2:18 p.m.
I have proposed a patch to makedumpfile to (optionally) exclude from
a dump the page structures representing excluded pages.
The idea being that a system with terabytes of system memory has
millions of pages of page structures. And most of them are unneeded.
That patch thread begins here:
http://marc.info/?l=kexec&m=138853299130125&w=2
Dave [Anderson] raised these crash-related issues;
Although I'm sure you tested this, I find it amazing that
only the "kmem -[fF]" option is the only command option
that is affected?
If I'm not mistaken, this would be the first time that legitimate
kernel data would be excluded from the dump, and the user would
have no obvious way of knowing that it had been done, correct?
If it were encoded in the in the header somewhere, at least a
warning message could be printed during crash initialization.
...
Right, but look at all of the other page struct offsets in addition to
page.lru that are used. The page.flags usage comes to mind, and for
example, what would "kmem -p" display for the missing pages?
Or "kmem <address>"? And would "kmem -i" display invalid
data?
I'm just speculating off the top of my head, but the page structure is
such a fundamental data structure with several of its fields being used,
just enter "help -o page_" to see all of its potential member usages.
So I am submitting two patches for your consideration, should the patch
to exclude unused vmemmap pages be taken into makedumpfile.
- [PATCH 1/2] crash: initial note of excluded page structures
This one makes crash startup look like this:
This program has absolutely no warranty. Enter "help warranty" for
details.
NOTE: Unused vmemmap page structures are excluded from this dump.
GNU gdb (GDB) 7.6
Copyright (C) 2013 Free Software Foundation, Inc.
- [PATCH 1/2] crash: kmem warnings for excluded page structures
This patch modifies kmem options -f, -F, -s addr, -S addr, and -i.
Those are the only options that I could detect looking for excluded pages.
This patch applies on top of the first, and adds some warnings to the
output of these kmem options. For example:
crash> kmem -f
Note: kmem -f may fail because unused page structures are excluded from this dump.
NODE
0
ZONE NAME SIZE FREE MEM_MAP START_PADDR START_MAPNR
0 DMA 4095 3934 ffffea0000000038 1000 0
AREA SIZE FREE_AREA_STRUCT BLOCKS PAGES
0 4k ffff880000013068 2 2
...
crash> kmem -i
PAGES TOTAL PERCENTAGE
TOTAL MEM 128008147 488.3 GB ----
FREE 127599276 486.8 GB 99% of TOTAL MEM
USED 408871 1.6 GB 0% of TOTAL MEM
SHARED 11049 43.2 MB 0% of TOTAL MEM
BUFFERS 5722 22.4 MB 0% of TOTAL MEM
CACHED 44638 174.4 MB 0% of TOTAL MEM
SLAB 62139 242.7 MB 0% of TOTAL MEM
TOTAL SWAP 4893032 18.7 GB ----
SWAP USED 0 0 0% of TOTAL SWAP
SWAP FREE 4893032 18.7 GB 100% of TOTAL SWAP
Note: 1970727 free pages not found (excluded); results are incomplete.
Unused page structures are excluded from this dump.
-Cliff Wickman
cpw(a)sgi.com
2:23 p.m.
New subject: [PATCH 1/2] crash: initial note of excluded page structures
If makedumpfile excludes unused page structures it will flag that
fact in the dump header.
(There are about 3.67 million pages full of page structures for
every tera byte of system memory. The great bulk of those
page structures are not needed.)
Their exclusion is a makedumpfile option.
Crash will display a note during initialization if such structures
have been excluded. Crash commands that walk page freelists, for
example, will fail. So the note will help the user understand why.
Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
---
diskdump.c | 4 ++++
diskdump.h | 1 +
2 files changed, 5 insertions(+)
Index: crash-7.0.4/diskdump.c
===================================================================
--- crash-7.0.4.orig/diskdump.c
+++ crash-7.0.4/diskdump.c
@@ -749,6 +749,10 @@ restart:
dd->valid_pages[i]++;
}
+ if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
+ fprintf(fp,
+ "NOTE: Unused vmemmap page structures are excluded from this dump.\n");
+
return TRUE;
err:
Index: crash-7.0.4/diskdump.h
===================================================================
--- crash-7.0.4.orig/diskdump.h
+++ crash-7.0.4/diskdump.h
@@ -84,6 +84,7 @@ struct kdump_sub_header {
#define DUMP_DH_COMPRESSED_ZLIB 0x1 /* page is compressed with zlib */
#define DUMP_DH_COMPRESSED_LZO 0x2 /* page is compressed with lzo */
#define DUMP_DH_COMPRESSED_SNAPPY 0x4 /* page is compressed with snappy */
+#define DUMP_DH_EXCLUDED_VMEMMAP 0x8 /* unused vmemmap pages are excluded */
/* descriptor of each page for vmcore */
typedef struct page_desc {
2:27 p.m.
New subject: [PATCH 2/2] crash: kmem warnings for excluded page structures
This patch applies on top of patch "initial note of excluded page structures".
A time and diskspace option of makedumpfile is to exclude unused
vemmap page structures.
Certain crash commands will fail if they search free page
lists.
They fail with some hint of the problem:
9 2048k ffff8800000133b0 kmem: page excluded: kernel virtual address:
ffffea0000007028 type: "first list entry"
This patch further explains to the user why certain kmem options fail:
kmem -f
kmem -F
kmem -s|-S addr
Or may give incomplete results:
kmem -i
Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
---
diskdump.c | 5 ++++-
memory.c | 33 ++++++++++++++++++++++++++++++---
2 files changed, 34 insertions(+), 4 deletions(-)
Index: crash-7.0.4/memory.c
===================================================================
--- crash-7.0.4.orig/memory.c
+++ crash-7.0.4/memory.c
@@ -20,6 +20,8 @@
#include <sys/mman.h>
#include <ctype.h>
#include <netinet/in.h>
+extern int vmemmmap_excluded;
+long not_found_excluded = 0;
struct meminfo { /* general purpose memory information structure */
ulong cache; /* used by the various memory searching/dumping */
@@ -2160,6 +2162,7 @@ readmem(ulonglong addr, int memtype, voi
goto readmem_error;
case PAGE_EXCLUDED:
+ not_found_excluded++;
RETURN_ON_PARTIAL_READ();
if (PRINT_ERROR_MESSAGE)
error(INFO, PAGE_EXCLUDED_ERRMSG, memtype_string(memtype, 0),
addr, type);
@@ -4370,7 +4373,9 @@ cmd_kmem(void)
meminfo.spec_addr = value[i];
meminfo.flags = ADDRESS_SPECIFIED;
if (meminfo.calls++)
- fprintf(fp, "\n");
+ fprintf(fp, "\nNote: kmem -f may fail because
"
+ "unused page structures are "
+ "excluded from this dump.\n");
vt->dump_free_pages(&meminfo);
fflag++;
}
@@ -4417,6 +4422,10 @@ cmd_kmem(void)
meminfo.flags = ADDRESS_SPECIFIED;
if (Sflag && (vt->flags & KMALLOC_SLUB))
meminfo.flags |= VERBOSE;
+ if (vmemmmap_excluded)
+ fprintf(fp, "Note: kmem -s|-S may fail because "
+ "unused page structures are "
+ "excluded from this dump.\n");
if (meminfo.calls++)
fprintf(fp, "\n");
vt->dump_kmem_cache(&meminfo);
@@ -4483,17 +4492,35 @@ cmd_kmem(void)
}
- if (iflag == 1)
+ if (iflag == 1) {
+ not_found_excluded = 0;
dump_kmeminfo();
+ if (not_found_excluded) {
+ fprintf(fp, "\nNote: %ld free pages not found (excluded); "
+ "results are incomplete.\n", not_found_excluded);
+ if (vmemmmap_excluded)
+ fprintf(fp, "Unused page structures are "
+ "excluded from this dump.\n");
+ }
+ }
if (pflag == 1)
dump_mem_map(&meminfo);
- if (fflag == 1)
+ if (fflag == 1) {
+ if (vmemmmap_excluded)
+ fprintf(fp, "Note: kmem -f may fail because "
+ "unused page structures are "
+ "excluded from this dump.\n");
vt->dump_free_pages(&meminfo);
+ }
if (Fflag == 1) {
meminfo.flags = VERBOSE;
+ if (vmemmmap_excluded)
+ fprintf(fp, "Note: kmem -F may fail because "
+ "unused page structures are "
+ "excluded from this dump.\n");
vt->dump_free_pages(&meminfo);
}
Index: crash-7.0.4/diskdump.c
===================================================================
--- crash-7.0.4.orig/diskdump.c
+++ crash-7.0.4/diskdump.c
@@ -25,6 +25,7 @@
#include "defs.h"
#include "diskdump.h"
+int vmemmmap_excluded = 0;
#define BITMAP_SECT_LEN 4096
@@ -749,9 +750,11 @@ restart:
dd->valid_pages[i]++;
}
- if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
+ if (header->status & DUMP_DH_EXCLUDED_VMEMMAP) {
fprintf(fp,
"NOTE: Unused vmemmap page structures are excluded from this dump.\n");
+ vmemmmap_excluded = 1;
+ }
return TRUE;
2:32 p.m.
----- Original Message -----
I have proposed a patch to makedumpfile to (optionally) exclude from
a dump the page structures representing excluded pages.
The idea being that a system with terabytes of system memory has
millions of pages of page structures. And most of them are unneeded.
That patch thread begins here:
http://marc.info/?l=kexec&m=138853299130125&w=2
Dave [Anderson] raised these crash-related issues;
Although I'm sure you tested this, I find it amazing that
only the "kmem -[fF]" option is the only command option
that is affected?
