The va_128TBswitch selftest is designed and implemented for PowerPC and x86 architectures which support a 128TB switch, up to 256TB of virtual address space and hugepage sizes of 16MB and 2MB respectively. Arm64 platforms on the other hand support a 256Tb switch, up to 4PB of virtual address space and a default hugepage size of 512MB when 64k pagesize is enabled.
These architectural differences require introducing support for arm64 platforms, after which a more generic naming convention is suggested. The in code comments are amended to provide a more platform independent explanation of the working of the code and nr_hugepages are configured as required. Finally, the file running the testcase is modified in order to prevent skipping of hugetlb testcases of va_high_addr_switch.
This series has been tested on 6.3.0-rc3 kernel, both on arm64 and x86 platforms.
Cc: Andrew Morton akpm@linux-foundation.org Cc: Aneesh Kumar K.V aneesh.kumar@linux.ibm.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Cc: Shuah Khan shuah@kernel.org Cc: linux-mm@kvack.org Cc: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Chaitanya S Prakash (5): selftests/mm: Add support for arm64 platform on va switch selftests/mm: Rename va_128TBswitch to va_high_addr_switch selftests/mm: Add platform independent in code comments selftests/mm: Configure nr_hugepages for arm64 selftests/mm: Run hugetlb testcases of va switch
tools/testing/selftests/mm/Makefile | 4 +- tools/testing/selftests/mm/run_vmtests.sh | 12 +++++- ...va_128TBswitch.c => va_high_addr_switch.c} | 41 +++++++++++++++---- ..._128TBswitch.sh => va_high_addr_switch.sh} | 6 ++- 4 files changed, 49 insertions(+), 14 deletions(-) rename tools/testing/selftests/mm/{va_128TBswitch.c => va_high_addr_switch.c} (86%) rename tools/testing/selftests/mm/{va_128TBswitch.sh => va_high_addr_switch.sh} (89%)
Arm64 platforms have the ability to support 64kb pagesize, 512MB default hugepage size and up to 4PB of virtual address space. The address switch occurs at 256TB as opposed to 128TB. Hence, the necessary support has been added.
Cc: Andrew Morton akpm@linux-foundation.org Cc: Aneesh Kumar K.V aneesh.kumar@linux.ibm.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Cc: Shuah Khan shuah@kernel.org Cc: linux-mm@kvack.org Cc: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org Signed-off-by: Chaitanya S Prakash chaitanyas.prakash@arm.com --- tools/testing/selftests/mm/va_128TBswitch.c | 26 +++++++++++++++++++-- 1 file changed, 24 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/mm/va_128TBswitch.c b/tools/testing/selftests/mm/va_128TBswitch.c index 1d2068989883..6679213effed 100644 --- a/tools/testing/selftests/mm/va_128TBswitch.c +++ b/tools/testing/selftests/mm/va_128TBswitch.c @@ -17,6 +17,13 @@ * This will work with 16M and 2M hugepage size */ #define HUGETLB_SIZE (16 << 20) +#elif __aarch64__ +/* + * The default hugepage size for 64k base pagesize + * is 512MB. + */ +#define PAGE_SIZE (64 << 10) +#define HUGETLB_SIZE (512 << 20) #else #define PAGE_SIZE (4 << 10) #define HUGETLB_SIZE (2 << 20) @@ -26,9 +33,22 @@ * >= 128TB is the hint addr value we used to select * large address space. */ -#define ADDR_SWITCH_HINT (1UL << 47) + +#define ADDR_MARK_128TB (1UL << 47) +#define ADDR_MARK_256TB (1UL << 48) + +#define HIGH_ADDR_128TB ((void *) (1UL << 48)) +#define HIGH_ADDR_256TB ((void *) (1UL << 49)) + #define LOW_ADDR ((void *) (1UL << 30)) -#define HIGH_ADDR ((void *) (1UL << 48)) + +#ifdef __aarch64__ +#define ADDR_SWITCH_HINT ADDR_MARK_256TB +#define HIGH_ADDR HIGH_ADDR_256TB +#else +#define ADDR_SWITCH_HINT ADDR_MARK_128TB +#define HIGH_ADDR HIGH_ADDR_128TB +#endif
struct testcase { void *addr; @@ -270,6 +290,8 @@ static int supported_arch(void) return 1; #elif defined(__x86_64__) return 1; +#elif defined(__aarch64__) + return 1; #else return 0; #endif
As the initial selftest only took into consideration PowperPC and x86 architectures, on adding support for arm64, a platform independent naming convention is chosen.
