Message ID | 20190514090100.5795-1-po-hsu.lin@canonical.com |
---|---|
Headers | show |
Series | Remove Power9 copy_unaligned test | expand |
On 5/14/19 2:00 AM, Po-Hsu Lin wrote: > BugLink: https://bugs.launchpad.net/bugs/1813118 > > == Justification == > The copy_unaligned / paste_unaligned / paste_last_unaligned test in > powerpc/alignment will fail on a Power9 node as which implements ISA > 3.0B, but the test was designed base on ISA 3.0. > > From the commit message: > This is a test of the ISA 3.0 "copy" instruction. That instruction has > an L field, which if set to 1 specifies that "the instruction > identifies the beginning of a move group" (pp 858). That's also > referred to as "copy first" vs "copy". > > In ISA 3.0B the copy instruction does not have an L field, and the > corresponding bit in the instruction must be set to 1. > > This test is generating a "copy" instruction, not a "copy first", and > so on Power9 (which implements 3.0B), this results in an illegal > instruction. > > So just drop the test entirely. We still have copy_first_unaligned to > test the "copy first" behaviour. > > == Fix == > * 83039f22 (selftests/powerpc: Remove Power9 copy_unaligned test) > > Patch can be cherry-picked into C, for Bionic it needs patches for bug > 1828935 to land first (8d191587) to make it a clean cherry-pick. > > It's already landed in D and onward. > > == Test == > Tested on a Power9 server with Cosmic, the broken tests will be removed > correctly. > > == Regression Potential == > Low. > Code change limited to testing tools and just removing broken tests. > We still have copy_first_unaligned to test the "copy first" behaviour. > > > Michael Ellerman (1): > selftests/powerpc: Remove Power9 copy_unaligned test > > .../testing/selftests/powerpc/alignment/.gitignore | 1 - > tools/testing/selftests/powerpc/alignment/Makefile | 2 +- > .../selftests/powerpc/alignment/copy_unaligned.c | 41 ---------------------- > 3 files changed, 1 insertion(+), 43 deletions(-) > delete mode 100644 tools/testing/selftests/powerpc/alignment/copy_unaligned.c > Removes broken test and retains same test coverage. Acked-by: Connor Kuehl <connor.kuehl@canonical.com>
On 5/14/19 11:00 AM, Po-Hsu Lin wrote: > BugLink: https://bugs.launchpad.net/bugs/1813118 > > == Justification == > The copy_unaligned / paste_unaligned / paste_last_unaligned test in > powerpc/alignment will fail on a Power9 node as which implements ISA > 3.0B, but the test was designed base on ISA 3.0. > > From the commit message: > This is a test of the ISA 3.0 "copy" instruction. That instruction has > an L field, which if set to 1 specifies that "the instruction > identifies the beginning of a move group" (pp 858). That's also > referred to as "copy first" vs "copy". > > In ISA 3.0B the copy instruction does not have an L field, and the > corresponding bit in the instruction must be set to 1. > > This test is generating a "copy" instruction, not a "copy first", and > so on Power9 (which implements 3.0B), this results in an illegal > instruction. > > So just drop the test entirely. We still have copy_first_unaligned to > test the "copy first" behaviour. > > == Fix == > * 83039f22 (selftests/powerpc: Remove Power9 copy_unaligned test) > > Patch can be cherry-picked into C, for Bionic it needs patches for bug > 1828935 to land first (8d191587) to make it a clean cherry-pick. > > It's already landed in D and onward. > > == Test == > Tested on a Power9 server with Cosmic, the broken tests will be removed > correctly. > > == Regression Potential == > Low. > Code change limited to testing tools and just removing broken tests. > We still have copy_first_unaligned to test the "copy first" behaviour. > > > Michael Ellerman (1): > selftests/powerpc: Remove Power9 copy_unaligned test > > .../testing/selftests/powerpc/alignment/.gitignore | 1 - > tools/testing/selftests/powerpc/alignment/Makefile | 2 +- > .../selftests/powerpc/alignment/copy_unaligned.c | 41 ---------------------- > 3 files changed, 1 insertion(+), 43 deletions(-) > delete mode 100644 tools/testing/selftests/powerpc/alignment/copy_unaligned.c > Acked-by: Kleber Sacilotto de Souza <kleber.souza@canonical.com>
On 5/14/19 11:00 AM, Po-Hsu Lin wrote: > BugLink: https://bugs.launchpad.net/bugs/1813118 > > == Justification == > The copy_unaligned / paste_unaligned / paste_last_unaligned test in > powerpc/alignment will fail on a Power9 node as which implements ISA > 3.0B, but the test was designed base on ISA 3.0. > > From the commit message: > This is a test of the ISA 3.0 "copy" instruction. That instruction has > an L field, which if set to 1 specifies that "the instruction > identifies the beginning of a move group" (pp 858). That's also > referred to as "copy first" vs "copy". > > In ISA 3.0B the copy instruction does not have an L field, and the > corresponding bit in the instruction must be set to 1. > > This test is generating a "copy" instruction, not a "copy first", and > so on Power9 (which implements 3.0B), this results in an illegal > instruction. > > So just drop the test entirely. We still have copy_first_unaligned to > test the "copy first" behaviour. > > == Fix == > * 83039f22 (selftests/powerpc: Remove Power9 copy_unaligned test) > > Patch can be cherry-picked into C, for Bionic it needs patches for bug > 1828935 to land first (8d191587) to make it a clean cherry-pick. > > It's already landed in D and onward. > > == Test == > Tested on a Power9 server with Cosmic, the broken tests will be removed > correctly. > > == Regression Potential == > Low. > Code change limited to testing tools and just removing broken tests. > We still have copy_first_unaligned to test the "copy first" behaviour. > > > Michael Ellerman (1): > selftests/powerpc: Remove Power9 copy_unaligned test > > .../testing/selftests/powerpc/alignment/.gitignore | 1 - > tools/testing/selftests/powerpc/alignment/Makefile | 2 +- > .../selftests/powerpc/alignment/copy_unaligned.c | 41 ---------------------- > 3 files changed, 1 insertion(+), 43 deletions(-) > delete mode 100644 tools/testing/selftests/powerpc/alignment/copy_unaligned.c > Applied to {bionic,cosmic}/master-next branch. Thanks, Kleber