- Merge SLE12 changelog.
- Patches that were never actually applied to Factory:
* makedumpfile-x86_64-xen-vtop.patch (included in 1.6.2)
* makedumpfile-Fix-elf_info-file_size-if-segment-excluded.patch
(included in 1.6.2)
- makedumpfile-Fix-elf_info-file_size-if-segment-excluded.patch:
elf_info: Fix file_size if segment is excluded (bsc#1068925).
- makedumpfile-x86_64-xen-vtop.patch: Fix the use of Xen physical
and machine addresses (bsc#1014136, bsc#1068694).
- makedumpfile-is_cache_page-helper.patch: Add is_cache_page()
helper to check if a page belongs to the cache (bsc#1088354).
- makedumpfile-check-PG_swapbacked.patch: Check PG_swapbacked for
swap cache pages (bsc#1088354).
- makedumpfile-do-not-print-ETA-if-progress-is-0.patch: Do not
print ETA value if current progress is 0 (bsc#1084936).
- Use %license instead of %doc [bsc#1082318]
- makedumpfile-sadump-fix-PTI-enabled-kernels.patch: sadump: Fix a
problem of PTI enabled kernel (bsc#1085826).
- makedumpfile-always-use-bigger-SECTION_MAP_MASK.patch: Always use
bigger SECTION_MAP_MASK (bsc#1066811, bsc#1067703).
- Update to 1.6.3
* Support kernels up to 4.14.8 (bsc#1068864).
* 86_64: handle renamed init_level4_pgt -> init_top_pgt
* Fix SECTION_MAP_MASK for kernel >= v.13
* book3s/ppc64: Lower the max real address to 53 bits for
kernels >= v4.11
* Support symbol __cpu_online_mask
* ppc64: update hash page table geometry
- Drop upstreamed patches:
* makedumpfile-Fix-SECTION_MAP_MASK-for-kernel-v.13.patch
* makedumpfile-handle-renamed-init_level4_pgt-init_top_pgt.patch
* makedumpfile-ppc64-update-hash-page-table-geometry.patch
* makedumpfile-book3s-ppc64-Lower-the-max-real-address-to-53-bits.patch
* makedumpfile-__cpu_online_mask-symbol.patch
* makedumpfile-vtop4_x86_64_pagetable.patch
* makedumpfile-fix-KASLR-for-sadump.patch
* makedumpfile-fix-KASLR-for-sadump-while-kdump.patch
* makedumpfile-support-4.12.patch
- Drop SLE12-specific patches:
* makedumpfile-ppc64-update-hash-page-table-geometry.patch
* makedumpfile-Revert-Clean-up-unused-KERNEL_IMAGE_SIZE.patch
* makedumpfile-Revert-x86_64-kill-some-unused-init.patch
* makedumpfile-Revert-kill-is_vmalloc_addr_x86_64.patch
* makedumpfile-Revert-x86_64-translate-all-VA-to-PA-using-pgt.patch
* makedumpfile-Revert-Calculate-page_offset-from-pt_load.patch
- makedumpfile-__cpu_online_mask-symbol.patch: Support symbol
__cpu_online_mask (FATE#323473, bsc#1070291).
- makedumpfile-vtop4_x86_64_pagetable.patch: Introduce
vtop4_x86_64_pagetable (FATE#323473, bsc#1070291).
- makedumpfile-fix-KASLR-for-sadump.patch: Fix a KASLR problem of
sadump (FATE#323473, bsc#1070291).
- makedumpfile-fix-KASLR-for-sadump-while-kdump.patch: sadump: Fix
a KASLR problem of sadump while kdump is working (FATE#323473,
bsc#1070291).
- makedumpfile-Revert-Clean-up-unused-KERNEL_IMAGE_SIZE.patch:
Revert "Clean up unused KERNEL_IMAGE_SIZE" (bsc#1068925,
bsc#1099121).
- makedumpfile-Revert-x86_64-kill-some-unused-init.patch: Revert
"x86_64: kill some unused initialization" (bsc#1068925,
bsc#1099121).
- makedumpfile-Revert-kill-is_vmalloc_addr_x86_64.patch: Revert
"x86_64: kill is_vmalloc_addr_x86_64()" (bsc#1068925,
bsc#1099121).
