage and the code, it might be better to add a local variable for the mask value to __vtop4_x86_64() function and mask it without condition, for example

  unsigned long sme_mask = ~0UL;

  if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)
      sme_mask = ~(NUMBER(sme_mask));
  ...
  pud_paddr = pgd & ENTRY_MASK & sme_mask;

to avoid adding lots of 'if' statements.

Thanks,
Kazu

Quoted 22 lines

and, what I'm wondering is whether it doesn't need to take hugepages into account such as this

392 if (pmd_pte & _PAGE_PSE) /* 2MB pages */
393 return (pmd_pte & ENTRY_MASK & PMD_MASK) +
394 (vaddr & ~PMD_MASK);
"arch/x86_64.c"

Thanks,
Kazu

Quoted 6 lines
makedumpfile.c +0 -1
 	pgbuf = cache_search(pgaddr, read_size);
 	if (!pgbuf) {
 		++cache_miss;
--

2.17.1

Quoted 4 lines

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#7 lore raw

Re: [PATCH 2/2] Remove the memory encryption mask to obtain the true physical address

Quoted 62 lines · Kazuhito Hagio
Quoted 54 lines · Kazuhito Hagio
Quoted 44 lines · Lianbo Jiang

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

---
 arch/x86_64.c  | 3 +++
 makedumpfile.c | 1 +
 2 files changed, 4 insertions(+)
arch/x86_64.c +3 -0
index 537fb78..7651d36 100644
@@ -346,6 +346,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 			return NOT_PADDR;
 		}
 		pud_paddr  = pgd & ENTRY_MASK;
+		pud_paddr = pud_paddr & ~(NUMBER(sme_mask));
 	}
 
 	/*
@@ -371,6 +372,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PMD.
 	 */
 	pmd_paddr  = pud_pte & ENTRY_MASK;
+	pmd_paddr = pmd_paddr & ~(NUMBER(sme_mask));
 	pmd_paddr += pmd_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pmd_paddr, &pmd_pte, sizeof pmd_pte)) {
 		ERRMSG("Can't get pmd_pte (pmd_paddr:%lx).\n", pmd_paddr);
@@ -391,6 +393,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PTE.
 	 */
 	pte_paddr  = pmd_pte & ENTRY_MASK;
+	pte_paddr = pte_paddr & ~(NUMBER(sme_mask));
 	pte_paddr += pte_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pte_paddr, &pte, sizeof pte)) {
 		ERRMSG("Can't get pte (pte_paddr:%lx).\n", pte_paddr);
makedumpfile.c +0 -0
index a03aaa1..81c7bb4 100644

makedumpfile.c next_page: · line 980

Quoted 4 lines
makedumpfile.c +1 -0
@@ -977,6 +977,7 @@ next_page:
 	read_size = MIN(info->page_size - PAGEOFFSET(paddr), size);
 
 	pgaddr = PAGEBASE(paddr);
+	pgaddr = pgaddr & ~(NUMBER(sme_mask));

Since NUMBER(sme_mask) is initialized with -1 (NOT_FOUND_NUMBER), if the sme_mask is not in vmcoreinfo, ~(NUMBER(sme_mask)) will be 0. So the four lines added above need

Quoted 2 lines

if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)

    ...

Considering hugepage and the code, it might be better to add a local variable for the mask value to __vtop4_x86_64() function and mask it without condition, for example

Quoted 9 lines
  unsigned long sme_mask = ~0UL;

  if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)
      sme_mask = ~(NUMBER(sme_mask));
  ...
  pud_paddr = pgd & ENTRY_MASK & sme_mask;

to avoid adding lots of 'if' statements.

Good idea. Thank you, Kazu.

