2:17 a.m.
There are few comments which are not relevant anymore. Update the
comments to reflect the current situation.
Signed-off-by: Mika Westerberg <ext-mika.1.westerberg(a)nokia.com>
---
arm.c | 3 ---
unwind_arm.c | 5 ++---
2 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/arm.c b/arm.c
index 4e6ab3d..1752f62 100644
--- a/arm.c
+++ b/arm.c
@@ -1048,9 +1048,6 @@ arm_get_frame(struct bt_info *bt, ulong *pcp, ulong *spp)
/*
* Get the starting point for the active cpu in a diskdump.
- *
- * Note that we currently support only UP machines. In future we might want to
- * support SMP machines as well.
*/
static int
arm_get_dumpfile_stack_frame(struct bt_info *bt, ulong *nip, ulong *ksp)
diff --git a/unwind_arm.c b/unwind_arm.c
index 58fb5da..18fdad9 100644
--- a/unwind_arm.c
+++ b/unwind_arm.c
@@ -250,9 +250,8 @@ read_module_unwind_table(struct unwind_table *tbl, ulong addr)
idx_size = idx_stop - idx_start;
/*
- * We know the size of the index table. Allocate memory for the table
- * (including the EH table) and read the contents from the kernel
- * memory.
+ * We know the size of the index table. Allocate memory for
+ * the table and read the contents from the kernel memory.
*/
tbl->idx = calloc(idx_size, 1);
if (!tbl->idx)
--
1.7.3.2
2:17 a.m.
New subject: [PATCH 2/2] ARM: add support for the new page table format introduced in 2.6.38
Starting from 2.6.38 ARM hardware and Linux page tables switched place. There
were also some smaller changes to the PTE bits. This patch makes crash to cope
with those changes.
Signed-off-by: Mika Westerberg <ext-mika.1.westerberg(a)nokia.com>
---
Dave,
Are you interested in having a reference vmlinux/vmcore pair for .38 ARM
kernels? I can try to arrange such if needed and provide download url in
a private email.
Regards,
MW
arm.c | 63 ++++++++++++++++++++++++++++++++++++++++++++++++++++++---------
defs.h | 1 +
2 files changed, 55 insertions(+), 9 deletions(-)
diff --git a/arm.c b/arm.c
index 1752f62..0347166 100644
--- a/arm.c
+++ b/arm.c
@@ -6,7 +6,7 @@
* Jan Karlsson <jan.karlsson(a)sonyericsson.com>
* Mika Westerberg <ext-mika.1.westerberg(a)nokia.com>
*
- * Copyright (C) 2010 Nokia Corporation
+ * Copyright (C) 2010-2011 Nokia Corporation
* Copyright (C) 2010 Sony Ericsson. All rights reserved.
*
* This program is free software; you can redistribute it and/or modify
@@ -88,8 +88,12 @@ pmd_page_addr(ulong pmd)
{
ulong ptr;
- ptr = pmd & ~(PTRS_PER_PTE * sizeof(void *) - 1);
- ptr += PTRS_PER_PTE * sizeof(void *);
+ if (machdep->flags & PGTABLE_V2) {
+ ptr = PAGEBASE(pmd);
+ } else {
+ ptr = pmd & ~(PTRS_PER_PTE * sizeof(void *) - 1);
+ ptr += PTRS_PER_PTE * sizeof(void *);
+ }
return (ulong *)ptr;
}
@@ -102,16 +106,21 @@ pmd_page_addr(ulong pmd)
#define L_PTE_FILE (1 << 2)
#define L_PTE_DIRTY (1 << 6)
#define L_PTE_WRITE (1 << 7)
+#define L_PTE_RDONLY L_PTE_WRITE
#define L_PTE_USER (1 << 8)
#define L_PTE_EXEC (1 << 9)
+#define L_PTE_XN L_PTE_EXEC
#define L_PTE_SHARED (1 << 10)
#define pte_val(pte) (pte)
#define pte_present(pte) (pte_val(pte) & L_PTE_PRESENT)
#define pte_write(pte) (pte_val(pte) & L_PTE_WRITE)
+#define pte_rdonly(pte) (pte_val(pte) & L_PTE_RDONLY)
#define pte_dirty(pte) (pte_val(pte) & L_PTE_DIRTY)
#define pte_young(pte) (pte_val(pte) & L_PTE_YOUNG)
+#define pte_exec(pte) (pte_val(pte) & L_PTE_EXEC)
+#define pte_xn(pte) (pte_val(pte) & L_PTE_XN)
/*
* Following stuff is taken directly from the kernel sources. These are used in
@@ -241,6 +250,16 @@ arm_init(int when)
break;
case POST_GDB:
+ /*
+ * Starting from 2.6.38 hardware and Linux page tables
+ * were reordered. See also mainline kernel commit
+ * d30e45eeabe (ARM: pgtable: switch order of Linux vs
+ * hardware page tables).
