| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch adds a -e option to makedumpfile.
The -e option excludes kernel pages that contain nothing but kernel page
structures for pages that are not being included in the dump.
The -e option only works in non-cyclic mode, which its use implies.
And only applies to the x86_64 architecture.
The -e requires the use of --work-dir, as it will create a pfn file in that
work directory. The --work-dir should probably be set up by the distro procedures
which determine the mount point of the root device.
This patch formerly applied after patch:
[PATCH V2] makedumpfile: make --work-dir easier to use
but now it stands alone.
I have tested on large memory systems to demonstrate the importance
of this feature to such systems. See some numbers below.
The most dramatic demonstration was on a 32TB system where the patch
reduced the process from 2 hours to 26 minutes. The size of the dump
would probably have been over 30GB (but I ran out of disk space). It was
reduced to 5.4GB.
A page structure (56 bytes) exists for every 4096-byte page.
This amounts to 3.67 million pages, or about 14GB, per terabyte of system memory!
Without -e an idle 2-terabyte system can be dumped (compressed) to a file of
about 3.6G.
With -e that is reduced to about 456M. And the time and space savings
multiply for each additional terabyte of memory in the system.
Experimental time/size results: (basically idle systems)
Memory Size With -e Without -e
(sec.) (sec.)
(using a sles11sp3 kernel that does not provide mmap of /proc/vmcore:)
1TB 52 244M 257 1.7G
2TB 128 456M 526 3.6G
8TB 780 1.6G 3400 13.8G
16TB 2600 3.1G 9800 (extrapolated, 2:40 is too long to wait)
(using a sles11sp3 kernel that provides mmap of /proc/vmcore:)
16TB 900 3.8G not done
32TB 6000 5.4G not done
(using a sles11sp3 kernel that provides mmap of /proc/vmcore:)
32TB 1600 5.4G 7300 (extrapolated)
(ran out of 19G space before 1/2 done)
The only disadvantage is that various options of the crash 'kmem' command (that
walk lists of page structures) will not work.
Version 7.0.9 of crash is already patched to issue a warning about such commands
when the dump is flagged DUMP_DH_EXCLUDED_VMEMMAP.
Sorry that this patch is large. The vmemmap page scan is done by some very large
functions, and they are all interrelated. I didn't see any point to breaking
them into several inter-dependent patches.
Signed-off-by: Cliff Wickman <cpw@sgi.com>
|
| |
|
|
|
|
|
|
|
|
| |
ENOSPC error.
kdump-compressed:
Dump the bitmap before any page header and page data. This format use
"status" of "disk_dump_header" to indicate that it has been modified.
Signed-of-by: Wang Xiao <wangx.fnst@cn.fujitsu.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch will fix a bug of makedumpfile doesn't work correctly on system
has over 44-bit addressing in compression dump mode.
This bug was posted here:
http://lists.infradead.org/pipermail/kexec/2013-September/009587.html
This patch will add 3 new fields in struct kdump_sub_header.
unsigned long long start_pfn_64; /* header_version 6 and later */
unsigned long long end_pfn_64; /* header_version 6 and later */
unsigned long long max_mapnr_64; /* header_version 6 and later */
And the old "unsigned int max_mapnr" in struct disk_dump_header will
not be used anymore, but still be there for compatibility purpose.
The max_mapnr_64 only exists in strcut kdump_sub_header, and that only
for compressed kdump format, so for ELF format kdump files (non-compressed),
only the max_mapnr is available, so it still may be truncated for addresses
exceed 44bit (above 16TB).
This patch will change the header_version to 6.
The corresponding patch for crash utility can be found here:
http://lists.infradead.org/pipermail/kexec/2013-October/009750.html
This patch doesn't change sadump_header.
Changelog:
v4:
- Do not change max_mapnr_64 in kdump_sub_header in memory for old kernel.
v3:
- Change notes for max_mapnr, start_pfn and end_pfn as obsolete.
- Remove "(32bit)" from debug messages of max_mapnr, start_pfn and end_pfn.
- Set the 32bit start_pfn and end_pfn to UINT_MAX.
- Remove bitmap writting enhancement to another seperate patch.
- Change type of len_bitmap in struct DumpInfo back to unsigned long.
v2:
- Rename max_mapnr in struct kdump_sub_header to max_mapnr_64.
- Change type of max_mapnr_64 from unsigned long to unsigned long long.
In x86 PAE mode on x86_32 kernel, the address may exceeds 44bit limit.
- Add start_pfn_64, end_pfn_64 for struct kdump_sub_header.
- Only print 64bit start_pfn_64, end_pfn_64 and max_mapnr_64
debug messages for disk dump header version >= 6.
- Change type of bitmap_len in struct DumpInfo, from unsigned long to
unsigned long long.
- Enhance bitmap writting function in reassemble_kdump_header().
Prevent bitmap writting failure if the size of bitmap is too large to
fit a sigle write.
v1:
- http://lists.infradead.org/pipermail/kexec/2013-September/009662.html
Signed-off-by: Jingbai Ma <jingbai.ma@hp.com>
Tested-by: Lisa Mitchell <lisa.mitchell@hp.com>
|
| |
|
|
| |
Signed-off-by: Atsushi Kumagai <kumagai-atsushi@mxc.nes.nec.co.jp>
|
| |
|
|
| |
Signed-off-by: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com>
|
| |
|
|
|
|
|
|
| |
Add -l option as one of the command-line options users can specify. If
-l option is specified, then makedumpfile generates dumpfile in
kdump-compressed format with lzo compression by each page.