If I'm not mistaken, this would be the first time that legitimate
kernel data would be excluded from the dump, and the user would
have no obvious way of knowing that it had been done, correct?
If it were encoded in the in the header somewhere, at least a
warning message could be printed during crash initialization.
...
Right, but look at all of the other page struct offsets in addition to
page.lru that are used. The page.flags usage comes to mind, and for
example, what would "kmem -p" display for the missing pages?
Or "kmem <address>"? And would "kmem -i" display invalid
data?
I'm just speculating off the top of my head, but the page structure is
such a fundamental data structure with several of its fields being used,
just enter "help -o page_" to see all of its potential member usages.
So I am submitting two patches for your consideration, should the patch
to exclude unused vmemmap pages be taken into makedumpfile.
- [PATCH 1/2] crash: initial note of excluded page structures
This one makes crash startup look like this:
This program has absolutely no warranty. Enter "help warranty" for
details.
NOTE: Unused vmemmap page structures are excluded from this dump.
GNU gdb (GDB) 7.6
Copyright (C) 2013 Free Software Foundation, Inc.
- [PATCH 1/2] crash: kmem warnings for excluded page structures
This patch modifies kmem options -f, -F, -s addr, -S addr, and -i.
Those are the only options that I could detect looking for excluded pages.
This patch applies on top of the first, and adds some warnings to the
output of these kmem options. For example:
crash> kmem -f
Note: kmem -f may fail because unused page structures are excluded from
this dump.
NODE
0
ZONE NAME SIZE FREE MEM_MAP START_PADDR
START_MAPNR
0 DMA 4095 3934 ffffea0000000038 1000 0
AREA SIZE FREE_AREA_STRUCT BLOCKS PAGES
0 4k ffff880000013068 2 2
...
crash> kmem -i
PAGES TOTAL PERCENTAGE
TOTAL MEM 128008147 488.3 GB ----
FREE 127599276 486.8 GB 99% of TOTAL MEM
USED 408871 1.6 GB 0% of TOTAL MEM
SHARED 11049 43.2 MB 0% of TOTAL MEM
BUFFERS 5722 22.4 MB 0% of TOTAL MEM
CACHED 44638 174.4 MB 0% of TOTAL MEM
SLAB 62139 242.7 MB 0% of TOTAL MEM
TOTAL SWAP 4893032 18.7 GB ----
SWAP USED 0 0 0% of TOTAL SWAP
SWAP FREE 4893032 18.7 GB 100% of TOTAL SWAP
Note: 1970727 free pages not found (excluded); results are incomplete.
Unused page structures are excluded from this dump.
-Cliff Wickman
cpw(a)sgi.com
Cliff,
Can you make this patch far simpler? I would prefer that an
error message *follows* the gdb banner, i.e., where a number of
current warnings get displayed? You seem to be showing it just
above the gdb banner, where it kind of gets lost.
Secondly, I'm not sure where/how you are determining the 1970727
pages that are excluded. Is it possible to put that in the
early warning message?
Thirdly, I'm not convinced that the kmem locations where you
are adding per-option warning messages are going to be the only
place where problems may arise. For that reason, I would
not even bother putting them in all those various locations,
but rather listing the "known" commands that may fail in the
early warning message. So do something like:
$ crash vmlinux vmcore
crash 7.0.4
Copyright (C) 2002-2013 Red Hat, Inc.
Copyright (C) 2004, 2005, 2006, 2010 IBM Corporation
Copyright (C) 1999-2006 Hewlett-Packard Co
Copyright (C) 2005, 2006, 2011, 2012 Fujitsu Limited
Copyright (C) 2006, 2007 VA Linux Systems Japan K.K.
Copyright (C) 2005, 2011 NEC Corporation
Copyright (C) 1999, 2002, 2007 Silicon Graphics, Inc.
Copyright (C) 1999, 2000, 2001, 2002 Mission Critical Linux, Inc.
This program is free software, covered by the GNU General Public License,
and you are welcome to change it and/or distribute copies of it under
certain conditions. Enter "help copying" to see the conditions.
This program has absolutely no warranty. Enter "help warranty" for details.
GNU gdb (GDB) 7.6
Copyright (C) 2013 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-unknown-linux-gnu"...
WARNING: ominous warning message indicating (number) page structures have been
excluded, and that the kmem command will be affected...
KERNEL: /usr/lib/debug/lib/modules/3.9.10-100.fc17.x86_64/vmlinux
DUMPFILE: /dev/crash
CPUS: 4
DATE: Wed Jan 8 15:27:14 2014
UPTIME: 57 days, 05:38:32
...
Dave
3:06 p.m.
On Wed, Jan 08, 2014 at 03:32:01PM -0500, Dave Anderson wrote:
----- Original Message -----
>
>
>
> I have proposed a patch to makedumpfile to (optionally) exclude from
> a dump the page structures representing excluded pages.
> The idea being that a system with terabytes of system memory has
> millions of pages of page structures. And most of them are unneeded.
>
> That patch thread begins here:
> http://marc.info/?l=kexec&m=138853299130125&w=2
>
> Dave [Anderson] raised these crash-related issues;
> Although I'm sure you tested this, I find it amazing that
> only the "kmem -[fF]" option is the only command option
> that is affected?
> If I'm not mistaken, this would be the first time that legitimate
> kernel data would be excluded from the dump, and the user would
> have no obvious way of knowing that it had been done, correct?
> If it were encoded in the in the header somewhere, at least a
> warning message could be printed during crash initialization.
> ...
> Right, but look at all of the other page struct offsets in addition to
> page.lru that are used. The page.flags usage comes to mind, and for
> example, what would "kmem -p" display for the missing pages?
> Or "kmem <address>"? And would "kmem -i" display
invalid data?
> I'm just speculating off the top of my head, but the page structure is
> such a fundamental data structure with several of its fields being used,
> just enter "help -o page_" to see all of its potential member usages.
>
> So I am submitting two patches for your consideration, should the patch
> to exclude unused vmemmap pages be taken into makedumpfile.
>
> - [PATCH 1/2] crash: initial note of excluded page structures
> This one makes crash startup look like this:
> This program has absolutely no warranty. Enter "help warranty" for
> details.
>
> NOTE: Unused vmemmap page structures are excluded from this dump.
> GNU gdb (GDB) 7.6
> Copyright (C) 2013 Free Software Foundation, Inc.
>
> - [PATCH 1/2] crash: kmem warnings for excluded page structures
> This patch modifies kmem options -f, -F, -s addr, -S addr, and -i.
> Those are the only options that I could detect looking for excluded pages.
>
> This patch applies on top of the first, and adds some warnings to the
> output of these kmem options. For example:
>
> crash> kmem -f
> Note: kmem -f may fail because unused page structures are excluded from
> this dump.
> NODE
> 0
> ZONE NAME SIZE FREE MEM_MAP START_PADDR
> START_MAPNR
> 0 DMA 4095 3934 ffffea0000000038 1000 0
> AREA SIZE FREE_AREA_STRUCT BLOCKS PAGES
> 0 4k ffff880000013068 2 2
> ...
>
> crash> kmem -i
> PAGES TOTAL PERCENTAGE
> TOTAL MEM 128008147 488.3 GB ----
> FREE 127599276 486.8 GB 99% of TOTAL MEM
> USED 408871 1.6 GB 0% of TOTAL MEM
> SHARED 11049 43.2 MB 0% of TOTAL MEM
> BUFFERS 5722 22.4 MB 0% of TOTAL MEM
> CACHED 44638 174.4 MB 0% of TOTAL MEM
> SLAB 62139 242.7 MB 0% of TOTAL MEM
>
> TOTAL SWAP 4893032 18.7 GB ----
> SWAP USED 0 0 0% of TOTAL SWAP
> SWAP FREE 4893032 18.7 GB 100% of TOTAL SWAP
>
> Note: 1970727 free pages not found (excluded); results are incomplete.
> Unused page structures are excluded from this dump.
>
> -Cliff Wickman
> cpw(a)sgi.com
Cliff,
Can you make this patch far simpler? I would prefer that an
error message *follows* the gdb banner, i.e., where a number of
current warnings get displayed? You seem to be showing it just
above the gdb banner, where it kind of gets lost.
Okay I'll move the message.
Secondly, I'm not sure where/how you are determining the 1970727
pages that are excluded. Is it possible to put that in the
early warning message?
The 1970727 pages were counted by the second patch. That is, I
put a counter in the readmem path that counted 'excluded' errors.
So there were 1970727 such errors during the execution of the
kmem -i.
It is probably not worth it to make a place for a new number
in the dump header. Knowing how many million pages were
excluded won't help if 'your' problem page was one of them.
Thirdly, I'm not convinced that the kmem locations where you
are adding per-option warning messages are going to be the only
place where problems may arise. For that reason, I would
not even bother putting them in all those various locations,
but rather listing the "known" commands that may fail in the
early warning message. So do something like:
Drop the 2nd patch then. If you think it might be more misleading
then helpful I'm fine with just the initialization warning.
$ crash vmlinux vmcore
crash 7.0.4
Copyright (C) 2002-2013 Red Hat, Inc.
Copyright (C) 2004, 2005, 2006, 2010 IBM Corporation
Copyright (C) 1999-2006 Hewlett-Packard Co
Copyright (C) 2005, 2006, 2011, 2012 Fujitsu Limited
Copyright (C) 2006, 2007 VA Linux Systems Japan K.K.
Copyright (C) 2005, 2011 NEC Corporation
Copyright (C) 1999, 2002, 2007 Silicon Graphics, Inc.
Copyright (C) 1999, 2000, 2001, 2002 Mission Critical Linux, Inc.