- makedumpfile-Revert-x86_64-translate-all-VA-to-PA-using-pgt.patch:
Revert "x86_64: translate all VA to PA using page table values"
(bsc#1068925, bsc#1099121).
- makedumpfile-Revert-Calculate-page_offset-from-pt_load.patch:
Revert "x86_64: Calculate page_offset from pt_load"
(bsc#1068925, bsc#1040469, bsc#1099121).
- makedumpfile-ppc64-update-hash-page-table-geometry.patch:
Kernel commit f6eedbba7a26 ("powerpc/mm/hash: Increase VA range to 128TB")
updated hash page table geometry. A modified version of this commit is
included in SLES12 SP3. Make the corresponding changes in makedumpfile tool
for filtering dump appropriately (bsc#1068485)
- ppc64 Can't convert a virtual address (bsc#1067703)
* Added patches: makedumpfile-ppc64-update-hash-page-table-geometry.patch
makedumpfile-book3s-ppc64-Lower-the-max-real-address-to-53-bits.patch
* Refresh makedumpfile-Fix-SECTION_MAP_MASK-for-kernel-v.13.patch to also
apply to SLE15 (4.12 kernel) due to backport of 2d070eab2e82 (bsc#1067703)
- Handled renaming of init_level4_pgt to init_top_pgt (bsc#1066770).
* Added patch: makedumpfile-handle-renamed-init_level4_pgt-init_top_pgt.patch
- add makedumpfile-Fix-SECTION_MAP_MASK-for-kernel-v.13.patch (bnc#1066811)
- Update to 1.6.2
* Fix the use of Xen physical and machine addresses (bsc#1014136)
* Fix memory leak in get_kcore_dump_loads()
* Support kernels up to 4.11.7
* Consider not page-size aligned phys_end for paddr_to_pfn()
* Add runtime kaslr offset if it exists
- Update to 1.6.1 (FATE#322011).
* Enhance support for aarch64
* Enhance support for ppc64
* Support kernels up to 4.8
- Drop upstreamed patch
* makedumpfile-_count-_refcount-rename.patch
- Merge with updates on SLE12 SP2 (FATE#318012, bsc#992885,
bsc#999869).
- Rename Support-_count-_refcount-rename-in-struct-p.patch to
makedumpfile-_count-_refcount-rename.patch.
- Silence rpmlint errors about devel files in non-devel package;
despite their .c suffix, the provided eppic scripts are intended
for production, not development.
- Build and install the eppic extension.
- makedumpfile-override-libtinfo.patch: Allow to override the tinfo
library used for eppic.
- Update to 1.6.0 (FATE#320955).
* Exclude page structures of non-dumped pages.
- Drop upstreamed patch
* Looking-for-page.compound_order-compound_dtor-.patch
* Skip-examining-compound-tail-pages.patch
- Looking-for-page.compound_order-compound_dtor-.patch:
fix excluding hugepages (kernel 4.4 compatibility)
- Skip-examining-compound-tail-pages.patch
fix excluding compound tail pages (kernel 4.5 compatibility)
- Support-_count-_refcount-rename-in-struct-p.patch:
support 4.7 kernel (page._count renamed to page._refcount)
- Update to 1.5.9
* support for aarch64 (FATE#318444)
* Support kernels up to 4.1
* Enable compressed dump formats for Xen (FATE#316467).
- Drop upstreamed patch
* makedumpfile-add-aarch64.diff
- Use url for source
- Cleanup spec file with spec-cleaner
- Adjust usage of install (-c is ignored)
- makedumpfile-add-aarch64.diff: Add support for aarch64
This patch should be oboslet when switching to 1.5.9
- upgrade to makedumpfile-1.5.8
o Fair I/O workload assignment for --split
o Make incomplete dumpfile readable
o Support kernels up to 3.19
- Switch to dynamic linking: Since most libraries are no longer
available for static link, the remaining space savings are
no longer worth the effort.
- upgrade to makedumpfile-1.5.7
o Show memory usage of the running kernel
o Hugepage filtering
o Support kernels up to 3.16
- Drop following patches, which are now upstream:
o makedumpfile-generic-multi-page-excl.patch
o makedumpfile-remove-overrun-adj.patch
o makedumpfile-fix-free-bitmap_buffer_cyclic.patch
o makedumpfile-isCompoundHead.patch
o makedumpfile-exclude-compound-pages.patch
- makedumpfile-fix-free-bitmap_buffer_cyclic.patch: Fix free
bitmap_buffer_cyclic error.
