Back to bug 1391217

Who When What Removed Added
RHEL Program Management 2016-11-02 19:50:34 UTC Keywords Rebase
Mark Wielaard 2016-11-03 09:51:25 UTC Blocks 1390370
Blocks 1385006
Hanns-Joachim Uhl 2016-11-03 10:10:18 UTC CC hannsj_uhl
Joshua Miller 2017-01-05 14:18:22 UTC CC jomiller
Blocks 1371219
Martin Cermak 2017-01-11 20:39:09 UTC CC mcermak
Brian Gollaher 2017-01-11 20:43:30 UTC CC briang
Matt Newsome 2017-01-16 22:11:39 UTC Status NEW ASSIGNED
CC mnewsome
Hanns-Joachim Uhl 2017-01-27 16:05:23 UTC Target Release --- 7.4
Hardware Unspecified All
OS Unspecified Linux
Mark Wielaard 2017-03-01 22:58:24 UTC Status ASSIGNED MODIFIED
Fixed In Version valgrind-3.12.0-7.el7
errata-xmlrpc 2017-03-01 23:06:18 UTC Status MODIFIED ON_QA
Lenka Špačková 2017-05-17 10:31:45 UTC Docs Contact vslavik
Doc Type If docs needed, set a value Rebase: Bug Fixes and Enhancements
Mark Wielaard 2017-05-17 11:39:39 UTC Doc Text Rebase package(s) to version: 3.12.0 (from 3.11.0 in RHEL 7.3).

Highlights, important fixes, or notable enhancements:

Architecture specific changes:

- Support for power9 (ISA 3.0) has been added.
- Partial support for AMD FMA4 instructions has been added.
- aarch64: Support for v8 crypto and CRC instructions has been added.

Generic tool changes:

- memcheck has a new flag --ignore-range-below-sp to ignore memory accesses below the stack pointer. This is a generic replacement for the now deprecated --workaround-gcc296-bugs=yes (which is equivalent to --ignore-range-below-sp=1024-1)

- The --gen-suppressions=yes option now generates suppressions containing up to --num-callers frames. And the maximum number of callers in a suppression entry is now equal to the maximum size for --num-callers (500).

Performance improvements:

- The cost of instrumenting code blocks for the most common use case (memcheck on x86_64) has been reduced by 10%-15%.

- Improved performance for programs that do a lot of discarding of instruction address ranges of 8KB or less.
Vladimír Slávik 2017-05-19 16:29:03 UTC CC mjw
Doc Text Rebase package(s) to version: 3.12.0 (from 3.11.0 in RHEL 7.3).

Highlights, important fixes, or notable enhancements:

Architecture specific changes:

- Support for power9 (ISA 3.0) has been added.
- Partial support for AMD FMA4 instructions has been added.
- aarch64: Support for v8 crypto and CRC instructions has been added.

Generic tool changes:

- memcheck has a new flag --ignore-range-below-sp to ignore memory accesses below the stack pointer. This is a generic replacement for the now deprecated --workaround-gcc296-bugs=yes (which is equivalent to --ignore-range-below-sp=1024-1)

- The --gen-suppressions=yes option now generates suppressions containing up to --num-callers frames. And the maximum number of callers in a suppression entry is now equal to the maximum size for --num-callers (500).

Performance improvements:

- The cost of instrumenting code blocks for the most common use case (memcheck on x86_64) has been reduced by 10%-15%.

- Improved performance for programs that do a lot of discarding of instruction address ranges of 8KB or less.
_valgrind_ rebased to version 3.12

The _valgrind_ package has been upgraded to upstream version 3.12, which provides a number of bug fixes and enhancements over the previous version. Notable changes include:

* A new option "--ignore-range-below-sp" has been added to the memcheck tool to ignore memory accesses below the stack pointer. This is a generic replacement for the now deprecated option "--workaround-gcc296-bugs=yes".
* The "--gen-suppressions=yes" option now generates suppressions containing amount of frames up to the value given by the "--num-callers" option. The maximum number of callers in a suppression entry is now equal to the maximum size for value given by the "--num-callers" option.
* The cost of instrumenting code blocks for the most common use case, the *memcheck* tool on the AMD64 and Intel 64 architectures, has been reduced.
* Performance has been improved for debugging programs which discard a lot of instruction address ranges of 8KB or less.
* Support for IBM Power 9 (ISA 3.0) architecture has been added.
* Partial support for AMD FMA4 instructions has been added.
* Support for cryptographic and CRC instructions on the 64-bit ARM architecture version 8 has been added.
Flags needinfo?(mjw)
Mark Wielaard 2017-05-19 17:23:38 UTC Flags needinfo?(mjw)
Vladimír Slávik 2017-05-19 17:32:36 UTC Doc Text _valgrind_ rebased to version 3.12

The _valgrind_ package has been upgraded to upstream version 3.12, which provides a number of bug fixes and enhancements over the previous version. Notable changes include:

* A new option "--ignore-range-below-sp" has been added to the memcheck tool to ignore memory accesses below the stack pointer. This is a generic replacement for the now deprecated option "--workaround-gcc296-bugs=yes".
* The "--gen-suppressions=yes" option now generates suppressions containing amount of frames up to the value given by the "--num-callers" option. The maximum number of callers in a suppression entry is now equal to the maximum size for value given by the "--num-callers" option.
* The cost of instrumenting code blocks for the most common use case, the *memcheck* tool on the AMD64 and Intel 64 architectures, has been reduced.
* Performance has been improved for debugging programs which discard a lot of instruction address ranges of 8KB or less.
* Support for IBM Power 9 (ISA 3.0) architecture has been added.
* Partial support for AMD FMA4 instructions has been added.
* Support for cryptographic and CRC instructions on the 64-bit ARM architecture version 8 has been added.
_valgrind_ rebased to version 3.12

The _valgrind_ package has been upgraded to upstream version 3.12, which provides a number of bug fixes and enhancements over the previous version. Notable changes include:

* A new option "--ignore-range-below-sp" has been added to the memcheck tool to ignore memory accesses below the stack pointer. This is a generic replacement for the now deprecated option "--workaround-gcc296-bugs=yes".
* The maximum number of callers in a suppression entry generated by the "--gen-suppressions=yes" option is now equal to the value given by the "--num-callers" option.
* The cost of instrumenting code blocks for the most common use case, the *memcheck* tool on the AMD64 and Intel 64 architectures, has been reduced.
* Performance has been improved for debugging programs which discard a lot of instruction address ranges of 8KB or less.
* Support for IBM Power 9 (ISA 3.0) architecture has been added.
* Partial support for AMD FMA4 instructions has been added.
* Support for cryptographic and CRC instructions on the 64-bit ARM architecture version 8 has been added.
Ondrej Hudlicky 2017-05-31 15:20:54 UTC CC ohudlick
Miloš Prchlík 2017-06-01 11:20:32 UTC Status ON_QA VERIFIED
CC mprchlik
QA Contact qe-baseos-tools mprchlik
Matt Newsome 2017-06-27 19:27:20 UTC Blocks 1465627
errata-xmlrpc 2017-08-01 22:07:18 UTC Status VERIFIED CLOSED
Resolution --- ERRATA
Last Closed 2017-08-01 18:07:18 UTC

Back to bug 1391217