Quoted 31 lines

Thanks,
Kazu

Quoted 26 lines

and, what I'm wondering is whether it doesn't need to take hugepages into account such as this

392 if (pmd_pte & _PAGE_PSE) /* 2MB pages */
393 return (pmd_pte & ENTRY_MASK & PMD_MASK) +
394 (vaddr & ~PMD_MASK);
"arch/x86_64.c"

Thanks,
Kazu

Quoted 6 lines
makedumpfile.c +0 -1
 	pgbuf = cache_search(pgaddr, read_size);
 	if (!pgbuf) {
 		++cache_miss;
--

2.17.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#8 lore raw

Re: [PATCH 2/2] Remove the memory encryption mask to obtain the true physical address

In reply to Kazuhito Hagio
Quoted 54 lines · Kazuhito Hagio
Quoted 44 lines · Lianbo Jiang

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

---
 arch/x86_64.c  | 3 +++
 makedumpfile.c | 1 +
 2 files changed, 4 insertions(+)
arch/x86_64.c +3 -0
index 537fb78..7651d36 100644
@@ -346,6 +346,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 			return NOT_PADDR;
 		}
 		pud_paddr  = pgd & ENTRY_MASK;
+		pud_paddr = pud_paddr & ~(NUMBER(sme_mask));
 	}
 
 	/*
@@ -371,6 +372,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PMD.
 	 */
 	pmd_paddr  = pud_pte & ENTRY_MASK;
+	pmd_paddr = pmd_paddr & ~(NUMBER(sme_mask));
 	pmd_paddr += pmd_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pmd_paddr, &pmd_pte, sizeof pmd_pte)) {
 		ERRMSG("Can't get pmd_pte (pmd_paddr:%lx).\n", pmd_paddr);
@@ -391,6 +393,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PTE.
 	 */
 	pte_paddr  = pmd_pte & ENTRY_MASK;
+	pte_paddr = pte_paddr & ~(NUMBER(sme_mask));
 	pte_paddr += pte_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pte_paddr, &pte, sizeof pte)) {
 		ERRMSG("Can't get pte (pte_paddr:%lx).\n", pte_paddr);
makedumpfile.c +0 -0
index a03aaa1..81c7bb4 100644

makedumpfile.c next_page: · line 980

Quoted 4 lines
makedumpfile.c +1 -0
@@ -977,6 +977,7 @@ next_page:
 	read_size = MIN(info->page_size - PAGEOFFSET(paddr), size);
 
 	pgaddr = PAGEBASE(paddr);
+	pgaddr = pgaddr & ~(NUMBER(sme_mask));

Since NUMBER(sme_mask) is initialized with -1 (NOT_FOUND_NUMBER), if the sme_mask is not in vmcoreinfo, ~(NUMBER(sme_mask)) will be 0. So the four lines added above need

Quoted 3 lines

if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)

    ...

Thank you very much for pointing out my mistake.

I will improve it and post again.

Quoted 8 lines

and, what I'm wondering is whether it doesn't need to take hugepages into account such as this

392 if (pmd_pte & _PAGE_PSE) /* 2MB pages */
393 return (pmd_pte & ENTRY_MASK & PMD_MASK) +
394 (vaddr & ~PMD_MASK);
"arch/x86_64.c"

This is a good question. Theoretically, it should be modified accordingly for huge pages case.

But makedumpfile still works well without this change. And i'm sure that the huge pages are enabled in crashed kernel. This is very strange.

Thanks.
Lianbo

Quoted 12 lines

Thanks,
Kazu

Quoted 6 lines
makedumpfile.c +0 -1
 	pgbuf = cache_search(pgaddr, read_size);
 	if (!pgbuf) {
 		++cache_miss;
--

2.17.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#9 lore raw

Re: [PATCH 2/2] Remove the memory encryption mask to obtain the true physical address

+ Tom

Quoted 72 lines · lijiang
Quoted 54 lines · Kazuhito Hagio
Quoted 44 lines · Lianbo Jiang

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

---
 arch/x86_64.c  | 3 +++
 makedumpfile.c | 1 +
 2 files changed, 4 insertions(+)
arch/x86_64.c +3 -0
index 537fb78..7651d36 100644
@@ -346,6 +346,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 			return NOT_PADDR;
 		}
 		pud_paddr  = pgd & ENTRY_MASK;
+		pud_paddr = pud_paddr & ~(NUMBER(sme_mask));
 	}
 