- makedumpfile-generic-multi-page-excl.patch: Generic handling of
multi-page exclusions (bnc#873232).
- makedumpfile-remove-overrun-adj.patch: Get rid of overrun
adjustments (bnc#873232).
- makedumpfile-isCompoundHead.patch: Add isCompoundHead() macro to
check for compound pages (bnc#873232).
- makedumpfile-exclude-compound-pages.patch: Treat compound pages
as a single entity (bnc#873232).
- upgrade to makedumpfile-1.5.6
o support for Linux 3.13
o include sample eppic scripts
o eppic: Add support for module data structures
o ability to filter Xen Dom0 dumps (bnc#864910, bnc#829646).
- makedumpfile-fix-sprintf-append.patch: now upstream
- makedumpfile-kernel-3.12-supported.patch: now upstream
- enable snappy compression (FATE#315726).
- makedumpfile-fix-sprintf-append.patch: Fix string append in
dump_log_entry() (bnc#865596).
- makedumpfile-kernel-3.12-supported.patch: Mark kernel 3.12 as
supported.
- upgrade to makedumpfile-1.5.5
o Add eppic support for module data structures
o makedumpfile header to show LZO/snappy/zlib
o Add --non-mmap option to disable mmap() manually
o many bugfixes
- Drop makedumpfile-powerpc.patch: now upstream
- Drop makedumpfile-raise-44bit-physaddr-limit.patch: now upstream
- enable ppc64le
- makedumpfile-raise-44bit-physaddr-limit.patch: Fix max_mapnr
issue on system has over 44-bit addressing (bnc#841145).
- makedumpfile-powerpc.patch: fix builf on powerpc
- makedumpfile.spec: enable powerpc32
- upgrade to makedumpfile-1.5.4
o support kernels up to 3.9
o speed improvements with mmap(2)
o bugfixes
- makedumpfile-handle-structured-log_buf: upstream.
- Added support for structured logging added in Linux v3.5. (bnc#801063)
- upgrade to makedumpfile-1.5.1
o Support for Xen4 Dom0 dumps
o Introduce mem_map array logic
- makedumpfile-x86-return-in-nonvoid-function.patch: dropped.
- upgrade to makedumpfile-1.5.0
o implement cyclic mode
o update necessary elfutils version
o tested for kernels up to 3.4.8
- include makedumpfile.conf(5) man page in the package.
- makedumpfile-x86-return-in-nonvoid-function.patch: add a missing
return statement.
- upgrade to makedumpfile-1.4.4.
- activate LZO support
- makedumpfile-supports-3.0.patch: dropped (upstreams).
- README.static: Explain why static linking of some libraries is
preferred.
- license update: GPL-2.0
SPDX format (http://www.spdx.org/licenses)
- Fix build on older distros.
- fix the specfile License tag: since ppc64.c and s390x.c are
licensed under GPL-2.0 only, the whole package cannot be GPL-2.0
or later (bnc#727022).
- upgrade to makedumpfile-1.4.0.
- makedumpfile-supports-3.0.patch: Mark all kernel 3.0 releases
as supported.
- Update to 1.3.6
* Features
o Use TMPDIR environment variable for temporary files
o Add linux-2.6.32 - 2.6.34 support
* Bugfixes
o Fix buffer overflow when writing dh->signature
o handle !SPARSEMEM_EX properly
* Code Cleanup
o Fix 'struct kdump_sub_header' member in IMPLEMENTATION
o fix typo in IMPLEMENTATION
o add maintainers
o add TMPDIR description
o add .gitignore file
o makedumpfile: works on 2.6.32
o the initialization method is operated if it is possible
- makedumpfile-fix-buffer-overflow.diff: dropped (upstream).
- SPARC porting: use ExclusiveArch instead of ExcludeArch
- Fix buffer overflow that prevents build in Factory.
- Update to 1.3.5
* Features
o Store vmcoreinfo data into a dumpfile in the kdump-
compressed format.
o Read vmcoreinfo data from a kdump-compressed dumpfile.
o Support kdump-compressed format for input file.
o Add dump filtering on an x86_64 xen domain-0.
o Save utsname data into disk_dump_header.
o Add linux-2.6.30 - 2.6.31 support.