This program is free software, covered by the GNU General Public License,
and you are welcome to change it and/or distribute copies of it under
certain conditions. Enter "help copying" to see the conditions.
This program has absolutely no warranty. Enter "help warranty" for details.
GNU gdb (GDB) 7.6
Copyright (C) 2013 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-unknown-linux-gnu"...
WARNING: ominous warning message indicating (number) page structures have been
excluded, and that the kmem command will be affected...
KERNEL: /usr/lib/debug/lib/modules/3.9.10-100.fc17.x86_64/vmlinux
DUMPFILE: /dev/crash
CPUS: 4
DATE: Wed Jan 8 15:27:14 2014
UPTIME: 57 days, 05:38:32
...
Dave
--
Cliff Wickman
SGI
cpw(a)sgi.com
(651) 683-3824
3:18 p.m.
----- Original Message -----
On Wed, Jan 08, 2014 at 03:32:01PM -0500, Dave Anderson wrote:
>
>
> ----- Original Message -----
> >
> >
> >
> > I have proposed a patch to makedumpfile to (optionally) exclude from
> > a dump the page structures representing excluded pages.
> > The idea being that a system with terabytes of system memory has
> > millions of pages of page structures. And most of them are unneeded.
> >
> > That patch thread begins here:
> > http://marc.info/?l=kexec&m=138853299130125&w=2
> >
> > Dave [Anderson] raised these crash-related issues;
> > Although I'm sure you tested this, I find it amazing that
> > only the "kmem -[fF]" option is the only command option
> > that is affected?
> > If I'm not mistaken, this would be the first time that legitimate
> > kernel data would be excluded from the dump, and the user would
> > have no obvious way of knowing that it had been done, correct?
> > If it were encoded in the in the header somewhere, at least a
> > warning message could be printed during crash initialization.
> > ...
> > Right, but look at all of the other page struct offsets in addition to
> > page.lru that are used. The page.flags usage comes to mind, and for
> > example, what would "kmem -p" display for the missing pages?
> > Or "kmem <address>"? And would "kmem -i" display
invalid data?
> > I'm just speculating off the top of my head, but the page structure is
> > such a fundamental data structure with several of its fields being
> > used,
> > just enter "help -o page_" to see all of its potential member
usages.
> >
> > So I am submitting two patches for your consideration, should the patch
> > to exclude unused vmemmap pages be taken into makedumpfile.
> >
> > - [PATCH 1/2] crash: initial note of excluded page structures
> > This one makes crash startup look like this:
> > This program has absolutely no warranty. Enter "help warranty"
for
> > details.
> >
> > NOTE: Unused vmemmap page structures are excluded from this dump.
> > GNU gdb (GDB) 7.6
> > Copyright (C) 2013 Free Software Foundation, Inc.
> >
> > - [PATCH 1/2] crash: kmem warnings for excluded page structures
> > This patch modifies kmem options -f, -F, -s addr, -S addr, and -i.
> > Those are the only options that I could detect looking for excluded
> > pages.
> >
> > This patch applies on top of the first, and adds some warnings to the
> > output of these kmem options. For example:
> >
> > crash> kmem -f
> > Note: kmem -f may fail because unused page structures are excluded
> > from
> > this dump.
> > NODE
> > 0
> > ZONE NAME SIZE FREE MEM_MAP START_PADDR
> > START_MAPNR
> > 0 DMA 4095 3934 ffffea0000000038 1000 0
> > AREA SIZE FREE_AREA_STRUCT BLOCKS PAGES
> > 0 4k ffff880000013068 2 2
> > ...
> >
> > crash> kmem -i
> > PAGES TOTAL PERCENTAGE
> > TOTAL MEM 128008147 488.3 GB ----
> > FREE 127599276 486.8 GB 99% of TOTAL MEM
> > USED 408871 1.6 GB 0% of TOTAL MEM
> > SHARED 11049 43.2 MB 0% of TOTAL MEM
> > BUFFERS 5722 22.4 MB 0% of TOTAL MEM
> > CACHED 44638 174.4 MB 0% of TOTAL MEM
> > SLAB 62139 242.7 MB 0% of TOTAL MEM
> >
> > TOTAL SWAP 4893032 18.7 GB ----
> > SWAP USED 0 0 0% of TOTAL SWAP
> > SWAP FREE 4893032 18.7 GB 100% of TOTAL SWAP
> >
> > Note: 1970727 free pages not found (excluded); results are incomplete.
> > Unused page structures are excluded from this dump.
> >
> > -Cliff Wickman
> > cpw(a)sgi.com
>
> Cliff,
>
> Can you make this patch far simpler? I would prefer that an
> error message *follows* the gdb banner, i.e., where a number of
> current warnings get displayed? You seem to be showing it just
> above the gdb banner, where it kind of gets lost.
Okay I'll move the message.
> Secondly, I'm not sure where/how you are determining the 1970727
> pages that are excluded. Is it possible to put that in the
> early warning message?
The 1970727 pages were counted by the second patch. That is, I
put a counter in the readmem path that counted 'excluded' errors.
So there were 1970727 such errors during the execution of the
kmem -i.
It is probably not worth it to make a place for a new number
in the dump header. Knowing how many million pages were
excluded won't help if 'your' problem page was one of them.
>
> Thirdly, I'm not convinced that the kmem locations where you
> are adding per-option warning messages are going to be the only
> place where problems may arise. For that reason, I would
> not even bother putting them in all those various locations,
> but rather listing the "known" commands that may fail in the
> early warning message. So do something like:
Drop the 2nd patch then. If you think it might be more misleading
then helpful I'm fine with just the initialization warning.
OK -- I just don't want to clutter the code with a bunch of per-option
warning messages, but would rather have an ominous, as-wordy-as-you'd-like-it,
message delivered right up front.
Take a look at "ERASEINFO_DATA" in the crash sources, and do a similar thing.
It's the same concept, where in that case, sensitive kernel data may have been
erased from the vmcore.
Thanks,
Dave
4:24 p.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
Version 2
- Moves the warning to this point:
...
This GDB was configured as "x86_64-unknown-linux-gnu"...
WARNING: All unused vmemmap page structures are excluded from this dump.
This will cause failures of the kmem command if it attempts to
walk any list of free pages (with options -f -F -i -s and -S).
SYSTEM MAP: /boot/System.map-2.6.32-cpw
...
- Drop patch 2, which added warnings to individual kmem options.
- Feel free to change the wording of the warning.
And this patch is contingent upon the acceptance of a change to the
makedumpfile command.
http://marc.info/?l=kexec&m=138853299130125&w=2
If makedumpfile excludes unused page structures it will flag that
fact in the dump header.
(There are about 3.67 million pages full of page structures for
every tera byte of system memory. The great bulk of those
page structures are not needed.)
Their exclusion is a makedumpfile option.
Crash will display a note during initialization if such structures
have been excluded. Crash commands that walk page freelists, for
example, will fail. So the note will help the user understand why.
Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
---
defs.h | 1 +
diskdump.c | 3 +++
diskdump.h | 1 +
main.c | 6 ++++++
4 files changed, 11 insertions(+)
Index: crash-7.0.4/diskdump.c
===================================================================
--- crash-7.0.4.orig/diskdump.c
+++ crash-7.0.4/diskdump.c
@@ -749,6 +749,9 @@ restart:
dd->valid_pages[i]++;
}
+ if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
+ pc->flags2 |= VMEXCLUDED;
+
return TRUE;
err:
Index: crash-7.0.4/diskdump.h
===================================================================
--- crash-7.0.4.orig/diskdump.h
+++ crash-7.0.4/diskdump.h
@@ -84,6 +84,7 @@ struct kdump_sub_header {
#define DUMP_DH_COMPRESSED_ZLIB 0x1 /* page is compressed with zlib */
#define DUMP_DH_COMPRESSED_LZO 0x2 /* page is compressed with lzo */
#define DUMP_DH_COMPRESSED_SNAPPY 0x4 /* page is compressed with snappy */
+#define DUMP_DH_EXCLUDED_VMEMMAP 0x8 /* unused vmemmap pages are excluded */
/* descriptor of each page for vmcore */
typedef struct page_desc {
Index: crash-7.0.4/defs.h
===================================================================
--- crash-7.0.4.orig/defs.h
+++ crash-7.0.4/defs.h
@@ -505,6 +505,7 @@ struct program_context {
#define VMCOREINFO (0x400ULL)
#define ALLOW_FP (0x800ULL)
#define REM_PAUSED_F (0x1000ULL)
+#define VMEXCLUDED (0x2000ULL)
#define REMOTE_PAUSED() (pc->flags2 & REM_PAUSED_F)
char *cleanup;
char *namelist_orig;
Index: crash-7.0.4/main.c
===================================================================
--- crash-7.0.4.orig/main.c
+++ crash-7.0.4/main.c
@@ -662,6 +662,12 @@ main_loop(void)
} else
SIGACTION(SIGINT, restart, &pc->sigaction, NULL);
+ if (pc->flags2 & VMEXCLUDED)
+ fprintf(fp,
+ "WARNING: All unused vmemmap page structures are excluded from this
dump.\n"
+ " This will cause failures of the kmem command if it attempts to\n"
+ " walk any list of free pages (with options -f -F -i -s and
-S).\n\n");
+
/*
* Display system statistics and current context.
*/
10:40 a.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
----- Original Message -----
Version 2
- Moves the warning to this point:
...
This GDB was configured as "x86_64-unknown-linux-gnu"...
WARNING: All unused vmemmap page structures are excluded from this
dump.
This will cause failures of the kmem command if it attempts to
walk any list of free pages (with options -f -F -i -s and -S).
SYSTEM MAP: /boot/System.map-2.6.32-cpw
...
- Drop patch 2, which added warnings to individual kmem options.
- Feel free to change the wording of the warning.
And this patch is contingent upon the acceptance of a change to the
makedumpfile command.
http://marc.info/?l=kexec&m=138853299130125&w=2
If makedumpfile excludes unused page structures it will flag that
fact in the dump header.
(There are about 3.67 million pages full of page structures for
every tera byte of system memory. The great bulk of those
page structures are not needed.)
Their exclusion is a makedumpfile option.
Crash will display a note during initialization if such structures
have been excluded. Crash commands that walk page freelists, for
example, will fail. So the note will help the user understand why.
Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
---
defs.h | 1 +
diskdump.c | 3 +++
diskdump.h | 1 +
main.c | 6 ++++++
4 files changed, 11 insertions(+)
Index: crash-7.0.4/diskdump.c
===================================================================
--- crash-7.0.4.orig/diskdump.c
+++ crash-7.0.4/diskdump.c
@@ -749,6 +749,9 @@ restart:
dd->valid_pages[i]++;
}
+ if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
+ pc->flags2 |= VMEXCLUDED;
+
return TRUE;
err:
Index: crash-7.0.4/diskdump.h
===================================================================
--- crash-7.0.4.orig/diskdump.h
+++ crash-7.0.4/diskdump.h
@@ -84,6 +84,7 @@ struct kdump_sub_header {
#define DUMP_DH_COMPRESSED_ZLIB 0x1 /* page is compressed with zlib */
#define DUMP_DH_COMPRESSED_LZO 0x2 /* page is compressed with lzo */
#define DUMP_DH_COMPRESSED_SNAPPY 0x4 /* page is compressed with snappy
*/
+#define DUMP_DH_EXCLUDED_VMEMMAP 0x8 /* unused vmemmap pages are
excluded */
/* descriptor of each page for vmcore */
typedef struct page_desc {
Index: crash-7.0.4/defs.h
===================================================================
--- crash-7.0.4.orig/defs.h
+++ crash-7.0.4/defs.h
@@ -505,6 +505,7 @@ struct program_context {
#define VMCOREINFO (0x400ULL)
#define ALLOW_FP (0x800ULL)
#define REM_PAUSED_F (0x1000ULL)
+#define VMEXCLUDED (0x2000ULL)
#define REMOTE_PAUSED() (pc->flags2 & REM_PAUSED_F)
char *cleanup;
char *namelist_orig;
Index: crash-7.0.4/main.c
===================================================================
--- crash-7.0.4.orig/main.c
+++ crash-7.0.4/main.c
@@ -662,6 +662,12 @@ main_loop(void)
} else
SIGACTION(SIGINT, restart, &pc->sigaction, NULL);
+ if (pc->flags2 & VMEXCLUDED)
+ fprintf(fp,
+ "WARNING: All unused vmemmap page structures are excluded from this
dump.\n"
+ " This will cause failures of the kmem command if it attempts to\n"
+ " walk any list of free pages (with options -f -F -i -s and
-S).\n\n");
+
/*
* Display system statistics and current context.
*/
This patch looks reasonable.
BTW, what happens if you enter "kmem <address>" alone with no option?
It should fail as well, no?
Dave
11 a.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
On Thu, Jan 09, 2014 at 11:40:59AM -0500, Dave Anderson wrote:
----- Original Message -----
>
> Version 2
> - Moves the warning to this point:
> ...
> This GDB was configured as "x86_64-unknown-linux-gnu"...
>
> WARNING: All unused vmemmap page structures are excluded from this
> dump.
> This will cause failures of the kmem command if it attempts to
> walk any list of free pages (with options -f -F -i -s and -S).
>
> SYSTEM MAP: /boot/System.map-2.6.32-cpw
> ...
>
> - Drop patch 2, which added warnings to individual kmem options.
>
> - Feel free to change the wording of the warning.
>
> And this patch is contingent upon the acceptance of a change to the
> makedumpfile command.
> http://marc.info/?l=kexec&m=138853299130125&w=2
>
>
> If makedumpfile excludes unused page structures it will flag that
> fact in the dump header.
> (There are about 3.67 million pages full of page structures for
> every tera byte of system memory. The great bulk of those
> page structures are not needed.)
> Their exclusion is a makedumpfile option.
>
> Crash will display a note during initialization if such structures
> have been excluded. Crash commands that walk page freelists, for
> example, will fail. So the note will help the user understand why.
>
> Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
> ---
> defs.h | 1 +
> diskdump.c | 3 +++
> diskdump.h | 1 +
> main.c | 6 ++++++
> 4 files changed, 11 insertions(+)
>
> Index: crash-7.0.4/diskdump.c
> ===================================================================
> --- crash-7.0.4.orig/diskdump.c
> +++ crash-7.0.4/diskdump.c
> @@ -749,6 +749,9 @@ restart:
> dd->valid_pages[i]++;
> }
>
> + if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
> + pc->flags2 |= VMEXCLUDED;
> +
> return TRUE;
>
> err:
> Index: crash-7.0.4/diskdump.h
> ===================================================================
> --- crash-7.0.4.orig/diskdump.h
> +++ crash-7.0.4/diskdump.h
> @@ -84,6 +84,7 @@ struct kdump_sub_header {
> #define DUMP_DH_COMPRESSED_ZLIB 0x1 /* page is compressed with zlib */
> #define DUMP_DH_COMPRESSED_LZO 0x2 /* page is compressed with lzo */
> #define DUMP_DH_COMPRESSED_SNAPPY 0x4 /* page is compressed with snappy
> */
> +#define DUMP_DH_EXCLUDED_VMEMMAP 0x8 /* unused vmemmap pages are
> excluded */
>
> /* descriptor of each page for vmcore */
> typedef struct page_desc {
> Index: crash-7.0.4/defs.h
> ===================================================================
> --- crash-7.0.4.orig/defs.h
> +++ crash-7.0.4/defs.h
> @@ -505,6 +505,7 @@ struct program_context {
> #define VMCOREINFO (0x400ULL)
> #define ALLOW_FP (0x800ULL)
> #define REM_PAUSED_F (0x1000ULL)
> +#define VMEXCLUDED (0x2000ULL)
> #define REMOTE_PAUSED() (pc->flags2 & REM_PAUSED_F)
> char *cleanup;
> char *namelist_orig;
> Index: crash-7.0.4/main.c
> ===================================================================
> --- crash-7.0.4.orig/main.c
> +++ crash-7.0.4/main.c
> @@ -662,6 +662,12 @@ main_loop(void)
> } else
> SIGACTION(SIGINT, restart, &pc->sigaction, NULL);
>
> + if (pc->flags2 & VMEXCLUDED)
> + fprintf(fp,
> + "WARNING: All unused vmemmap page structures are excluded from this
dump.\n"
> + " This will cause failures of the kmem command if it attempts
to\n"
> + " walk any list of free pages (with options -f -F -i -s and
-S).\n\n");
> +
> /*
> * Display system statistics and current context.
> */
>
This patch looks reasonable.
BTW, what happens if you enter "kmem <address>" alone with no option?
It should fail as well, no?
Yes, it does.
crash> kmem 133360
kmem: page excluded: kernel virtual address: ffffea0000004350 type:
"page.lru.next"
crash> kmem ffff8897baae7540
CACHE NAME OBJSIZE ALLOCATED TOTAL SLABS SSIZE
ffff88c7bec70040 task_struct 2656 4578 5361 1787 8k
SLAB MEMORY TOTAL ALLOCATED FREE
ffff8897baae6040 ffff8897baae6080 3 2 1
FREE / [ALLOCATED]
[ffff8897baae7540]
kmem: page excluded: kernel virtual address: ffffea0000007028 type: "first list
entry"
Another way to search a freelist, that I didn't notice.
Do you want to add that to the warning?
-Cliff
--
Cliff Wickman
SGI
cpw(a)sgi.com
(651) 683-3824
12:21 p.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
----- Original Message -----
On Thu, Jan 09, 2014 at 11:40:59AM -0500, Dave Anderson wrote:
>
>
> ----- Original Message -----
> >
> > Version 2
> > - Moves the warning to this point:
> > ...
> > This GDB was configured as "x86_64-unknown-linux-gnu"...
> >
> > WARNING: All unused vmemmap page structures are excluded from this
> > dump.
> > This will cause failures of the kmem command if it
> > attempts to
> > walk any list of free pages (with options -f -F -i -s and
> > -S).
> >
> > SYSTEM MAP: /boot/System.map-2.6.32-cpw
> > ...
> >
> > - Drop patch 2, which added warnings to individual kmem options.
> >
> > - Feel free to change the wording of the warning.
> >
> > And this patch is contingent upon the acceptance of a change to the
> > makedumpfile command.
> > http://marc.info/?l=kexec&m=138853299130125&w=2
> >
> >
> > If makedumpfile excludes unused page structures it will flag that
> > fact in the dump header.
> > (There are about 3.67 million pages full of page structures for
> > every tera byte of system memory. The great bulk of those
> > page structures are not needed.)
> > Their exclusion is a makedumpfile option.
> >
> > Crash will display a note during initialization if such structures
> > have been excluded. Crash commands that walk page freelists, for
> > example, will fail. So the note will help the user understand why.
> >
> > Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
> > ---
> > defs.h | 1 +
> > diskdump.c | 3 +++
> > diskdump.h | 1 +
> > main.c | 6 ++++++
> > 4 files changed, 11 insertions(+)
> >
> > Index: crash-7.0.4/diskdump.c
> > ===================================================================
> > --- crash-7.0.4.orig/diskdump.c
> > +++ crash-7.0.4/diskdump.c
> > @@ -749,6 +749,9 @@ restart:
> > dd->valid_pages[i]++;
> > }
> >
> > + if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
> > + pc->flags2 |= VMEXCLUDED;
> > +
> > return TRUE;
> >
> > err:
> > Index: crash-7.0.4/diskdump.h
> > ===================================================================
> > --- crash-7.0.4.orig/diskdump.h
> > +++ crash-7.0.4/diskdump.h
> > @@ -84,6 +84,7 @@ struct kdump_sub_header {
> > #define DUMP_DH_COMPRESSED_ZLIB 0x1 /* page is compressed with zlib
> > */
> > #define DUMP_DH_COMPRESSED_LZO 0x2 /* page is compressed with lzo
> > */
> > #define DUMP_DH_COMPRESSED_SNAPPY 0x4 /* page is compressed with
> > snappy
> > */
> > +#define DUMP_DH_EXCLUDED_VMEMMAP 0x8 /* unused vmemmap pages are
> > excluded */
> >
> > /* descriptor of each page for vmcore */
> > typedef struct page_desc {
> > Index: crash-7.0.4/defs.h
> > ===================================================================
> > --- crash-7.0.4.orig/defs.h
> > +++ crash-7.0.4/defs.h
> > @@ -505,6 +505,7 @@ struct program_context {
> > #define VMCOREINFO (0x400ULL)
> > #define ALLOW_FP (0x800ULL)
> > #define REM_PAUSED_F (0x1000ULL)
> > +#define VMEXCLUDED (0x2000ULL)
> > #define REMOTE_PAUSED() (pc->flags2 & REM_PAUSED_F)
> > char *cleanup;
> > char *namelist_orig;
> > Index: crash-7.0.4/main.c
> > ===================================================================
> > --- crash-7.0.4.orig/main.c
> > +++ crash-7.0.4/main.c
> > @@ -662,6 +662,12 @@ main_loop(void)
> > } else
> > SIGACTION(SIGINT, restart, &pc->sigaction, NULL);
> >
> > + if (pc->flags2 & VMEXCLUDED)
> > + fprintf(fp,
> > + "WARNING: All unused vmemmap page structures are excluded from
> > this dump.\n"
> > + " This will cause failures of the kmem command if it attempts
> > to\n"
> > + " walk any list of free pages (with options -f -F -i -s and
> > -S).\n\n");
> > +
> > /*
> > * Display system statistics and current context.
> > */
> >
>
> This patch looks reasonable.
>
> BTW, what happens if you enter "kmem <address>" alone with no
option?
> It should fail as well, no?
Yes, it does.
crash> kmem 133360
kmem: page excluded: kernel virtual address: ffffea0000004350 type:
"page.lru.next"
crash> kmem ffff8897baae7540
CACHE NAME OBJSIZE ALLOCATED TOTAL SLABS
SSIZE
ffff88c7bec70040 task_struct 2656 4578 5361 1787
8k
SLAB MEMORY TOTAL ALLOCATED FREE
ffff8897baae6040 ffff8897baae6080 3 2 1
FREE / [ALLOCATED]
[ffff8897baae7540]
kmem: page excluded: kernel virtual address: ffffea0000007028 type: "first
list entry"
Another way to search a freelist, that I didn't notice.
Do you want to add that to the warning?
I'm thinking that there might be others as well. Like how does "kmem -p"
handle
the missing page structs?
It might be better to just generally warn that results may be unpredictable
because any command that accesses page structures, such as kmem, may fail.
By listing specific options, it makes it sound like that they are the only
ones, so it may be better to leave it open-ended.
Dave
2:54 p.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
On Thu, Jan 09, 2014 at 01:21:35PM -0500, Dave Anderson wrote:
----- Original Message -----
> On Thu, Jan 09, 2014 at 11:40:59AM -0500, Dave Anderson wrote:
> >
> >
> > ----- Original Message -----
> > >
> > > Version 2
> > > - Moves the warning to this point:
> > > ...
> > > This GDB was configured as "x86_64-unknown-linux-gnu"...
> > >
> > > WARNING: All unused vmemmap page structures are excluded from this
> > > dump.
> > > This will cause failures of the kmem command if it
> > > attempts to
> > > walk any list of free pages (with options -f -F -i -s and
> > > -S).
> > >
> > > SYSTEM MAP: /boot/System.map-2.6.32-cpw
> > > ...
> > >
> > > - Drop patch 2, which added warnings to individual kmem options.
> > >
> > > - Feel free to change the wording of the warning.
> > >
> > > And this patch is contingent upon the acceptance of a change to the
> > > makedumpfile command.
> > > http://marc.info/?l=kexec&m=138853299130125&w=2
> > >
> > >
> > > If makedumpfile excludes unused page structures it will flag that
> > > fact in the dump header.
> > > (There are about 3.67 million pages full of page structures for
> > > every tera byte of system memory. The great bulk of those
> > > page structures are not needed.)
> > > Their exclusion is a makedumpfile option.
> > >
> > > Crash will display a note during initialization if such structures
> > > have been excluded. Crash commands that walk page freelists, for
> > > example, will fail. So the note will help the user understand why.
> > >
> > > Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
> > > ---
> > > defs.h | 1 +
> > > diskdump.c | 3 +++
> > > diskdump.h | 1 +
> > > main.c | 6 ++++++
> > > 4 files changed, 11 insertions(+)
> > >
> > > Index: crash-7.0.4/diskdump.c
> > > ===================================================================
> > > --- crash-7.0.4.orig/diskdump.c
> > > +++ crash-7.0.4/diskdump.c
> > > @@ -749,6 +749,9 @@ restart:
> > > dd->valid_pages[i]++;
> > > }
> > >
> > > + if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
> > > + pc->flags2 |= VMEXCLUDED;
> > > +
> > > return TRUE;
> > >
> > > err:
> > > Index: crash-7.0.4/diskdump.h
> > > ===================================================================
> > > --- crash-7.0.4.orig/diskdump.h
> > > +++ crash-7.0.4/diskdump.h
> > > @@ -84,6 +84,7 @@ struct kdump_sub_header {
> > > #define DUMP_DH_COMPRESSED_ZLIB 0x1 /* page is compressed with zlib
> > > */
> > > #define DUMP_DH_COMPRESSED_LZO 0x2 /* page is compressed with lzo
> > > */
> > > #define DUMP_DH_COMPRESSED_SNAPPY 0x4 /* page is compressed with
> > > snappy
> > > */
> > > +#define DUMP_DH_EXCLUDED_VMEMMAP 0x8 /* unused vmemmap pages are
> > > excluded */
> > >
> > > /* descriptor of each page for vmcore */
> > > typedef struct page_desc {
> > > Index: crash-7.0.4/defs.h
> > > ===================================================================
> > > --- crash-7.0.4.orig/defs.h
> > > +++ crash-7.0.4/defs.h
> > > @@ -505,6 +505,7 @@ struct program_context {
> > > #define VMCOREINFO (0x400ULL)
> > > #define ALLOW_FP (0x800ULL)
> > > #define REM_PAUSED_F (0x1000ULL)
> > > +#define VMEXCLUDED (0x2000ULL)
> > > #define REMOTE_PAUSED() (pc->flags2 & REM_PAUSED_F)
> > > char *cleanup;
> > > char *namelist_orig;
> > > Index: crash-7.0.4/main.c
> > > ===================================================================
> > > --- crash-7.0.4.orig/main.c
> > > +++ crash-7.0.4/main.c
> > > @@ -662,6 +662,12 @@ main_loop(void)
> > > } else
> > > SIGACTION(SIGINT, restart, &pc->sigaction, NULL);
> > >
> > > + if (pc->flags2 & VMEXCLUDED)
> > > + fprintf(fp,
> > > + "WARNING: All unused vmemmap page structures are excluded
from
> > > this dump.\n"
> > > + " This will cause failures of the kmem command if it
attempts
> > > to\n"
> > > + " walk any list of free pages (with options -f -F -i -s
and
> > > -S).\n\n");
> > > +
> > > /*
> > > * Display system statistics and current context.
> > > */
> > >
> >
> > This patch looks reasonable.
> >
> > BTW, what happens if you enter "kmem <address>" alone with no
option?
> > It should fail as well, no?
>
> Yes, it does.
>
> crash> kmem 133360
> kmem: page excluded: kernel virtual address: ffffea0000004350 type:
> "page.lru.next"
>
> crash> kmem ffff8897baae7540
> CACHE NAME OBJSIZE ALLOCATED TOTAL SLABS
> SSIZE
> ffff88c7bec70040 task_struct 2656 4578 5361 1787
> 8k
> SLAB MEMORY TOTAL ALLOCATED FREE
> ffff8897baae6040 ffff8897baae6080 3 2 1
> FREE / [ALLOCATED]
> [ffff8897baae7540]
>
> kmem: page excluded: kernel virtual address: ffffea0000007028 type: "first
> list entry"
>
> Another way to search a freelist, that I didn't notice.
> Do you want to add that to the warning?
I'm thinking that there might be others as well. Like how does "kmem -p"
handle
the missing page structs?
kmem -p gives the appearance of working, but is unable to read the
mem_map pages. It is giving false results.
This is an example of what you feared.
If we do this:
---
memory.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: crash-7.0.4/memory.c
===================================================================
--- crash-7.0.4.orig/memory.c
+++ crash-7.0.4/memory.c
@@ -5657,7 +5657,7 @@ fill_mem_map_cache(ulong pp, ulong ppend
* Try to read it in one fell swoop.
*/
if (readmem(pp, KVADDR, page_cache, SIZE(page) * PGMM_CACHED,
- "page struct cache", RETURN_ON_ERROR|QUIET))
+ "page struct cache", FAULT_ON_ERROR))
return;
/*
Then we see the immediate failure:
crash> kmem -p
PAGE PHYSICAL MAPPING INDEX CNT FLAGS
kmem: page excluded: kernel virtual address: ffffea0000004000 type: "page struct
cache"
What should be done about that?
I wonder why the reads of the mem_map were done with RETURN_ON_ERROR|QUIET
to begin with?
-Cliff
It might be better to just generally warn that results may be unpredictable
because any command that accesses page structures, such as kmem, may fail.
By listing specific options, it makes it sound like that they are the only
ones, so it may be better to leave it open-ended.
Dave
--
Cliff Wickman
SGI
cpw(a)sgi.com
(651) 683-3824
3:11 p.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
----- Original Message -----
On Thu, Jan 09, 2014 at 01:21:35PM -0500, Dave Anderson wrote:
>
>
> ----- Original Message -----
> > On Thu, Jan 09, 2014 at 11:40:59AM -0500, Dave Anderson wrote:
> > >
> > >
> > > ----- Original Message -----
> > > >
> > > > Version 2
> > > > - Moves the warning to this point:
> > > > ...
> > > > This GDB was configured as
"x86_64-unknown-linux-gnu"...
> > > >
> > > > WARNING: All unused vmemmap page structures are excluded from
> > > > this
> > > > dump.
> > > > This will cause failures of the kmem command if it
> > > > attempts to
> > > > walk any list of free pages (with options -f -F -i -s
> > > > and
> > > > -S).
> > > >
> > > > SYSTEM MAP: /boot/System.map-2.6.32-cpw
> > > > ...
> > > >
> > > > - Drop patch 2, which added warnings to individual kmem options.
> > > >
> > > > - Feel free to change the wording of the warning.
> > > >
> > > > And this patch is contingent upon the acceptance of a change to the
> > > > makedumpfile command.
> > > > http://marc.info/?l=kexec&m=138853299130125&w=2
> > > >
> > > >
> > > > If makedumpfile excludes unused page structures it will flag that
> > > > fact in the dump header.
> > > > (There are about 3.67 million pages full of page structures for
> > > > every tera byte of system memory. The great bulk of those
> > > > page structures are not needed.)
> > > > Their exclusion is a makedumpfile option.
> > > >
> > > > Crash will display a note during initialization if such structures
> > > > have been excluded. Crash commands that walk page freelists, for
> > > > example, will fail. So the note will help the user understand why.
> > > >
> > > > Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
> > > > ---
> > > > defs.h | 1 +
> > > > diskdump.c | 3 +++
> > > > diskdump.h | 1 +
> > > > main.c | 6 ++++++
> > > > 4 files changed, 11 insertions(+)
> > > >
> > > > Index: crash-7.0.4/diskdump.c
> > > > ===================================================================
> > > > --- crash-7.0.4.orig/diskdump.c
> > > > +++ crash-7.0.4/diskdump.c
> > > > @@ -749,6 +749,9 @@ restart:
> > > > dd->valid_pages[i]++;
> > > > }
> > > >
> > > > + if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
> > > > + pc->flags2 |= VMEXCLUDED;
> > > > +
> > > > return TRUE;
> > > >
> > > > err:
> > > > Index: crash-7.0.4/diskdump.h
> > > > ===================================================================
> > > > --- crash-7.0.4.orig/diskdump.h
> > > > +++ crash-7.0.4/diskdump.h
> > > > @@ -84,6 +84,7 @@ struct kdump_sub_header {
> > > > #define DUMP_DH_COMPRESSED_ZLIB 0x1 /* page is compressed with
> > > > zlib
> > > > */
> > > > #define DUMP_DH_COMPRESSED_LZO 0x2 /* page is compressed with
> > > > lzo
> > > > */
> > > > #define DUMP_DH_COMPRESSED_SNAPPY 0x4 /* page is compressed with
> > > > snappy
> > > > */
> > > > +#define DUMP_DH_EXCLUDED_VMEMMAP 0x8 /* unused vmemmap pages
are
> > > > excluded */
> > > >
> > > > /* descriptor of each page for vmcore */
> > > > typedef struct page_desc {
> > > > Index: crash-7.0.4/defs.h
> > > > ===================================================================
> > > > --- crash-7.0.4.orig/defs.h
> > > > +++ crash-7.0.4/defs.h
> > > > @@ -505,6 +505,7 @@ struct program_context {
> > > > #define VMCOREINFO (0x400ULL)
> > > > #define ALLOW_FP (0x800ULL)
> > > > #define REM_PAUSED_F (0x1000ULL)
> > > > +#define VMEXCLUDED (0x2000ULL)
> > > > #define REMOTE_PAUSED() (pc->flags2 & REM_PAUSED_F)
> > > > char *cleanup;
> > > > char *namelist_orig;
> > > > Index: crash-7.0.4/main.c
> > > > ===================================================================
> > > > --- crash-7.0.4.orig/main.c
> > > > +++ crash-7.0.4/main.c
> > > > @@ -662,6 +662,12 @@ main_loop(void)
> > > > } else
> > > > SIGACTION(SIGINT, restart, &pc->sigaction, NULL);
> > > >
> > > > + if (pc->flags2 & VMEXCLUDED)
> > > > + fprintf(fp,
> > > > + "WARNING: All unused vmemmap page structures are
excluded
> > > > from
> > > > this dump.\n"
> > > > + " This will cause failures of the kmem command if it
> > > > attempts
> > > > to\n"
> > > > + " walk any list of free pages (with options -f -F -i
-s and
> > > > -S).\n\n");
> > > > +
> > > > /*
> > > > * Display system statistics and current context.
> > > > */
> > > >
> > >
> > > This patch looks reasonable.
> > >
> > > BTW, what happens if you enter "kmem <address>" alone with
no option?
> > > It should fail as well, no?
> >
> > Yes, it does.
> >
> > crash> kmem 133360
> > kmem: page excluded: kernel virtual address: ffffea0000004350 type:
> > "page.lru.next"
> >
> > crash> kmem ffff8897baae7540
> > CACHE NAME OBJSIZE ALLOCATED TOTAL SLABS
> > SSIZE
> > ffff88c7bec70040 task_struct 2656 4578 5361 1787
> > 8k
> > SLAB MEMORY TOTAL ALLOCATED FREE
> > ffff8897baae6040 ffff8897baae6080 3 2 1
> > FREE / [ALLOCATED]
> > [ffff8897baae7540]
> >
> > kmem: page excluded: kernel virtual address: ffffea0000007028 type:
"first list entry"
> >
> > Another way to search a freelist, that I didn't notice.
> > Do you want to add that to the warning?
>
> I'm thinking that there might be others as well. Like how does "kmem
-p" handle
> the missing page structs?
kmem -p gives the appearance of working, but is unable to read the
mem_map pages. It is giving false results.
This is an example of what you feared.
If we do this:
---
memory.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: crash-7.0.4/memory.c
===================================================================
--- crash-7.0.4.orig/memory.c
+++ crash-7.0.4/memory.c
@@ -5657,7 +5657,7 @@ fill_mem_map_cache(ulong pp, ulong ppend
* Try to read it in one fell swoop.
*/
if (readmem(pp, KVADDR, page_cache, SIZE(page) * PGMM_CACHED,
- "page struct cache", RETURN_ON_ERROR|QUIET))
+ "page struct cache", FAULT_ON_ERROR))
return;
/*
Then we see the immediate failure:
crash> kmem -p
PAGE PHYSICAL MAPPING INDEX CNT FLAGS
kmem: page excluded: kernel virtual address: ffffea0000004000 type: "page
struct cache"
What should be done about that?
I'd leave it alone -- at least it doesn't fail miserably like the other commands.
I wonder why the reads of the mem_map were done with
RETURN_ON_ERROR|QUIET
to begin with?
It's done to guard against a vmemmap boundary (real page structs vs. unmapped
page structs) that happens to occur within the 512-page-size chunk of memory
that's being cached. So first it tries to read the full chunk in one readmem(),
but if that fails, it breaks up the read requests, zeroing out the unmapped pages.
Dave
-Cliff
>
> It might be better to just generally warn that results may be unpredictable
> because any command that accesses page structures, such as kmem, may fail.
> By listing specific options, it makes it sound like that they are the only
> ones, so it may be better to leave it open-ended.
>
> Dave
--
Cliff Wickman
SGI
cpw(a)sgi.com
(651) 683-3824
4:22 p.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
On Thu, Jan 09, 2014 at 04:11:40PM -0500, Dave Anderson wrote:
----- Original Message -----
> On Thu, Jan 09, 2014 at 01:21:35PM -0500, Dave Anderson wrote:
> >
> >
> > ----- Original Message -----
> > > On Thu, Jan 09, 2014 at 11:40:59AM -0500, Dave Anderson wrote:
> > > >
> > > >
> > > > ----- Original Message -----
> > > > >
> > > > > Version 2
> > > > > - Moves the warning to this point:
> > > > > ...
> > > > > This GDB was configured as
"x86_64-unknown-linux-gnu"...
> > > > >
> > > > > WARNING: All unused vmemmap page structures are excluded
from
> > > > > this
> > > > > dump.
> > > > > This will cause failures of the kmem command if
it
> > > > > attempts to
> > > > > walk any list of free pages (with options -f -F
-i -s
> > > > > and
> > > > > -S).
> > > > >
> > > > > SYSTEM MAP: /boot/System.map-2.6.32-cpw
> > > > > ...
> > > > >
> > > > > - Drop patch 2, which added warnings to individual kmem
options.
> > > > >
> > > > > - Feel free to change the wording of the warning.
> > > > >
> > > > > And this patch is contingent upon the acceptance of a change to
the
> > > > > makedumpfile command.
> > > > > http://marc.info/?l=kexec&m=138853299130125&w=2
> > > > >
> > > > >
> > > > > If makedumpfile excludes unused page structures it will flag
that
> > > > > fact in the dump header.
> > > > > (There are about 3.67 million pages full of page structures for
> > > > > every tera byte of system memory. The great bulk of those
> > > > > page structures are not needed.)
> > > > > Their exclusion is a makedumpfile option.
> > > > >
> > > > > Crash will display a note during initialization if such
structures
> > > > > have been excluded. Crash commands that walk page freelists,
for
> > > > > example, will fail. So the note will help the user understand
why.
> > > > >
> > > > > Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
> > > > > ---
> > > > > defs.h | 1 +
> > > > > diskdump.c | 3 +++
> > > > > diskdump.h | 1 +
> > > > > main.c | 6 ++++++
> > > > > 4 files changed, 11 insertions(+)
> > > > >
> > > > > Index: crash-7.0.4/diskdump.c
> > > > >
===================================================================
> > > > > --- crash-7.0.4.orig/diskdump.c
> > > > > +++ crash-7.0.4/diskdump.c
> > > > > @@ -749,6 +749,9 @@ restart:
> > > > > dd->valid_pages[i]++;
> > > > > }
> > > > >
> > > > > + if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
> > > > > + pc->flags2 |= VMEXCLUDED;
> > > > > +
> > > > > return TRUE;
> > > > >
> > > > > err:
> > > > > Index: crash-7.0.4/diskdump.h
> > > > >
===================================================================
> > > > > --- crash-7.0.4.orig/diskdump.h
> > > > > +++ crash-7.0.4/diskdump.h
> > > > > @@ -84,6 +84,7 @@ struct kdump_sub_header {
> > > > > #define DUMP_DH_COMPRESSED_ZLIB 0x1 /* page is compressed
with
> > > > > zlib
> > > > > */
> > > > > #define DUMP_DH_COMPRESSED_LZO 0x2 /* page is compressed
with
> > > > > lzo
> > > > > */
> > > > > #define DUMP_DH_COMPRESSED_SNAPPY 0x4 /* page is compressed
with
> > > > > snappy
> > > > > */
> > > > > +#define DUMP_DH_EXCLUDED_VMEMMAP 0x8 /* unused vmemmap
pages are
> > > > > excluded */
> > > > >
> > > > > /* descriptor of each page for vmcore */
> > > > > typedef struct page_desc {
> > > > > Index: crash-7.0.4/defs.h
> > > > >
===================================================================
> > > > > --- crash-7.0.4.orig/defs.h
> > > > > +++ crash-7.0.4/defs.h
> > > > > @@ -505,6 +505,7 @@ struct program_context {
> > > > > #define VMCOREINFO (0x400ULL)
> > > > > #define ALLOW_FP (0x800ULL)
> > > > > #define REM_PAUSED_F (0x1000ULL)
> > > > > +#define VMEXCLUDED (0x2000ULL)
> > > > > #define REMOTE_PAUSED() (pc->flags2 & REM_PAUSED_F)
> > > > > char *cleanup;
> > > > > char *namelist_orig;
> > > > > Index: crash-7.0.4/main.c
> > > > >
===================================================================
> > > > > --- crash-7.0.4.orig/main.c
> > > > > +++ crash-7.0.4/main.c
> > > > > @@ -662,6 +662,12 @@ main_loop(void)
> > > > > } else
> > > > > SIGACTION(SIGINT, restart, &pc->sigaction, NULL);
> > > > >
> > > > > + if (pc->flags2 & VMEXCLUDED)
> > > > > + fprintf(fp,
> > > > > + "WARNING: All unused vmemmap page structures are
excluded
> > > > > from
> > > > > this dump.\n"
> > > > > + " This will cause failures of the kmem command if
it
> > > > > attempts
> > > > > to\n"
> > > > > + " walk any list of free pages (with options -f -F
-i -s and
> > > > > -S).\n\n");
> > > > > +
> > > > > /*
> > > > > * Display system statistics and current context.
> > > > > */
> > > > >
> > > >
> > > > This patch looks reasonable.
> > > >
> > > > BTW, what happens if you enter "kmem <address>" alone
with no option?
> > > > It should fail as well, no?
> > >
> > > Yes, it does.
> > >
> > > crash> kmem 133360
> > > kmem: page excluded: kernel virtual address: ffffea0000004350 type:
> > > "page.lru.next"
> > >
> > > crash> kmem ffff8897baae7540
> > > CACHE NAME OBJSIZE ALLOCATED TOTAL SLABS
> > > SSIZE
> > > ffff88c7bec70040 task_struct 2656 4578 5361 1787
> > > 8k
> > > SLAB MEMORY TOTAL ALLOCATED FREE
> > > ffff8897baae6040 ffff8897baae6080 3 2 1
> > > FREE / [ALLOCATED]
> > > [ffff8897baae7540]
> > >
> > > kmem: page excluded: kernel virtual address: ffffea0000007028 type:
"first list entry"
> > >
> > > Another way to search a freelist, that I didn't notice.
> > > Do you want to add that to the warning?
> >
> > I'm thinking that there might be others as well. Like how does "kmem
-p" handle
> > the missing page structs?
>
> kmem -p gives the appearance of working, but is unable to read the
> mem_map pages. It is giving false results.
> This is an example of what you feared.
>
> If we do this:
>
> ---
> memory.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> Index: crash-7.0.4/memory.c
> ===================================================================
> --- crash-7.0.4.orig/memory.c
> +++ crash-7.0.4/memory.c
> @@ -5657,7 +5657,7 @@ fill_mem_map_cache(ulong pp, ulong ppend
> * Try to read it in one fell swoop.
> */
> if (readmem(pp, KVADDR, page_cache, SIZE(page) * PGMM_CACHED,
> - "page struct cache", RETURN_ON_ERROR|QUIET))
> + "page struct cache", FAULT_ON_ERROR))
> return;
>
> /*
>
> Then we see the immediate failure:
>
> crash> kmem -p
> PAGE PHYSICAL MAPPING INDEX CNT FLAGS
> kmem: page excluded: kernel virtual address: ffffea0000004000 type: "page
> struct cache"
>
> What should be done about that?
I'd leave it alone -- at least it doesn't fail miserably like the other
commands.
> I wonder why the reads of the mem_map were done with RETURN_ON_ERROR|QUIET
> to begin with?
It's done to guard against a vmemmap boundary (real page structs vs. unmapped
page structs) that happens to occur within the 512-page-size chunk of memory
that's being cached. So first it tries to read the full chunk in one readmem(),
but if that fails, it breaks up the read requests, zeroing out the unmapped pages.
Dave
Okay. I'll just make issue a more general warning like:
WARNING: All unused vmemmap page structures are excluded from this dump.
This will cause failures in any command that accesses page
structures of pages that are not included in the dump. This
is particularly likely when using several options of the
kmem command.
If that looks reasonable to you.
-Cliff
>
> -Cliff
>
> >
> > It might be better to just generally warn that results may be unpredictable
> > because any command that accesses page structures, such as kmem, may fail.
> > By listing specific options, it makes it sound like that they are the only
> > ones, so it may be better to leave it open-ended.
> >
> > Dave
>
> --
> Cliff Wickman
> SGI
> cpw(a)sgi.com
> (651) 683-3824
>
--
Cliff Wickman
SGI
cpw(a)sgi.com
(651) 683-3824
12:10 a.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
On 2014/01/10 7:23:32, crash-utility-bounces(a)redhat.com wrote:
On Thu, Jan 09, 2014 at 04:11:40PM -0500, Dave Anderson wrote:
> > > > > > Version 2
> > > > > > - Moves the warning to this point:
> > > > > > ...
> > > > > > This GDB was configured as
"x86_64-unknown-linux-gnu"...
> > > > > >
> > > > > > WARNING: All unused vmemmap page structures are
excluded from
> > > > > > this
> > > > > > dump.
> > > > > > This will cause failures of the kmem command
if it
> > > > > > attempts to
> > > > > > walk any list of free pages (with options -f
-F -i -s
> > > > > > and
> > > > > > -S).
> > > > > >
> > > > > > SYSTEM MAP: /boot/System.map-2.6.32-cpw
> > > > > > ...
> > > > > >
> > > > > > - Drop patch 2, which added warnings to individual kmem
options.
> > > > > >
> > > > > > - Feel free to change the wording of the warning.
> > > > > >
> > > > > > And this patch is contingent upon the acceptance of a
change to the
> > > > > > makedumpfile command.
> > > > > > http://marc.info/?l=kexec&m=138853299130125&w=2
> > > > > >
> > > > > >
> > > > > > If makedumpfile excludes unused page structures it will
flag that
> > > > > > fact in the dump header.
> > > > > > (There are about 3.67 million pages full of page structures
for
> > > > > > every tera byte of system memory. The great bulk of
those
> > > > > > page structures are not needed.)
> > > > > > Their exclusion is a makedumpfile option.
> > > > > >
> > > > > > Crash will display a note during initialization if such
structures
> > > > > > have been excluded. Crash commands that walk page
freelists, for
> > > > > > example, will fail. So the note will help the user
understand why.
> > > > > >
> > > > > > Signed-off-by: Cliff Wickman <cpw(a)sgi.com>
> > > > > > ---
> > > > > > defs.h | 1 +
> > > > > > diskdump.c | 3 +++
> > > > > > diskdump.h | 1 +
> > > > > > main.c | 6 ++++++
> > > > > > 4 files changed, 11 insertions(+)
> > > > > >
> > > > > > Index: crash-7.0.4/diskdump.c
> > > > > >
===================================================================
> > > > > > --- crash-7.0.4.orig/diskdump.c
> > > > > > +++ crash-7.0.4/diskdump.c
> > > > > > @@ -749,6 +749,9 @@ restart:
> > > > > > dd->valid_pages[i]++;
> > > > > > }
> > > > > >
> > > > > > + if (header->status & DUMP_DH_EXCLUDED_VMEMMAP)
> > > > > > + pc->flags2 |= VMEXCLUDED;
> > > > > > +
> > > > > > return TRUE;
> > > > > >
> > > > > > err:
> > > > > > Index: crash-7.0.4/diskdump.h
> > > > > >
===================================================================
> > > > > > --- crash-7.0.4.orig/diskdump.h
> > > > > > +++ crash-7.0.4/diskdump.h
> > > > > > @@ -84,6 +84,7 @@ struct kdump_sub_header {
> > > > > > #define DUMP_DH_COMPRESSED_ZLIB 0x1 /* page is
compressed with
> > > > > > zlib
> > > > > > */
> > > > > > #define DUMP_DH_COMPRESSED_LZO 0x2 /* page is
compressed with
> > > > > > lzo
> > > > > > */
> > > > > > #define DUMP_DH_COMPRESSED_SNAPPY 0x4 /* page is
compressed with
> > > > > > snappy
> > > > > > */
> > > > > > +#define DUMP_DH_EXCLUDED_VMEMMAP 0x8 /* unused vmemmap
pages are
> > > > > > excluded */
> > > > > >
> > > > > > /* descriptor of each page for vmcore */
> > > > > > typedef struct page_desc {
> > > > > > Index: crash-7.0.4/defs.h
> > > > > >
===================================================================
> > > > > > --- crash-7.0.4.orig/defs.h
> > > > > > +++ crash-7.0.4/defs.h
> > > > > > @@ -505,6 +505,7 @@ struct program_context {
> > > > > > #define VMCOREINFO (0x400ULL)
> > > > > > #define ALLOW_FP (0x800ULL)
> > > > > > #define REM_PAUSED_F (0x1000ULL)
> > > > > > +#define VMEXCLUDED (0x2000ULL)
> > > > > > #define REMOTE_PAUSED() (pc->flags2 &
REM_PAUSED_F)
> > > > > > char *cleanup;
> > > > > > char *namelist_orig;
> > > > > > Index: crash-7.0.4/main.c
> > > > > >
===================================================================
> > > > > > --- crash-7.0.4.orig/main.c
> > > > > > +++ crash-7.0.4/main.c
> > > > > > @@ -662,6 +662,12 @@ main_loop(void)
> > > > > > } else
> > > > > > SIGACTION(SIGINT, restart, &pc->sigaction,
NULL);
> > > > > >
> > > > > > + if (pc->flags2 & VMEXCLUDED)
> > > > > > + fprintf(fp,
> > > > > > + "WARNING: All unused vmemmap page structures
are excluded
> > > > > > from
> > > > > > this dump.\n"
> > > > > > + " This will cause failures of the kmem
command if it
> > > > > > attempts
> > > > > > to\n"
> > > > > > + " walk any list of free pages (with options
-f -F -i -s and
> > > > > > -S).\n\n");
> > > > > > +
> > > > > > /*
> > > > > > * Display system statistics and current
context.
> > > > > > */
> > > > > >
> > > > >
> > > > > This patch looks reasonable.
> > > > >
> > > > > BTW, what happens if you enter "kmem <address>"
alone with no option?
> > > > > It should fail as well, no?
> > > >
> > > > Yes, it does.
> > > >
> > > > crash> kmem 133360
> > > > kmem: page excluded: kernel virtual address: ffffea0000004350 type:
> > > > "page.lru.next"
> > > >
> > > > crash> kmem ffff8897baae7540
> > > > CACHE NAME OBJSIZE ALLOCATED TOTAL
SLABS
> > > > SSIZE
> > > > ffff88c7bec70040 task_struct 2656 4578 5361
1787
> > > > 8k
> > > > SLAB MEMORY TOTAL ALLOCATED FREE
> > > > ffff8897baae6040 ffff8897baae6080 3 2 1
> > > > FREE / [ALLOCATED]
> > > > [ffff8897baae7540]
> > > >
> > > > kmem: page excluded: kernel virtual address: ffffea0000007028 type:
"first list entry"
> > > >
> > > > Another way to search a freelist, that I didn't notice.
> > > > Do you want to add that to the warning?
> > >
> > > I'm thinking that there might be others as well. Like how does
"kmem -p" handle
> > > the missing page structs?
> >
> > kmem -p gives the appearance of working, but is unable to read the
> > mem_map pages. It is giving false results.
> > This is an example of what you feared.
> >
> > If we do this:
> >
> > ---
> > memory.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > Index: crash-7.0.4/memory.c
> > ===================================================================
> > --- crash-7.0.4.orig/memory.c
> > +++ crash-7.0.4/memory.c
> > @@ -5657,7 +5657,7 @@ fill_mem_map_cache(ulong pp, ulong ppend
> > * Try to read it in one fell swoop.
> > */
> > if (readmem(pp, KVADDR, page_cache, SIZE(page) * PGMM_CACHED,
> > - "page struct cache", RETURN_ON_ERROR|QUIET))
> > + "page struct cache", FAULT_ON_ERROR))
> > return;
> >
> > /*
> >
> > Then we see the immediate failure:
> >
> > crash> kmem -p
> > PAGE PHYSICAL MAPPING INDEX CNT FLAGS
> > kmem: page excluded: kernel virtual address: ffffea0000004000 type:
"page
> > struct cache"
> >
> > What should be done about that?
>
> I'd leave it alone -- at least it doesn't fail miserably like the other
commands.
>
> > I wonder why the reads of the mem_map were done with RETURN_ON_ERROR|QUIET
> > to begin with?
>
> It's done to guard against a vmemmap boundary (real page structs vs. unmapped
> page structs) that happens to occur within the 512-page-size chunk of memory
> that's being cached. So first it tries to read the full chunk in one
readmem(),
> but if that fails, it breaks up the read requests, zeroing out the unmapped pages.
>
> Dave
Okay. I'll just make issue a more general warning like:
WARNING: All unused vmemmap page structures are excluded from this dump.
This will cause failures in any command that accesses page
structures of pages that are not included in the dump. This
is particularly likely when using several options of the
kmem command.
If that looks reasonable to you.
Dave, the excluding vmemmap obviously affect a user's investigation,
is this really acceptable for you ?
There is no chance to re-capture the same dump image,
I think we should be more carefully about filtering out
since it's an irreversible change.
How many users want to get such a broken dump image even if
it could be gotten faster ?
Why we capture dump images, it's for analyzing, of course.
At least, we should supply alternatives to the affected commands.
Thanks
Atsushi Kumagai
-Cliff
>
> >
> > -Cliff
> >
> > >
> > > It might be better to just generally warn that results may be
unpredictable
> > > because any command that accesses page structures, such as kmem, may
fail.
> > > By listing specific options, it makes it sound like that they are the
only
> > > ones, so it may be better to leave it open-ended.
> > >
> > > Dave
> >
> > --
> > Cliff Wickman
> > SGI
> > cpw(a)sgi.com
> > (651) 683-3824
> >
--
Cliff Wickman
SGI
cpw(a)sgi.com
(651) 683-3824
--
Crash-utility mailing list
Crash-utility(a)redhat.com
https://www.redhat.com/mailman/listinfo/crash-utility
8:02 a.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
----- Original Message -----
Dave, the excluding vmemmap obviously affect a user's investigation,
is this really acceptable for you ?
There is no chance to re-capture the same dump image,
I think we should be more carefully about filtering out
since it's an irreversible change.
How many users want to get such a broken dump image even if
it could be gotten faster ?
Why we capture dump images, it's for analyzing, of course.
At least, we should supply alternatives to the affected commands.
Thanks
Atsushi Kumagai
No, I don't like this at all. That is why I immediately objected
to doing the exclusion by default based upon memory size.
But I understand Cliff's request. I also don't think too much of
the eraseinfo/eppic capability either, but I understand why it's
considered necessary.
And so if you do accept it as a feature, I want the up-front warning
message crash utility message to explicitly state that the page-exclusion
is as a direct result of user/administrator actions. Perhaps the warning
message should also indicate that the new "makedumpfile -x" option (I forget
the option letter) was used. At least then the end-user will have
somebody to "blame", and will not consider it a problem with the tools
that were used.
And I absolutely do not want the crash source code sprinkled with a
bunch of alternative code paths to handle what will be a very
rarely-used option.
Thanks,
Dave
7:49 a.m.
New subject: [PATCH 1/1 V2] crash: initial note of excluded page structures
----- Original Message -----
Okay. I'll just make issue a more general warning like:
WARNING: All unused vmemmap page structures are excluded from this dump.
This will cause failures in any command that accesses page
structures of pages that are not included in the dump. This
is particularly likely when using several options of the
kmem command.
If that looks reasonable to you.
-Cliff
The term "unused vmemmap page structures" is confusing, given that they
were certainly used during the kernel's life, and still serve a purpose.
I think the message should make it perfectly clear that the vmemmap page
exclusion was purposely done by the user/administrator.
How about something like: "the vmemmap page structures that are associated
with the memory pages excluded by the "makedumpfile -d <level>" option
have
been excluded from this dump"?
Dave
3830
days inactive
3832
days old
devel@lists.crash-utility.osci.io
15 comments
4 participants
participants (4)
-
Atsushi Kumagai -
Cliff Wickman -
cpw
-
Dave Anderson