 	/*
@@ -371,6 +372,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PMD.
 	 */
 	pmd_paddr  = pud_pte & ENTRY_MASK;
+	pmd_paddr = pmd_paddr & ~(NUMBER(sme_mask));
 	pmd_paddr += pmd_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pmd_paddr, &pmd_pte, sizeof pmd_pte)) {
 		ERRMSG("Can't get pmd_pte (pmd_paddr:%lx).\n", pmd_paddr);
@@ -391,6 +393,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PTE.
 	 */
 	pte_paddr  = pmd_pte & ENTRY_MASK;
+	pte_paddr = pte_paddr & ~(NUMBER(sme_mask));
 	pte_paddr += pte_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pte_paddr, &pte, sizeof pte)) {
 		ERRMSG("Can't get pte (pte_paddr:%lx).\n", pte_paddr);
makedumpfile.c +0 -0
index a03aaa1..81c7bb4 100644

makedumpfile.c next_page: · line 980

Quoted 4 lines
makedumpfile.c +1 -0
@@ -977,6 +977,7 @@ next_page:
 	read_size = MIN(info->page_size - PAGEOFFSET(paddr), size);
 
 	pgaddr = PAGEBASE(paddr);
+	pgaddr = pgaddr & ~(NUMBER(sme_mask));

Since NUMBER(sme_mask) is initialized with -1 (NOT_FOUND_NUMBER), if the sme_mask is not in vmcoreinfo, ~(NUMBER(sme_mask)) will be 0. So the four lines added above need

Quoted 3 lines

if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)

    ...

Thank you very much for pointing out my mistake.

I will improve it and post again.

Quoted 8 lines

and, what I'm wondering is whether it doesn't need to take hugepages into account such as this

392 if (pmd_pte & _PAGE_PSE) /* 2MB pages */
393 return (pmd_pte & ENTRY_MASK & PMD_MASK) +
394 (vaddr & ~PMD_MASK);
"arch/x86_64.c"

Quoted 21 lines

This is a good question. Theoretically, it should be modified accordingly for huge pages case.

But makedumpfile still works well without this change. And i'm sure that the huge pages are enabled in crashed kernel. This is very strange.

Thanks.
Lianbo

Quoted 12 lines

Thanks,
Kazu

Quoted 6 lines
makedumpfile.c +0 -1
 	pgbuf = cache_search(pgaddr, read_size);
 	if (!pgbuf) {
 		++cache_miss;
--

2.17.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#10 lore raw

Re: [PATCH 2/2] Remove the memory encryption mask to obtain the true physical address

Quoted 64 lines · On 1/24/19 9:55 PM, [email protected] wrote:

+ Tom

Quoted 62 lines · lijiang
Quoted 54 lines · Kazuhito Hagio
Quoted 44 lines · Lianbo Jiang

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

---
 arch/x86_64.c  | 3 +++
 makedumpfile.c | 1 +
 2 files changed, 4 insertions(+)
arch/x86_64.c +3 -0
index 537fb78..7651d36 100644
@@ -346,6 +346,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 			return NOT_PADDR;
 		}
 		pud_paddr  = pgd & ENTRY_MASK;
+		pud_paddr = pud_paddr & ~(NUMBER(sme_mask));
 	}
 