Cc: Andrew Morton akpm@linux-foundation.org Cc: Aneesh Kumar K.V aneesh.kumar@linux.ibm.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Cc: Shuah Khan shuah@kernel.org Cc: linux-mm@kvack.org Cc: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org Signed-off-by: Chaitanya S Prakash chaitanyas.prakash@arm.com --- tools/testing/selftests/mm/Makefile | 4 ++-- tools/testing/selftests/mm/run_vmtests.sh | 2 +- .../selftests/mm/{va_128TBswitch.c => va_high_addr_switch.c} | 0 .../mm/{va_128TBswitch.sh => va_high_addr_switch.sh} | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) rename tools/testing/selftests/mm/{va_128TBswitch.c => va_high_addr_switch.c} (100%) rename tools/testing/selftests/mm/{va_128TBswitch.sh => va_high_addr_switch.sh} (98%)
diff --git a/tools/testing/selftests/mm/Makefile b/tools/testing/selftests/mm/Makefile index c31d952cff68..1246552d4b86 100644 --- a/tools/testing/selftests/mm/Makefile +++ b/tools/testing/selftests/mm/Makefile @@ -91,7 +91,7 @@ endif endif
ifneq (,$(filter $(MACHINE),arm64 ia64 mips64 parisc64 ppc64 riscv64 s390x sh64 sparc64 x86_64)) -TEST_GEN_FILES += va_128TBswitch +TEST_GEN_FILES += va_high_addr_switch TEST_GEN_FILES += virtual_address_range TEST_GEN_FILES += write_to_hugetlbfs endif @@ -100,7 +100,7 @@ TEST_PROGS := run_vmtests.sh
TEST_FILES := test_vmalloc.sh TEST_FILES += test_hmm.sh -TEST_FILES += va_128TBswitch.sh +TEST_FILES += va_high_addr_switch.sh
include ../lib.mk
diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh index 8984e0bb58c7..eb9823a6959a 100644 --- a/tools/testing/selftests/mm/run_vmtests.sh +++ b/tools/testing/selftests/mm/run_vmtests.sh @@ -223,7 +223,7 @@ if [ $VADDR64 -ne 0 ]; then CATEGORY="hugevm" run_test ./virtual_address_range
# virtual address 128TB switch test - CATEGORY="hugevm" run_test ./va_128TBswitch.sh + CATEGORY="hugevm" run_test ./va_high_addr_switch.sh fi # VADDR64
# vmalloc stability smoke test diff --git a/tools/testing/selftests/mm/va_128TBswitch.c b/tools/testing/selftests/mm/va_high_addr_switch.c similarity index 100% rename from tools/testing/selftests/mm/va_128TBswitch.c rename to tools/testing/selftests/mm/va_high_addr_switch.c diff --git a/tools/testing/selftests/mm/va_128TBswitch.sh b/tools/testing/selftests/mm/va_high_addr_switch.sh similarity index 98% rename from tools/testing/selftests/mm/va_128TBswitch.sh rename to tools/testing/selftests/mm/va_high_addr_switch.sh index 41580751dc51..3056788a27ac 100644 --- a/tools/testing/selftests/mm/va_128TBswitch.sh +++ b/tools/testing/selftests/mm/va_high_addr_switch.sh @@ -51,4 +51,4 @@ check_test_requirements() }
check_test_requirements -./va_128TBswitch +./va_high_addr_switch
The in code comments for the selftest were made on the basis of 128TB switch, an architecture feature specific to PowerPc and x86 platforms. Keeping in mind the support added for arm64 platforms which implements a 256TB switch, a more generic explanation has been provided.