+ */
+ if (THIS_KERNEL_VERSION > LINUX(2,6,37) ||
+ STRUCT_EXISTS("pteval_t"))
+ machdep->flags |= PGTABLE_V2;
+
if (symbol_exists("irq_desc"))
ARRAY_LENGTH_INIT(machdep->nr_irqs, irq_desc,
"irq_desc", NULL, 0);
@@ -313,6 +332,8 @@ arm_dump_machdep_table(ulong arg)
fprintf(fp, "%sKSYMS_START", others++ ? "|" : "");
if (machdep->flags & PHYS_BASE)
fprintf(fp, "%sPHYS_BASE", others++ ? "|" : "");
+ if (machdep->flags & PGTABLE_V2)
+ fprintf(fp, "%sPGTABLE_V2", others++ ? "|" : "");
fprintf(fp, ")\n");
fprintf(fp, " kvbase: %lx\n", machdep->kvbase);
@@ -823,12 +844,21 @@ arm_translate_pte(ulong pte, void *physaddr, ulonglong pae_pte)
if (pte) {
if (pte_present(pte))
fprintf(fp, "%sPRESENT", others++ ? "|" : "");
- if (pte_write(pte))
- fprintf(fp, "%sWRITE", others++ ? "|" : "");
if (pte_dirty(pte))
fprintf(fp, "%sDIRTY", others++ ? "|" : "");
if (pte_young(pte))
fprintf(fp, "%sYOUNG", others++ ? "|" : "");
+ if (machdep->flags & PGTABLE_V2) {
+ if (!pte_rdonly(pte))
+ fprintf(fp, "%sWRITE", others++ ? "|" : "");
+ if (!pte_xn(pte))
+ fprintf(fp, "%sEXEC", others++ ? "|" : "");
+ } else {
+ if (pte_write(pte))
+ fprintf(fp, "%sWRITE", others++ ? "|" : "");
+ if (pte_exec(pte))
+ fprintf(fp, "%sEXEC", others++ ? "|" : "");
+ }
} else {
fprintf(fp, "no mapping");
}
@@ -863,6 +893,8 @@ arm_vtop(ulong vaddr, ulong *pgd, physaddr_t *paddr, int verbose)
* that in ARM Linux we have following setup (see also
* arch/arm/include/asm/pgtable.h)
*
+ * Before 2.6.38
+ *
* PGD PTE
* +---------+
* | | 0 ----> +------------+
@@ -875,15 +907,28 @@ arm_vtop(ulong vaddr, ulong *pgd, physaddr_t *paddr, int verbose)
* | | 4095 | Linux pt 1 |
* +---------+ +------------+ +4096
*
+ * Starting from 2.6.38
+ *
+ * PGD PTE
+ * +---------+
+ * | | 0 ----> +------------+
+ * +- - - - -+ | Linux pt 0 |
+ * | | 4 ----> +------------+ +1024
+ * +- - - - -+ | Linux pt 1 |
+ * . . +------------+ +2048
+ * . . | h/w pt 0 |
+ * . . +------------+ +3072
+ * | | 4095 | h/w pt 1 |
+ * +---------+ +------------+ +4096
+ *
* So in Linux implementation we have two hardware pointers to second
- * level page tables. After these come "Linux" versions of the page
- * tables.
+ * level page tables. Depending on the kernel version, the "Linux" page
+ * tables either follow or precede the hardware tables.
*
* Linux PT entries contain bits that are not supported on hardware, for
* example "young" and "dirty" flags.
*
- * Our translation scheme only uses Linux PTEs here. Hardware entries
- * are 1024 bytes below Linux versions.
+ * Our translation scheme only uses Linux PTEs here.
*/
if (verbose)
diff --git a/defs.h b/defs.h
index 22f876b..e1b7d01 100755
--- a/defs.h
+++ b/defs.h
@@ -4097,6 +4097,7 @@ struct arm_pt_regs {
#define KSYMS_START (0x1)
#define PHYS_BASE (0x2)
+#define PGTABLE_V2 (0x4)
struct machine_specific {
ulong phys_base;
--
1.7.3.2
9:19 a.m.
New subject: [PATCH 2/2] ARM: add support for the new page table format introduced in 2.6.38
----- Original Message -----
Starting from 2.6.38 ARM hardware and Linux page tables switched
place. There
were also some smaller changes to the PTE bits. This patch makes crash to cope
with those changes.
Signed-off-by: Mika Westerberg <ext-mika.1.westerberg(a)nokia.com>
---
Queued for the next release.
Dave,
Are you interested in having a reference vmlinux/vmcore pair for .38 ARM
kernels? I can try to arrange such if needed and provide download url in
a private email.
Yes pease, another small one would be great.
Thanks,
Dave
Regards,
MW
arm.c | 63
++++++++++++++++++++++++++++++++++++++++++++++++++++++---------
defs.h | 1 +
2 files changed, 55 insertions(+), 9 deletions(-)
diff --git a/arm.c b/arm.c
index 1752f62..0347166 100644
--- a/arm.c
+++ b/arm.c
@@ -6,7 +6,7 @@
* Jan Karlsson <jan.karlsson(a)sonyericsson.com>
* Mika Westerberg <ext-mika.1.westerberg(a)nokia.com>
*
- * Copyright (C) 2010 Nokia Corporation
+ * Copyright (C) 2010-2011 Nokia Corporation
* Copyright (C) 2010 Sony Ericsson. All rights reserved.
*
* This program is free software; you can redistribute it and/or modify
@@ -88,8 +88,12 @@ pmd_page_addr(ulong pmd)
{
ulong ptr;
- ptr = pmd & ~(PTRS_PER_PTE * sizeof(void *) - 1);
- ptr += PTRS_PER_PTE * sizeof(void *);
+ if (machdep->flags & PGTABLE_V2) {
+ ptr = PAGEBASE(pmd);
+ } else {
+ ptr = pmd & ~(PTRS_PER_PTE * sizeof(void *) - 1);
+ ptr += PTRS_PER_PTE * sizeof(void *);
+ }
return (ulong *)ptr;
}
@@ -102,16 +106,21 @@ pmd_page_addr(ulong pmd)
#define L_PTE_FILE (1 << 2)
#define L_PTE_DIRTY (1 << 6)
#define L_PTE_WRITE (1 << 7)
+#define L_PTE_RDONLY L_PTE_WRITE
#define L_PTE_USER (1 << 8)
#define L_PTE_EXEC (1 << 9)
+#define L_PTE_XN L_PTE_EXEC
#define L_PTE_SHARED (1 << 10)
#define pte_val(pte) (pte)
#define pte_present(pte) (pte_val(pte) & L_PTE_PRESENT)
#define pte_write(pte) (pte_val(pte) & L_PTE_WRITE)
+#define pte_rdonly(pte) (pte_val(pte) & L_PTE_RDONLY)
#define pte_dirty(pte) (pte_val(pte) & L_PTE_DIRTY)
#define pte_young(pte) (pte_val(pte) & L_PTE_YOUNG)
+#define pte_exec(pte) (pte_val(pte) & L_PTE_EXEC)
+#define pte_xn(pte) (pte_val(pte) & L_PTE_XN)
/*
* Following stuff is taken directly from the kernel sources. These are
used in
@@ -241,6 +250,16 @@ arm_init(int when)
break;
case POST_GDB:
+ /*
+ * Starting from 2.6.38 hardware and Linux page tables
+ * were reordered. See also mainline kernel commit
+ * d30e45eeabe (ARM: pgtable: switch order of Linux vs
+ * hardware page tables).
+ */
+ if (THIS_KERNEL_VERSION > LINUX(2,6,37) ||
+ STRUCT_EXISTS("pteval_t"))
+ machdep->flags |= PGTABLE_V2;
+
if (symbol_exists("irq_desc"))
ARRAY_LENGTH_INIT(machdep->nr_irqs, irq_desc,
"irq_desc", NULL, 0);
@@ -313,6 +332,8 @@ arm_dump_machdep_table(ulong arg)
fprintf(fp, "%sKSYMS_START", others++ ? "|" : "");
if (machdep->flags & PHYS_BASE)
fprintf(fp, "%sPHYS_BASE", others++ ? "|" : "");
+ if (machdep->flags & PGTABLE_V2)
+ fprintf(fp, "%sPGTABLE_V2", others++ ? "|" : "");
fprintf(fp, ")\n");
fprintf(fp, " kvbase: %lx\n", machdep->kvbase);
@@ -823,12 +844,21 @@ arm_translate_pte(ulong pte, void *physaddr,
ulonglong pae_pte)
if (pte) {
if (pte_present(pte))
fprintf(fp, "%sPRESENT", others++ ? "|" : "");
- if (pte_write(pte))
- fprintf(fp, "%sWRITE", others++ ? "|" : "");
if (pte_dirty(pte))
fprintf(fp, "%sDIRTY", others++ ? "|" : "");
if (pte_young(pte))
fprintf(fp, "%sYOUNG", others++ ? "|" : "");
+ if (machdep->flags & PGTABLE_V2) {
+ if (!pte_rdonly(pte))
+ fprintf(fp, "%sWRITE", others++ ? "|" : "");
+ if (!pte_xn(pte))
+ fprintf(fp, "%sEXEC", others++ ? "|" : "");
+ } else {
+ if (pte_write(pte))
+ fprintf(fp, "%sWRITE", others++ ? "|" : "");
+ if (pte_exec(pte))
+ fprintf(fp, "%sEXEC", others++ ? "|" : "");
+ }
} else {
fprintf(fp, "no mapping");
}
@@ -863,6 +893,8 @@ arm_vtop(ulong vaddr, ulong *pgd, physaddr_t
*paddr, int verbose)
* that in ARM Linux we have following setup (see also
* arch/arm/include/asm/pgtable.h)
*
+ * Before 2.6.38
+ *
* PGD PTE
* +---------+
* | | 0 ----> +------------+
@@ -875,15 +907,28 @@ arm_vtop(ulong vaddr, ulong *pgd, physaddr_t
*paddr, int verbose)
* | | 4095 | Linux pt 1 |
* +---------+ +------------+ +4096
*
+ * Starting from 2.6.38
+ *
+ * PGD PTE
+ * +---------+
+ * | | 0 ----> +------------+
+ * +- - - - -+ | Linux pt 0 |
+ * | | 4 ----> +------------+ +1024
+ * +- - - - -+ | Linux pt 1 |
+ * . . +------------+ +2048
+ * . . | h/w pt 0 |
+ * . . +------------+ +3072
+ * | | 4095 | h/w pt 1 |
+ * +---------+ +------------+ +4096
+ *
* So in Linux implementation we have two hardware pointers to second
- * level page tables. After these come "Linux" versions of the page
- * tables.
+ * level page tables. Depending on the kernel version, the "Linux"
page
+ * tables either follow or precede the hardware tables.
*
* Linux PT entries contain bits that are not supported on hardware,
for
* example "young" and "dirty" flags.
*
- * Our translation scheme only uses Linux PTEs here. Hardware entries
- * are 1024 bytes below Linux versions.
+ * Our translation scheme only uses Linux PTEs here.
*/
if (verbose)
diff --git a/defs.h b/defs.h
index 22f876b..e1b7d01 100755
--- a/defs.h
+++ b/defs.h
@@ -4097,6 +4097,7 @@ struct arm_pt_regs {
#define KSYMS_START (0x1)
#define PHYS_BASE (0x2)
+#define PGTABLE_V2 (0x4)
struct machine_specific {
ulong phys_base;
--
1.7.3.2
--
Crash-utility mailing list
Crash-utility(a)redhat.com
https://www.redhat.com/mailman/listinfo/crash-utility
9:15 a.m.
----- Original Message -----
There are few comments which are not relevant anymore. Update the
comments to reflect the current situation.
Signed-off-by: Mika Westerberg <ext-mika.1.westerberg(a)nokia.com>
Now there's a nice safe patch -- queued for the next release.
Thanks,
Dave
---
arm.c | 3 ---
unwind_arm.c | 5 ++---
2 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/arm.c b/arm.c
index 4e6ab3d..1752f62 100644
--- a/arm.c
+++ b/arm.c
@@ -1048,9 +1048,6 @@ arm_get_frame(struct bt_info *bt, ulong *pcp,
ulong *spp)
/*
* Get the starting point for the active cpu in a diskdump.
- *
- * Note that we currently support only UP machines. In future we might want to
- * support SMP machines as well.
*/
static int
arm_get_dumpfile_stack_frame(struct bt_info *bt, ulong *nip, ulong *ksp)
diff --git a/unwind_arm.c b/unwind_arm.c
index 58fb5da..18fdad9 100644
--- a/unwind_arm.c
+++ b/unwind_arm.c
@@ -250,9 +250,8 @@ read_module_unwind_table(struct unwind_table *tbl,
ulong addr)
idx_size = idx_stop - idx_start;
/*
- * We know the size of the index table. Allocate memory for the table
- * (including the EH table) and read the contents from the kernel
- * memory.
+ * We know the size of the index table. Allocate memory for
+ * the table and read the contents from the kernel memory.
*/
tbl->idx = calloc(idx_size, 1);
if (!tbl->idx)
--
1.7.3.2
--
Crash-utility mailing list
Crash-utility(a)redhat.com
https://www.redhat.com/mailman/listinfo/crash-utility
5083
days inactive
5083
days old
devel@lists.crash-utility.osci.io
3 comments
2 participants
participants (2)
-
Dave Anderson -
Mika Westerberg