 	/*
@@ -371,6 +372,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PMD.
 	 */
 	pmd_paddr  = pud_pte & ENTRY_MASK;
+	pmd_paddr = pmd_paddr & ~(NUMBER(sme_mask));
 	pmd_paddr += pmd_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pmd_paddr, &pmd_pte, sizeof pmd_pte)) {
 		ERRMSG("Can't get pmd_pte (pmd_paddr:%lx).\n", pmd_paddr);
@@ -391,6 +393,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PTE.
 	 */
 	pte_paddr  = pmd_pte & ENTRY_MASK;
+	pte_paddr = pte_paddr & ~(NUMBER(sme_mask));
 	pte_paddr += pte_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pte_paddr, &pte, sizeof pte)) {
 		ERRMSG("Can't get pte (pte_paddr:%lx).\n", pte_paddr);
makedumpfile.c +0 -0
index a03aaa1..81c7bb4 100644

makedumpfile.c next_page: · line 980

Quoted 4 lines
makedumpfile.c +1 -0
@@ -977,6 +977,7 @@ next_page:
 	read_size = MIN(info->page_size - PAGEOFFSET(paddr), size);
 
 	pgaddr = PAGEBASE(paddr);
+	pgaddr = pgaddr & ~(NUMBER(sme_mask));

Since NUMBER(sme_mask) is initialized with -1 (NOT_FOUND_NUMBER), if the sme_mask is not in vmcoreinfo, ~(NUMBER(sme_mask)) will be 0. So the four lines added above need

Quoted 3 lines

if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)

    ...

Thank you very much for pointing out my mistake.

I will improve it and post again.

Might be worth creating a local variable that includes ENTRY_MASK and NUMBER(sme_mask) so that you make the check just once. Then use that variable in place of ENTRY_MASK in the remainder of the function so that the correct value is used throughout.

This would also cover the 5-level path which would make this future proof should AMD someday support 5-level paging.

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

and, what I'm wondering is whether it doesn't need to take hugepages into account such as this

Quoted 8 lines
Quoted 8 lines
Quoted 5 lines

392 if (pmd_pte & _PAGE_PSE) /* 2MB pages */
393 return (pmd_pte & ENTRY_MASK & PMD_MASK) +
394 (vaddr & ~PMD_MASK);
"arch/x86_64.c"

This is a good question. Theoretically, it should be modified accordingly for huge pages case.

Yes, this should also have the ~(NUMBER(sme_mask)) applied to it. You can probably add some debugging to see if you're hitting this case and whether the encryption bit (sme_mask) is set just to help understand what is occurring. This also goes for the 1GB page check above. However, if you use my suggestion of a local variable then you should be covered.

Thanks,
Tom

Quoted 11 lines
Quoted 11 lines

But makedumpfile still works well without this change. And i'm sure that the huge pages are enabled in crashed kernel. This is very strange.

Thanks.
Lianbo

Quoted 4 lines

Thanks,
Kazu

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
makedumpfile.c +0 -1
 	pgbuf = cache_search(pgaddr, read_size);
 	if (!pgbuf) {
 		++cache_miss;
--

2.17.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#11 lore raw

Re: [PATCH 2/2] Remove the memory encryption mask to obtain the true physical address

Quoted 66 lines · Thomas
Quoted 64 lines · On 1/24/19 9:55 PM, [email protected] wrote:

+ Tom

Quoted 62 lines · lijiang
Quoted 54 lines · Kazuhito Hagio
Quoted 44 lines · Lianbo Jiang

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

---
 arch/x86_64.c  | 3 +++
 makedumpfile.c | 1 +
 2 files changed, 4 insertions(+)
arch/x86_64.c +3 -0
index 537fb78..7651d36 100644
@@ -346,6 +346,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 			return NOT_PADDR;
 		}
 		pud_paddr  = pgd & ENTRY_MASK;
+		pud_paddr = pud_paddr & ~(NUMBER(sme_mask));
 	}
 