* Bugfix
o Fix some lacks of free(). (by Minoru Usui)
* Code Cleanup
o Set the calculated value to dh->sub_hdr_size.
o Fix some indents in write_kdump_header().
o Unify the name of kdump_sub_header.
o Fix the typo of DISKDUMP_HEADER_BLOCKS.
o Add the description about re-filtering into manpage.
o Fix the description in manpage and README.
o Add some comments to code for the readability.
o Remove VERSION_LINUX_2_6_XX.
o Add paddr_to_pfn() and pfn_to_paddr() macros.
o Merge the version information of makedumpfile.
o Merge the PT_NOTE methods of both elf64 and elf32.
o Reduce some indents in exclude_unnecessary_pages().
o Remove unnecessary spaces.
o read_disk_dump_header() and read_kdump_sub_header().
o Use DISKDUMP_HDADER_BLOCKS for the readability.
- Update to 1.3.3
o Add --split option
o Add --reassemble option
o Allow specifying multiple dump levels with "-d"
o Code cleanups
o Bugfixes
- do not use static zlib
- Update to 1.3.2
o Add --dump-dmesg option (Dumping dmesg buffer from /proc/vmcore
to a file)
o Fix a typo in README file.
o Code Cleanup
- Exclude 's390' and 's390x' from being built, too.
- Exclude 'ppc' from being built.
- Drop makedumpfile-64bit-kernel-on-32bit-userland.diff: Rejected
upstream. We need to use a 64 bit makedumpfile for 64 bit
kernels on PPC64 (bnc#447432).
- makedumpfile-64bit-kernel-on-32bit-userland.diff: Change return
type of read_vmcoreinfo_symbol() to 'unsigned long long'.
- Fix parsing of 64 bit VMCOREINFO on 32 bit platforms. For SUSE,
this only affects openSUSE on PPC because SLES PPC has a 64 bit
userland now. (bnc#447432).
- Update to 1.3.1
o Support linux-2.6.26 sparsemem on i386.
o Support linux-2.6.27 and linux-2.6.28.
- Remove makedumpfile-2.6.27.diff: Mainline.
- Don't require a C++ compiler. That was from the time where the
library was built within that package and statically linked in.
- Don't print the "The kernel version is not supported" message
for 2.6.27 (bnc#440225).
- Update to 1.3.0
* features
o Support i386 CONFIG_VMSPLIT_1G kernel.
o Add compiling options for calling lseek64() surely.
o Use gzip for manpage compression.
* bug fixes
o Fix the buffer handling problem for reading struct page.
o Fix readmem to read each page.
o Don't write undefined values to disk.
o Fix memory leaks.
o Fix/Add error messages for debugging.
o Fix the type of file_offset.
o Fix the error handling of kvtop_xen().
o Fix the comment for some values in page.flags.
* Code Cleanup
o Cleanup: Separate xen's paddr_to_offset from linux's one.
o Cleanup: Use the existing macros in kvtop_xen_x86().
o Cleanup: Delete unnecessary description.
o Update copyright date.
- update to 1.2.9
* features
o Support ia64 discontigmem kernels of linux-2.6.20+.
o Support x86_64 linux-2.6.27 kernel.
o Add makedumpfile's spec file.
o Add "--vtop" option for debugging.
o Add the debugging message for ia64 pgtable.
o Get information from vmcoreinfo of /proc/vmcore even if
-x/-i option.
* bug fixes
o Add free() for error handling.
o Fix NOT_FOUND value of SIZE(nodemask_t).
* code cleanup
o Use the terminal sizes of 80 for the IMPLEMENTATION file.
o Use static allocation instead of malloc().
o Add get_num_dumpable() function.
o Remove the unused pointer.
o Remove the unused code.
o Make the ELF methods simple.
o Merge the same code in write_elf_pages().
o Add "void" to some arguments.
o Separate the dependency code to machine and linux version.
o Cleanup vaddr_to_paddr() function.
- Remove new compression command for the manual page in the
Makefile because we use gzip, not zip, and Autobuild does the
compression automatically.
- update to 1.2.8
* Add the IMPLEMENTATION file for the kdump-compressed format.
- update to 1.2.7
* features
o Add xen-3.1.2 support
o handle x86_64 xen code/data relocation
o Add linux-2.6.26 support
o Add the progress bar for creating dump bitmap
o Shrink the time for creating dump bitmap
* bug fixes
o Fix vmcoreinfo generation on 32bit hosts for 64bit kernels
o Fix compile warning
o Fix illegal buffer access
o Improve error handing when SIZE(nodemask_t) is undefined
* code cleanup
o Remove trailing whitespace
o Remove unnecessary description in manpage
o Cleanup machine dependent method
o Cleanup open_dump_file() method
- drop makedumpfile-fix-64bit-on-32bit: merged mainline
- update to 1.2.6
o Extract vmcoreinfo from /proc/vmcore for Xen (by Itsuro ODA)
o Add linux-2.6.25 support
o Add '--help' option
o Fix the error code of Xen extracting feature (by Akio Takebe)
o Fix the ia64 problem that invalid data is copied around
overlapping segments.
o make create_dump_bitmap() simple
- update to 1.2.5
o follow the changes of page flags.
To follow the changes of page flags in linux kernel, a new
makedumpfile can get these values from the vmcoreinfo data.
o Code Cleanup
o fix error return values of vtop().
- update to 1.2.4
o Add linux-2.6.24 support.
o Add xen kernel support to the STDOUT feature (-F option).
o Fix the method looking for .debug_info section in
get_debug_info().
o Fix the problem that -d1 option cannot be specified without
-x/-i option.
- Fix generation of vmcoreinfo for ELF64 kernels on 32bit machines
- updated to 1.2.3
o Add the filtering feature for free_pages of linux-2.6.24.
- updated to 1.2.2
o Fix the problem that vmcoreinfo cannot be extracted from ELF32
/proc/vmcore.
- adding libelf0 libelf1 to BuildRequires to make the resolver
of the BuildService happy for <= 10.2
- updated to 1.2.1
o Add virtual memmap support for x86_64 sparsemem kernel.
o Add the message_level (16) for printing the report message.
o Add linux-2.6.23 support because makedumpfile was tested on
linux-2.6.23.
o Change the common/debug message output to STDOUT from
STDERR if not specifying '-F' option
o Some cleanups
- updated to 1.2.0
o extract vmcoreinfo from /proc/vmcore.
o linux-2.6.23-rc4-mm1 support.
o Add the "-d 1" option (skip zero-filled pages) for Xen.
o Add the feature which gets enum number.
o Some Cleanups
- updated to 1.1.6
o linux-2.6.22 support.
o Change the check method of free_pages for linux-2.6.21 or
later.
o Some cleanups
- updated to 1.1.5
o new feature that allows extracting the part of xen hypervisor
and domain-0 from /proc/vmcore.
o new option '-f' to overwrite existing dumpfile
o new option '--message-level "message-level"'
o fix the problem that makedumpfile complains about overlapping
memory segments.
o Rename makedumpfile's config file to "vmcoreinfo file"
o Change the method for getting the kernel's OSRELEASE if
generating a vmcoreinfo file.
o Some cleanups
- updated to 1.1.4
o Add ia64 DISCONTIGMEM support.
o Add ia64_vtop.
o Add __FUNCTION__ to ERRMSG.
o Cleanup README.
o Fix DISCONTIGMEM bitmap.
o Fix ELF output with overlapping sections.
- fix build
- link dynamically against libc to decrease binary size (initrd
has already libc included) but leave the other libraries linked
statically
- use external elfutils packages.
- updated to version 1.1.3:
* 2.6.21 kernel support
* fix handling of excluded zero-filled pages for crash
* crash can display dump_level
- add zlib-devel BuildRequires
- updated to version 1.1.2:
* don't exit even for the unsupported kernel version
* minor fixes
* documents, man page
- updated to version 1.1.1:
* x86-64 DISCONTIGMEM support
- updated to version 1.1.0:
* support SSH transfer
- updated to version 1.0.9:
* fix the calculation of page size.
- remove codes with problematic license from elfutils-0.124
(#232894)
- added COPYING file (from elfutils)
- fixed optimization flags to use $RPM_OPT_FLAGS
- fix a minor compile error with C99 standard.
- initial version: 1.0.8
- build statically with elfutils-1.2.4 since elfutils conflicts
with the existing libelf.