Cc: Andrew Morton akpm@linux-foundation.org Cc: Aneesh Kumar K.V aneesh.kumar@linux.ibm.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Cc: Shuah Khan shuah@kernel.org Cc: linux-mm@kvack.org Cc: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org Signed-off-by: Chaitanya S Prakash chaitanyas.prakash@arm.com --- tools/testing/selftests/mm/va_high_addr_switch.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/tools/testing/selftests/mm/va_high_addr_switch.c b/tools/testing/selftests/mm/va_high_addr_switch.c index 6679213effed..7cfaf4a74c57 100644 --- a/tools/testing/selftests/mm/va_high_addr_switch.c +++ b/tools/testing/selftests/mm/va_high_addr_switch.c @@ -30,8 +30,8 @@ #endif
/* - * >= 128TB is the hint addr value we used to select - * large address space. + * The hint addr value is used to allocate addresses + * beyond the high address switch boundary. */
#define ADDR_MARK_128TB (1UL << 47) @@ -73,9 +73,10 @@ static struct testcase testcases[] = { }, { /* - * We should never allocate at the requested address or above it - * The len cross the 128TB boundary. Without MAP_FIXED - * we will always search in the lower address space. + * Unless MAP_FIXED is specified, allocation based on hint + * addr is never at requested address or above it, which is + * beyond high address switch boundary in this case. Instead, + * a suitable allocation is found in lower address space. */ .addr = ((void *)(ADDR_SWITCH_HINT - PAGE_SIZE)), .size = 2 * PAGE_SIZE, @@ -85,8 +86,8 @@ static struct testcase testcases[] = { }, { /* - * Exact mapping at 128TB, the area is free we should get that - * even without MAP_FIXED. + * Exact mapping at high address switch boundary, should + * be obtained even without MAP_FIXED as area is free. */ .addr = ((void *)(ADDR_SWITCH_HINT)), .size = PAGE_SIZE,
Arm64 has a default hugepage size of 512MB when CONFIG_ARM64_64K_PAGES=y is enabled. While testing on arm64 platforms having up to 4PB of virtual address space, a minimum of 6 hugepages were required for all test cases to pass. Support for this requirement has been added.
Cc: Andrew Morton akpm@linux-foundation.org Cc: Aneesh Kumar K.V aneesh.kumar@linux.ibm.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Cc: Shuah Khan shuah@kernel.org Cc: linux-mm@kvack.org Cc: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org Signed-off-by: Chaitanya S Prakash chaitanyas.prakash@arm.com --- tools/testing/selftests/mm/run_vmtests.sh | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)
diff --git a/tools/testing/selftests/mm/run_vmtests.sh b/tools/testing/selftests/mm/run_vmtests.sh index eb9823a6959a..7616b3e4a0f0 100644 --- a/tools/testing/selftests/mm/run_vmtests.sh +++ b/tools/testing/selftests/mm/run_vmtests.sh @@ -222,8 +222,16 @@ CATEGORY="hugetlb" run_test ./thuge-gen if [ $VADDR64 -ne 0 ]; then CATEGORY="hugevm" run_test ./virtual_address_range
- # virtual address 128TB switch test + # va high address boundary switch test + ARCH_ARM64="arm64" + prev_nr_hugepages=$(cat /proc/sys/vm/nr_hugepages) + if [ "$ARCH" == "$ARCH_ARM64" ]; then + echo 6 > /proc/sys/vm/nr_hugepages + fi CATEGORY="hugevm" run_test ./va_high_addr_switch.sh + if [ "$ARCH" == "$ARCH_ARM64" ]; then + echo $prev_nr_hugepages > /proc/sys/vm/nr_hugepages + fi fi # VADDR64
# vmalloc stability smoke test
The va_high_addr_switch selftest is used to test mmap across 128TB boundary. It divides the selftest cases into two main categories on the basis of size. One set is used to create mappings that are multiples of PAGE_SIZE while the other creates mappings that are multiples of HUGETLB_SIZE.
In order to run the hugetlb testcases the binary must be appended with "--run-hugetlb" but the file that used to run the test only invokes the binary, thereby completely skipping the hugetlb testcases. Hence, the required statement has been added.
Cc: Andrew Morton akpm@linux-foundation.org Cc: Aneesh Kumar K.V aneesh.kumar@linux.ibm.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Cc: Shuah Khan shuah@kernel.org Cc: linux-mm@kvack.org Cc: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org Signed-off-by: Chaitanya S Prakash chaitanyas.prakash@arm.com --- tools/testing/selftests/mm/va_high_addr_switch.sh | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/tools/testing/selftests/mm/va_high_addr_switch.sh b/tools/testing/selftests/mm/va_high_addr_switch.sh index 3056788a27ac..45cae7cab27e 100644 --- a/tools/testing/selftests/mm/va_high_addr_switch.sh +++ b/tools/testing/selftests/mm/va_high_addr_switch.sh @@ -52,3 +52,7 @@ check_test_requirements()
check_test_requirements ./va_high_addr_switch + +# In order to run hugetlb testcases, "--run-hugetlb" must be appended +# to the binary. +./va_high_addr_switch --run-hugetlb
On Thu, Mar 23, 2023 at 04:22:38PM +0530, Chaitanya S Prakash wrote:
The va_128TBswitch selftest is designed and implemented for PowerPC and x86 architectures which support a 128TB switch, up to 256TB of virtual address space and hugepage sizes of 16MB and 2MB respectively. Arm64 platforms on the other hand support a 256Tb switch, up to 4PB of virtual address space and a default hugepage size of 512MB when 64k pagesize is enabled.
These architectural differences require introducing support for arm64 platforms, after which a more generic naming convention is suggested. The in code comments are amended to provide a more platform independent explanation of the working of the code and nr_hugepages are configured as required. Finally, the file running the testcase is modified in order to prevent skipping of hugetlb testcases of va_high_addr_switch.
This series has been tested on 6.3.0-rc3 kernel, both on arm64 and x86 platforms.
Cc: Andrew Morton akpm@linux-foundation.org Cc: Aneesh Kumar K.V aneesh.kumar@linux.ibm.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Cc: Shuah Khan shuah@kernel.org Cc: linux-mm@kvack.org Cc: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Chaitanya S Prakash (5): selftests/mm: Add support for arm64 platform on va switch selftests/mm: Rename va_128TBswitch to va_high_addr_switch selftests/mm: Add platform independent in code comments selftests/mm: Configure nr_hugepages for arm64 selftests/mm: Run hugetlb testcases of va switch
tools/testing/selftests/mm/Makefile | 4 +- tools/testing/selftests/mm/run_vmtests.sh | 12 +++++- ...va_128TBswitch.c => va_high_addr_switch.c} | 41 +++++++++++++++---- ..._128TBswitch.sh => va_high_addr_switch.sh} | 6 ++- 4 files changed, 49 insertions(+), 14 deletions(-) rename tools/testing/selftests/mm/{va_128TBswitch.c => va_high_addr_switch.c} (86%) rename tools/testing/selftests/mm/{va_128TBswitch.sh => va_high_addr_switch.sh} (89%)
The patchset looks sane to me, but I have question: why arm64 has switch on 256TB. The reason we have the switch is to keep system backward compatible.
Maybe it is better to make arm64 switch also on 128TB to make it compatible across architectures?
On Thu, Mar 23, 2023 at 02:14:36PM +0300, Kirill A. Shutemov wrote:
On Thu, Mar 23, 2023 at 04:22:38PM +0530, Chaitanya S Prakash wrote:
The va_128TBswitch selftest is designed and implemented for PowerPC and x86 architectures which support a 128TB switch, up to 256TB of virtual address space and hugepage sizes of 16MB and 2MB respectively. Arm64 platforms on the other hand support a 256Tb switch, up to 4PB of virtual address space and a default hugepage size of 512MB when 64k pagesize is enabled.
These architectural differences require introducing support for arm64 platforms, after which a more generic naming convention is suggested. The in code comments are amended to provide a more platform independent explanation of the working of the code and nr_hugepages are configured as required. Finally, the file running the testcase is modified in order to prevent skipping of hugetlb testcases of va_high_addr_switch.
This series has been tested on 6.3.0-rc3 kernel, both on arm64 and x86 platforms.
Cc: Andrew Morton akpm@linux-foundation.org Cc: Aneesh Kumar K.V aneesh.kumar@linux.ibm.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Cc: Shuah Khan shuah@kernel.org Cc: linux-mm@kvack.org Cc: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Chaitanya S Prakash (5): selftests/mm: Add support for arm64 platform on va switch selftests/mm: Rename va_128TBswitch to va_high_addr_switch selftests/mm: Add platform independent in code comments selftests/mm: Configure nr_hugepages for arm64 selftests/mm: Run hugetlb testcases of va switch
tools/testing/selftests/mm/Makefile | 4 +- tools/testing/selftests/mm/run_vmtests.sh | 12 +++++- ...va_128TBswitch.c => va_high_addr_switch.c} | 41 +++++++++++++++---- ..._128TBswitch.sh => va_high_addr_switch.sh} | 6 ++- 4 files changed, 49 insertions(+), 14 deletions(-) rename tools/testing/selftests/mm/{va_128TBswitch.c => va_high_addr_switch.c} (86%) rename tools/testing/selftests/mm/{va_128TBswitch.sh => va_high_addr_switch.sh} (89%)
The patchset looks sane to me, but I have question: why arm64 has switch on 256TB. The reason we have the switch is to keep system backward compatible.
It's the same reason, it's just that arm64 initially supported 48-bits / 256TB of user addresses (0x0000000000000000..0x0000ffffffffffff), while x86_64 supported 47-bits / 128TB (0x0000000000000000..0x00007fffffffffff).
Note: arm64 has separate page tables for the user / kernel halves of the VA space, which in practice means we get an extra bit of address range for each half.
Maybe it is better to make arm64 switch also on 128TB to make it compatible across architectures?
I don't think that's something that we can change; user addresses between 128TB and 256TB have been the case for over a decade now, and avoiding that by default could easily break something.
Thanks, Mark.
On Thu, 23 Mar 2023 16:22:38 +0530 Chaitanya S Prakash chaitanyas.prakash@arm.com wrote:
The va_128TBswitch selftest is designed and implemented for PowerPC and x86 architectures which support a 128TB switch, up to 256TB of virtual address space and hugepage sizes of 16MB and 2MB respectively. Arm64 platforms on the other hand support a 256Tb switch, up to 4PB of virtual address space and a default hugepage size of 512MB when 64k pagesize is enabled.
These architectural differences require introducing support for arm64 platforms, after which a more generic naming convention is suggested. The in code comments are amended to provide a more platform independent explanation of the working of the code and nr_hugepages are configured as required. Finally, the file running the testcase is modified in order to prevent skipping of hugetlb testcases of va_high_addr_switch.
This series has been tested on 6.3.0-rc3 kernel, both on arm64 and x86 platforms.
Would it make sense to get this series into the ARM tree, so it sees more testing on ARM?
On 3/24/23 03:19, Andrew Morton wrote:
On Thu, 23 Mar 2023 16:22:38 +0530 Chaitanya S Prakash chaitanyas.prakash@arm.com wrote:
The va_128TBswitch selftest is designed and implemented for PowerPC and x86 architectures which support a 128TB switch, up to 256TB of virtual address space and hugepage sizes of 16MB and 2MB respectively. Arm64 platforms on the other hand support a 256Tb switch, up to 4PB of virtual address space and a default hugepage size of 512MB when 64k pagesize is enabled.
These architectural differences require introducing support for arm64 platforms, after which a more generic naming convention is suggested. The in code comments are amended to provide a more platform independent explanation of the working of the code and nr_hugepages are configured as required. Finally, the file running the testcase is modified in order to prevent skipping of hugetlb testcases of va_high_addr_switch.
This series has been tested on 6.3.0-rc3 kernel, both on arm64 and x86 platforms.
Would it make sense to get this series into the ARM tree, so it sees more testing on ARM
Rather, it will be better for this series to go via the mm tree instead (via linux-next first) for better coverage on all platforms, this being a common test.
On 3/23/23 16:22, Chaitanya S Prakash wrote:
The va_128TBswitch selftest is designed and implemented for PowerPC and x86 architectures which support a 128TB switch, up to 256TB of virtual address space and hugepage sizes of 16MB and 2MB respectively. Arm64 platforms on the other hand support a 256Tb switch, up to 4PB of virtual address space and a default hugepage size of 512MB when 64k pagesize is enabled.
These architectural differences require introducing support for arm64 platforms, after which a more generic naming convention is suggested. The in code comments are amended to provide a more platform independent explanation of the working of the code and nr_hugepages are configured as required. Finally, the file running the testcase is modified in order to prevent skipping of hugetlb testcases of va_high_addr_switch.
This series has been tested on 6.3.0-rc3 kernel, both on arm64 and x86 platforms.
Gentle ping, any updates?
Cc: Andrew Morton akpm@linux-foundation.org Cc: Aneesh Kumar K.V aneesh.kumar@linux.ibm.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Cc: Shuah Khan shuah@kernel.org Cc: linux-mm@kvack.org Cc: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Chaitanya S Prakash (5): selftests/mm: Add support for arm64 platform on va switch selftests/mm: Rename va_128TBswitch to va_high_addr_switch selftests/mm: Add platform independent in code comments selftests/mm: Configure nr_hugepages for arm64 selftests/mm: Run hugetlb testcases of va switch
tools/testing/selftests/mm/Makefile | 4 +- tools/testing/selftests/mm/run_vmtests.sh | 12 +++++- ...va_128TBswitch.c => va_high_addr_switch.c} | 41 +++++++++++++++---- ..._128TBswitch.sh => va_high_addr_switch.sh} | 6 ++- 4 files changed, 49 insertions(+), 14 deletions(-) rename tools/testing/selftests/mm/{va_128TBswitch.c => va_high_addr_switch.c} (86%) rename tools/testing/selftests/mm/{va_128TBswitch.sh => va_high_addr_switch.sh} (89%)
linux-kselftest-mirror@lists.linaro.org