 	/*
@@ -371,6 +372,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PMD.
 	 */
 	pmd_paddr  = pud_pte & ENTRY_MASK;
+	pmd_paddr = pmd_paddr & ~(NUMBER(sme_mask));
 	pmd_paddr += pmd_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pmd_paddr, &pmd_pte, sizeof pmd_pte)) {
 		ERRMSG("Can't get pmd_pte (pmd_paddr:%lx).\n", pmd_paddr);
@@ -391,6 +393,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PTE.
 	 */
 	pte_paddr  = pmd_pte & ENTRY_MASK;
+	pte_paddr = pte_paddr & ~(NUMBER(sme_mask));
 	pte_paddr += pte_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pte_paddr, &pte, sizeof pte)) {
 		ERRMSG("Can't get pte (pte_paddr:%lx).\n", pte_paddr);
makedumpfile.c +0 -0
index a03aaa1..81c7bb4 100644

makedumpfile.c next_page: · line 980

Quoted 4 lines
makedumpfile.c +1 -0
@@ -977,6 +977,7 @@ next_page:
 	read_size = MIN(info->page_size - PAGEOFFSET(paddr), size);
 
 	pgaddr = PAGEBASE(paddr);
+	pgaddr = pgaddr & ~(NUMBER(sme_mask));

Since NUMBER(sme_mask) is initialized with -1 (NOT_FOUND_NUMBER), if the sme_mask is not in vmcoreinfo, ~(NUMBER(sme_mask)) will be 0. So the four lines added above need

Quoted 3 lines

if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)

    ...

Thank you very much for pointing out my mistake.

I will improve it and post again.

Quoted 8 lines

Might be worth creating a local variable that includes ENTRY_MASK and NUMBER(sme_mask) so that you make the check just once. Then use that variable in place of ENTRY_MASK in the remainder of the function so that the correct value is used throughout.

This would also cover the 5-level path which would make this future proof should AMD someday support 5-level paging.

Thank you, Tom. Makedumpfile will cover the 5-level path in next post, though AMD does not support 5-level paging yet.

Thanks.
Lianbo

Quoted 13 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

and, what I'm wondering is whether it doesn't need to take hugepages into account such as this

Quoted 8 lines
Quoted 8 lines
Quoted 5 lines

392 if (pmd_pte & _PAGE_PSE) /* 2MB pages */
393 return (pmd_pte & ENTRY_MASK & PMD_MASK) +
394 (vaddr & ~PMD_MASK);
"arch/x86_64.c"

This is a good question. Theoretically, it should be modified accordingly for huge pages case.

Quoted 28 lines

Yes, this should also have the ~(NUMBER(sme_mask)) applied to it. You can probably add some debugging to see if you're hitting this case and whether the encryption bit (sme_mask) is set just to help understand what is occurring. This also goes for the 1GB page check above. However, if you use my suggestion of a local variable then you should be covered.

Thanks,
Tom

Quoted 19 lines
Quoted 19 lines

But makedumpfile still works well without this change. And i'm sure that the huge pages are enabled in crashed kernel. This is very strange.

Thanks.
Lianbo

Quoted 12 lines

Thanks,
Kazu

Quoted 6 lines
makedumpfile.c +0 -1
 	pgbuf = cache_search(pgaddr, read_size);
 	if (!pgbuf) {
 		++cache_miss;
--

2.17.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#12 lore raw

Re: [PATCH 2/2] Remove the memory encryption mask to obtain the true physical address

Quoted 65 lines · lijiang
Quoted 64 lines · Thomas
Quoted 63 lines · On 1/24/19 9:55 PM, [email protected] wrote:

+ Tom

Quoted 61 lines · lijiang
Quoted 54 lines · Kazuhito Hagio
Quoted 44 lines · Lianbo Jiang

For AMD machine with SME feature, if SME is enabled in the first kernel, the crashed kernel's page table(pgd/pud/pmd/pte) contains the memory encryption mask, so makedumpfile needs to remove the memory encryption mask to obtain the true physical address.

