Ken'ichi Ohmichi wrote:
> Hello Dave,
>
> Dave Anderson wrote:
>
>>>I think that a new virtualization_init() function is a good idea.
>>>
>>>How about checking whether "Xen" exists in PT_NOTE section of an vmcore
>>>at virtualization_init()? I feel that it is safer than reading a xen_start_info
>>>pointer.
>>>
>>>$ readelf -n vmcore
>>>Notes at offset 0x00000120 with length 0x00001388:
>>> Owner Data size Description
>>> CORE 0x00000150 NT_PRSTATUS (prstatus structure)
>>> Xen 0x00000020 Unknown note type: (0x01000002) <----- HERE
>>> CORE 0x00000150 NT_PRSTATUS (prstatus structure)
>>> Xen 0x00000020 Unknown note type: (0x01000002) <----- HERE
>>> Xen 0x00000050 Unknown note type: (0x01000001) <----- HERE
>>> VMCOREINFO_XEN 0x00000fe4 Unknown note type: (0x00000000)
>>>$
>>
>>Well there may be a couple issues with that approach. First, we need
>>a way to handle/recognize it on live systems. And second, I believe we'll need
>>a way to differentiate a pv_ops/xen kernel from a "dedicated" xen kernel.
>>And my guess is that the vmcore notes will not differentiate between the
>>two "flavors" of xen kernel (unless there's an entry in the VMCOREINFO_XEN
>>that will somehow differentiate the two).
>
>
> Thank you for the good comment. I understand your plan, and I think it is
> better than mine.
Yeah, the more I look into it, the more it appears that support for
pv_ops/xen (as well as the other pv_ops clients) are going to require
significant changes.
For now, only "traditional xen" will be supported, and when pv_ops/xen
comes online, we'll take a look at how to merge support for it.
Here's my initial update to your patch -- it simply recognizes pv_ops
kernels, and disallows the setting of the ARCH_XEN in that case.
Dave
Index: defs.h
===================================================================
RCS file: /nfs/projects/cvs/crash/defs.h,v
retrieving revision 1.349
diff -u -r1.349 defs.h
--- defs.h 18 Aug 2008 20:38:04 -0000 1.349
+++ defs.h 3 Sep 2008 17:12:46 -0000
@@ -470,11 +470,13 @@
#define RELOC_SET (0x2000000)
#define RELOC_FORCE (0x4000000)
#define ARCH_OPENVZ (0x8000000)
+#define ARCH_PVOPS (0x10000000)
#define GCC_VERSION_DEPRECATED (GCC_3_2|GCC_3_2_3|GCC_2_96|GCC_3_3_2|GCC_3_3_3)
#define XEN() (kt->flags & ARCH_XEN)
#define OPENVZ() (kt->flags & ARCH_OPENVZ)
+#define PVOPS() (kt->flags & ARCH_PVOPS)
#define XEN_MACHINE_TO_MFN(m) ((ulonglong)(m) >> PAGESHIFT())
#define XEN_PFN_TO_PSEUDO(p) ((ulonglong)(p) << PAGESHIFT())
@@ -2124,6 +2126,8 @@
#define VMEMMAP_VADDR_2_6_24 0xffffe20000000000
#define VMEMMAP_END_2_6_24 0xffffe2ffffffffff
+#define PAGE_OFFSET_2_6_27 0xffff880000000000
+
#define USERSPACE_TOP_XEN 0x0000800000000000
#define PAGE_OFFSET_XEN 0xffff880000000000
#define VMALLOC_START_ADDR_XEN 0xffffc20000000000
@@ -2221,6 +2225,17 @@
#define PAGEBASE(X) (((ulong)(X)) & (ulong)machdep->pagemask)
+#define _CPU_PDA_READ2(CPU, BUFFER) \
+ ((readmem(symbol_value("_cpu_pda"), \
+ KVADDR, &cpu_pda_addr, sizeof(unsigned long), \
+ "_cpu_pda addr", FAULT_ON_ERROR)) && \
+ (readmem(cpu_pda_addr + ((CPU) * sizeof(void *)), \
+ KVADDR, &cpu_pda_addr, sizeof(unsigned long), \
+ "_cpu_pda addr", FAULT_ON_ERROR)) && \
+ (cpu_pda_addr) && \
+ (readmem(cpu_pda_addr, KVADDR, (BUFFER), SIZE(x8664_pda), \
+ "cpu_pda entry", FAULT_ON_ERROR)))
+
#define _CPU_PDA_READ(CPU, BUFFER) \
((STRNEQ("_cpu_pda", closest_symbol((symbol_value("_cpu_pda") + \
((CPU) * sizeof(unsigned long)))))) && \
@@ -3329,6 +3344,7 @@
long SIZE_option(long, long, char *, char *, int, char *, char *);
void dump_trace(ulong *);
int enumerator_value(char *, long *);
+void paravirt_init(void);
/*
* memory.c
Index: kernel.c
===================================================================
RCS file: /nfs/projects/cvs/crash/kernel.c,v
retrieving revision 1.186
diff -u -r1.186 kernel.c
--- kernel.c 6 Aug 2008 20:53:33 -0000 1.186
+++ kernel.c 3 Sep 2008 17:32:09 -0000
@@ -79,7 +79,8 @@
kt->end = symbol_value("_end");
/*
- * For the Xen architecture, default to writable page tables unless:
+ * For the traditional (non-pv_ops) Xen architecture, default to writable
+ * page tables unless:
*
* (1) it's an "xm save" CANONICAL_PAGE_TABLES dumpfile, or
* (2) the --shadow_page_tables option was explicitly entered.
@@ -88,7 +89,7 @@
* it's not an "xm save" canonical dumpfile, then we have no choice
* but to presume shadow page tables.
*/
- if (symbol_exists("xen_start_info")) {
+ if (!PVOPS() && symbol_exists("xen_start_info")) {
kt->flags |= ARCH_XEN;
if (!(kt->xen_flags & (SHADOW_PAGE_TABLES|CANONICAL_PAGE_TABLES)))
kt->xen_flags |= WRITABLE_PAGE_TABLES;
@@ -3887,6 +3888,8 @@
fprintf(fp, "%sARCH_XEN", others++ ? "|" : "");
if (kt->flags & ARCH_OPENVZ)
fprintf(fp, "%sARCH_OPENVZ", others++ ? "|" : "");
+ if (kt->flags & ARCH_PVOPS)
+ fprintf(fp, "%sARCH_PVOPS", others++ ? "|" : "");
if (kt->flags & NO_IKCONFIG)
fprintf(fp, "%sNO_IKCONFIG", others++ ? "|" : "");
if (kt->flags & DWARF_UNWIND)
@@ -6290,3 +6293,24 @@
break;
}
}
+
+/*
+ * With the evidence available, attempt to pre-determine whether
+ * this is a paravirt-capable kernel running as bare-metal, xen,
+ * kvm, etc.
+ *
+ * NOTE: Only bare-metal pv_ops kernels are supported so far.
+ */
+void
+paravirt_init(void)
+{
+ /*
+ * pv_init_ops appears to be (as of 2.6.27) an arch-common
+ * symbol. This may have to change.
+ */
+ if (kernel_symbol_exists("pv_init_ops")) {
+ if (CRASHDEBUG(1))
+ error(INFO, "pv_init_ops exists: ARCH_PVOPS\n");
+ kt->flags |= ARCH_PVOPS;
+ }
+}
Index: main.c
===================================================================
RCS file: /nfs/projects/cvs/crash/main.c,v
retrieving revision 1.87
diff -u -r1.87 main.c
--- main.c 18 Aug 2008 20:38:05 -0000 1.87
+++ main.c 3 Sep 2008 17:04:59 -0000
@@ -467,6 +467,7 @@
hq_init();
machdep_init(PRE_SYMTAB);
symtab_init();
+ paravirt_init();
machdep_init(PRE_GDB);
datatype_init();
Index: x86_64.c
===================================================================
RCS file: /nfs/projects/cvs/crash/x86_64.c,v
retrieving revision 1.102
diff -u -r1.102 x86_64.c
--- x86_64.c 23 Apr 2008 19:54:00 -0000 1.102
+++ x86_64.c 3 Sep 2008 17:19:26 -0000
@@ -147,7 +147,9 @@
case PRE_GDB:
if (!(machdep->flags & VM_FLAGS)) {
if (symbol_exists("xen_start_info")) {
- if (symbol_exists("low_pml4") &&
+ if (PVOPS())
+ machdep->flags |= VM_2_6_11;
+ else if (symbol_exists("low_pml4") &&
symbol_exists("swap_low_mappings"))
machdep->flags |= VM_XEN_RHEL4;
else
@@ -178,7 +180,6 @@
case VM_2_6_11:
/* 2.6.11 layout */
machdep->machspec->userspace_top = USERSPACE_TOP_2_6_11;
- machdep->machspec->page_offset = PAGE_OFFSET_2_6_11;
machdep->machspec->vmalloc_start_addr = VMALLOC_START_ADDR_2_6_11;
machdep->machspec->vmalloc_end = VMALLOC_END_2_6_11;
machdep->machspec->modules_vaddr = MODULES_VADDR_2_6_11;
@@ -190,6 +191,13 @@
if (symbol_exists("vmemmap_populate"))
machdep->flags |= VMEMMAP;
+ if (kernel_symbol_exists("end_pfn"))
+ /* 2.6.11 layout */
+ machdep->machspec->page_offset = PAGE_OFFSET_2_6_11;
+ else
+ /* 2.6.27 layout */
+ machdep->machspec->page_offset = PAGE_OFFSET_2_6_27;
+
machdep->uvtop = x86_64_uvtop_level4;
break;
@@ -534,7 +542,7 @@
static void
x86_64_cpu_pda_init(void)
{
- int i, cpus, nr_pda, cpunumber, _cpu_pda;
+ int i, cpus, nr_pda, cpunumber, _cpu_pda, _boot_cpu_pda;
char *cpu_pda_buf;
ulong level4_pgt, data_offset, cpu_pda_addr;
struct syment *sp, *nsp;
@@ -569,11 +577,21 @@
_cpu_pda = FALSE;
}
}
-
+ if (_cpu_pda) {
+ if (symbol_exists("_boot_cpu_pda"))
+ _boot_cpu_pda = TRUE;
+ else
+ _boot_cpu_pda = FALSE;
+ }
for (i = cpus = 0; i < nr_pda; i++) {
if (_cpu_pda) {
- if (!_CPU_PDA_READ(i, cpu_pda_buf))
- break;
+ if (_boot_cpu_pda) {
+ if (!_CPU_PDA_READ2(i, cpu_pda_buf))
+ break;
+ } else {
+ if (!_CPU_PDA_READ(i, cpu_pda_buf))
+ break;
+ }
} else {
if (!CPU_PDA_READ(i, cpu_pda_buf))
break;
@@ -3914,7 +3932,7 @@
int
x86_64_get_smp_cpus(void)
{
- int i, cpus, nr_pda, cpunumber, _cpu_pda;
+ int i, cpus, nr_pda, cpunumber, _cpu_pda, _boot_cpu_pda;
char *cpu_pda_buf;
ulong level4_pgt, cpu_pda_addr;
@@ -3940,10 +3958,21 @@
_cpu_pda = FALSE;
}
}
+ if (_cpu_pda) {
+ if (symbol_exists("_boot_cpu_pda"))
+ _boot_cpu_pda = TRUE;
+ else
+ _boot_cpu_pda = FALSE;
+ }
for (i = cpus = 0; i < nr_pda; i++) {
if (_cpu_pda) {
- if (!_CPU_PDA_READ(i, cpu_pda_buf))
- break;
+ if (_boot_cpu_pda) {
+ if (!_CPU_PDA_READ2(i, cpu_pda_buf))
+ break;
+ } else {
+ if (!_CPU_PDA_READ(i, cpu_pda_buf))
+ break;
+ }
} else {
if (!CPU_PDA_READ(i, cpu_pda_buf))
break;
--
Crash-utility mailing list
Crash-utility@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/crash-utility