Signed-off-by: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com>
|
| |
|
|
| |
Signed-off-by: Atsushi Kumagai <kumagai-atsushi@mxc.nes.nec.co.jp>
|
| |
|
|
| |
Signed-off-by: Ken'ichi Ohmichi <oomichi@mxs.nes.nec.co.jp>
|
| |
|
|
| |
Signed-off-by: Ken'ichi Ohmichi <oomichi@mxs.nes.nec.co.jp>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Include the erase information in the makedumpfile generated dumpfile. The
erased information contains only effective lines from the filter
configuration file. This patch supports inclusion of erase information in
kdump compressed, flatten and split dumpfiles. On re-filtering, the erase
information from input dumpfile (if any) will be carried over to resulting
dumpfile.
This patch introduces new members offset_eraseinfo, size_eraseinfo into the
sub header in compressed kdump file.
The erased information would contain *multiple* erased symbol names.
For example, the following is based on your above example.
struct S1 {
int a;
int b;
int c;
};
struct S2 {
struct S1 *mystruct1;
struct S1 *mystruct2;
};
struct S2 mystruct2;
#Filter command
erase mystruct2.mystruct1.a
erase mystruct2.mystruct1.c
erase mystruct2.mystruct2.b
#The erased information
erase mystruct2.mystruct1.a 4
erase mystruct2.mystruct1.c 4
erase mystruct2.mystruct2.b 4
The dumpfile image :
header dump data the erased information
+------------------+-----------------------+-------------------------------+
| ................ | ..................... | erase mystruct2.mystruct1.a 4 |
| offset_eraseinfo | ..................... | erase mystruct2.mystruct1.c 4 |
| size_eraseinfo | ..................... | erase mystruct2.mystruct2.b 4 |
+------------------+-------------------------------------------------------+
: offset_eraseinfo (offset in
dumpfile)
<------ size_eraseinfo ------->
Signed-off-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
---
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In case of s390x's DUMPFILE in the kdump-compressed format, the crash
utility gets the register information using CPU lowcore symbol.
All s390 dumps have the register information in memory on the lowcore
pages.
But in the future s390 could have a dump mechanism that does not store
the register information in the lowcore pages in the memory and will
only keep it on ELF note section.
So DUMPFILE in the kdump-compressed format needs ELF note section for
including the register information, and this patch adds new members
"offset_note" and "size_note" into kdump sub header for representing
the section.
Signed-off-by: Ken'ichi Ohmichi <oomichi@mxs.nes.nec.co.jp>
|
| |
|
|
| |
Signed-off-by: Ken'ichi Ohmichi <oomichi@mxs.nes.nec.co.jp>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
format.
By applying this patch, makedumpfile takes vmcoreinfo data from
/proc/vmcore and stores the data into a dumpfile in the kdump-
compressed fomat. By the next patch, makedumpfile takes the data
from a dumpfile in the kdump-compressed format and uses the data
for re-filtering.
Signed-off-by: Ken'ichi Ohmichi <oomichi@mxs.nes.nec.co.jp>
|
| |
|
|
| |
Signed-off-by: Ken'ichi Ohmichi <oomichi@mxs.nes.nec.co.jp>
|
| |
|
|
|
|
|
|
| |
This patch adds DISKDUMP_HDADER_BLOCKS and KDUMP_SUB_HEADER_BLOCKS
definitions, and they are used instead of values (1).
Signed-off-by: Ken'ichi Ohmichi <oomichi@mxs.nes.nec.co.jp>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A new option "--split":
Split the dump data to multiple dumpfiles in parallel.
If specifying dumpfiles on different storage devices, a device can share
I/O load with other devices and it reduces time for saving the dump data.
The file size of each dumpfile is smaller than the system memory size which
is divided by the number of dumpfiles.
This feature supports only the kdump-compressed format.
Example:
# makedumpfile --split -d 31 -x vmlinux /proc/vmcore dumpfile1 dumpfile2
Signed-off-by: Takao Indoh <tindoh@redhat.com>
Signed-off-by: Ken'ichi Ohmichi <oomichi@mxs.nes.nec.co.jp>
|
| |
|
|
|
|
|
| |
This is just a cleanup patch that removes trailing whitespace from
implementation files.
Signed-off-by: Bernhard Walle <bwalle@suse.de>
|
| |
|
|
|
|
|
|
|
|
|
| |
- The descriptors of the excluded zero-filled pages are changed so that the
crash utility (v4.0-3.22, or later) can distinguish between the excluded
zero-filled pages and other excluded pages. If the crash utility accesses
the excluded zero-filled page, it returns zero-filled page. If accessing
other excluded page, it displays the error message and fails.
- A member "dump_level" is added into struct kdump_sub_header so that the
crash utility (v4.0-3.22, or later) can display the dump_level in case of
kdump-compressed format.
|
| |
|
|
|
|
|
|
| |
'-F' means "Flatten dump data instead of lseek()".
For transporting dumpfile by SSH, there was one problem that lseek()
cannot affect to a remote host. Instead of lseek(), makedumpfile
outputs all dump data with a data header (fileoffset, data size).
This format is named "flattened format".
|
| |
|