---
 arch/x86_64.c  | 3 +++
 makedumpfile.c | 1 +
 2 files changed, 4 insertions(+)
arch/x86_64.c +3 -0
index 537fb78..7651d36 100644
@@ -346,6 +346,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 			return NOT_PADDR;
 		}
 		pud_paddr  = pgd & ENTRY_MASK;
+		pud_paddr = pud_paddr & ~(NUMBER(sme_mask));
 	}
 
 	/*
@@ -371,6 +372,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PMD.
 	 */
 	pmd_paddr  = pud_pte & ENTRY_MASK;
+	pmd_paddr = pmd_paddr & ~(NUMBER(sme_mask));
 	pmd_paddr += pmd_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pmd_paddr, &pmd_pte, sizeof pmd_pte)) {
 		ERRMSG("Can't get pmd_pte (pmd_paddr:%lx).\n", pmd_paddr);
@@ -391,6 +393,7 @@ __vtop4_x86_64(unsigned long vaddr, unsigned long pagetable)
 	 * Get PTE.
 	 */
 	pte_paddr  = pmd_pte & ENTRY_MASK;
+	pte_paddr = pte_paddr & ~(NUMBER(sme_mask));
 	pte_paddr += pte_index(vaddr) * sizeof(unsigned long);
 	if (!readmem(PADDR, pte_paddr, &pte, sizeof pte)) {
 		ERRMSG("Can't get pte (pte_paddr:%lx).\n", pte_paddr);
makedumpfile.c +0 -0
index a03aaa1..81c7bb4 100644

makedumpfile.c next_page: · line 980

Quoted 4 lines
makedumpfile.c +1 -0
@@ -977,6 +977,7 @@ next_page:
 	read_size = MIN(info->page_size - PAGEOFFSET(paddr), size);
 
 	pgaddr = PAGEBASE(paddr);
+	pgaddr = pgaddr & ~(NUMBER(sme_mask));

Since NUMBER(sme_mask) is initialized with -1 (NOT_FOUND_NUMBER), if the sme_mask is not in vmcoreinfo, ~(NUMBER(sme_mask)) will be 0. So the four lines added above need

Quoted 3 lines

if (NUMBER(sme_mask) != NOT_FOUND_NUMBER)

    ...

Thank you very much for pointing out my mistake.

Quoted 7 lines
Quoted 7 lines
Quoted 1 line
Quoted 1 line

I will improve it and post again.

Might be worth creating a local variable that includes ENTRY_MASK and NUMBER(sme_mask) so that you make the check just once. Then use that variable in place of ENTRY_MASK in the remainder of the function so that the correct value is used throughout.

Ok.

Quoted 7 lines
Quoted 3 lines

This would also cover the 5-level path which would make this future proof should AMD someday support 5-level paging.

Thank you, Tom. Makedumpfile will cover the 5-level path in next post, though AMD does not support 5-level paging yet.

I mean that i will improve this patch and cover the 5-level path in patch v2.

Thanks.

Quoted 22 lines

Thanks.
Lianbo

Quoted 19 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

and, what I'm wondering is whether it doesn't need to take hugepages into account such as this

Quoted 8 lines
Quoted 8 lines
Quoted 5 lines

392 if (pmd_pte & _PAGE_PSE) /* 2MB pages */
393 return (pmd_pte & ENTRY_MASK & PMD_MASK) +
394 (vaddr & ~PMD_MASK);
"arch/x86_64.c"

This is a good question. Theoretically, it should be modified accordingly for huge pages case.

Yes, this should also have the ~(NUMBER(sme_mask)) applied to it. You can probably add some debugging to see if you're hitting this case and whether the encryption bit (sme_mask) is set just to help understand what is occurring. This also goes for the 1GB page check above. However, if you use my suggestion of a local variable then you should be covered.

Thank you, Tom.

I will modify this patch and cover the huge pages case in patch v2.

Thanks.
Lianbo

Quoted 14 lines
Quoted 14 lines

Thanks,
Tom

Quoted 11 lines
Quoted 11 lines

But makedumpfile still works well without this change. And i'm sure that the huge pages are enabled in crashed kernel. This is very strange.

Thanks.
Lianbo

Quoted 4 lines

Thanks,
Kazu

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
makedumpfile.c +0 -1
 	pgbuf = cache_search(pgaddr, read_size);
 	if (!pgbuf) {
 		++cache_miss;
--

2.17.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec