linaro-kernel September 2022

linaro-kernel@lists.linaro.org

2 participants
5 discussions

[TCWG CI] Failure after v6.0-rc3-590-gaa3e1ec495ab: mm/hwpoison: move definitions of num_poisoned_pages_* to memory-failure.c

by ci_notify＠linaro.org

Failure after v6.0-rc3-590-gaa3e1ec495ab: mm/hwpoison: move definitions of num_poisoned_pages_* to memory-failure.c: Results changed to -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 7357 # First few build errors in logs: # 00:03:28 mm/memory-failure.c:77:20: error: static declaration of ‘num_poisoned_pages_inc’ follows non-static declaration # 00:03:28 make[1]: *** [scripts/Makefile.build:249: mm/memory-failure.o] Error 1 # 00:03:41 make: *** [Makefile:1854: mm] Error 2 from -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 8335 # linux build successful: all # linux boot successful: boot THIS IS THE END OF INTERESTING STUFF. BELOW ARE LINKS TO BUILDS, REPRODUCTION INSTRUCTIONS, AND THE RAW COMMIT. For latest status see comments in https://linaro.atlassian.net/browse/GNU-681 . Status of v6.0-rc3-590-gaa3e1ec495ab commit for tcwg_kernel: commit aa3e1ec495ab2fc3f22fc9240ad53a23b21fe0a1 Author: Naoya Horiguchi <naoya.horiguchi(a)nec.com> Date: Wed Sep 21 18:13:57 2022 +0900 mm/hwpoison: move definitions of num_poisoned_pages_* to memory-failure.c These interfaces will be used by drivers/base/core.c by later patch, so as a preparatory work move them to more common header file visible to the file. Link: https://lkml.kernel.org/r/20220921091359.25889-3-naoya.horiguchi@linux.dev Signed-off-by: Naoya Horiguchi <naoya.horiguchi(a)nec.com> Cc: David Hildenbrand <david(a)redhat.com> Cc: Jane Chu <jane.chu(a)oracle.com> Cc: Miaohe Lin <linmiaohe(a)huawei.com> Cc: Mike Kravetz <mike.kravetz(a)oracle.com> Cc: Muchun Song <songmuchun(a)bytedance.com> Cc: Oscar Salvador <osalvador(a)suse.de> Cc: Yang Shi <shy828301(a)gmail.com> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> * gnu-release-aarch64-next-defconfig ** Failure after v6.0-rc3-590-gaa3e1ec495ab: mm/hwpoison: move definitions of num_poisoned_pages_* to memory-failure.c: ** https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-de… Bad build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-de… Good build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-de… Reproduce current build: <cut> mkdir -p investigate-linux-aa3e1ec495ab2fc3f22fc9240ad53a23b21fe0a1 cd investigate-linux-aa3e1ec495ab2fc3f22fc9240ad53a23b21fe0a1 # Fetch scripts git clone https://git.linaro.org/toolchain/jenkins-scripts # Fetch manifests for bad and good builds mkdir -p bad/artifacts good/artifacts curl -o bad/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-de… --fail curl -o good/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-de… --fail # Reproduce bad build (cd bad; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) # Reproduce good build (cd good; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) </cut> Full commit (up to 1000 lines): <cut> commit aa3e1ec495ab2fc3f22fc9240ad53a23b21fe0a1 Author: Naoya Horiguchi <naoya.horiguchi(a)nec.com> Date: Wed Sep 21 18:13:57 2022 +0900 mm/hwpoison: move definitions of num_poisoned_pages_* to memory-failure.c These interfaces will be used by drivers/base/core.c by later patch, so as a preparatory work move them to more common header file visible to the file. Link: https://lkml.kernel.org/r/20220921091359.25889-3-naoya.horiguchi@linux.dev Signed-off-by: Naoya Horiguchi <naoya.horiguchi(a)nec.com> Cc: David Hildenbrand <david(a)redhat.com> Cc: Jane Chu <jane.chu(a)oracle.com> Cc: Miaohe Lin <linmiaohe(a)huawei.com> Cc: Mike Kravetz <mike.kravetz(a)oracle.com> Cc: Muchun Song <songmuchun(a)bytedance.com> Cc: Oscar Salvador <osalvador(a)suse.de> Cc: Yang Shi <shy828301(a)gmail.com> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- arch/parisc/kernel/pdt.c | 3 +-- include/linux/mm.h | 5 +++++ include/linux/swapops.h | 24 ++---------------------- mm/memory-failure.c | 10 ++++++++++ 4 files changed, 18 insertions(+), 24 deletions(-) diff --git a/arch/parisc/kernel/pdt.c b/arch/parisc/kernel/pdt.c index e391b175f5ec..fdc880e2575a 100644 --- a/arch/parisc/kernel/pdt.c +++ b/arch/parisc/kernel/pdt.c @@ -18,8 +18,7 @@ #include <linux/kthread.h> #include <linux/initrd.h> #include <linux/pgtable.h> -#include <linux/swap.h> -#include <linux/swapops.h> +#include <linux/mm.h> #include <asm/pdc.h> #include <asm/pdcpat.h> diff --git a/include/linux/mm.h b/include/linux/mm.h index c2277f5aba9e..80a2d800f272 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -3279,11 +3279,16 @@ extern atomic_long_t num_poisoned_pages __read_mostly; extern int soft_offline_page(unsigned long pfn, int flags); #ifdef CONFIG_MEMORY_FAILURE extern int __get_huge_page_for_hwpoison(unsigned long pfn, int flags); +extern void num_poisoned_pages_inc(void); #else static inline int __get_huge_page_for_hwpoison(unsigned long pfn, int flags) { return 0; } + +static inline void num_poisoned_pages_inc(void) +{ +} #endif #ifndef arch_memory_failure diff --git a/include/linux/swapops.h b/include/linux/swapops.h index 86b95ccb81bb..3ba9bf56899d 100644 --- a/include/linux/swapops.h +++ b/include/linux/swapops.h @@ -581,8 +581,6 @@ static inline int is_pmd_migration_entry(pmd_t pmd) #ifdef CONFIG_MEMORY_FAILURE -extern atomic_long_t num_poisoned_pages __read_mostly; - /* * Support for hardware poisoned pages */ @@ -597,17 +595,7 @@ static inline int is_hwpoison_entry(swp_entry_t entry) return swp_type(entry) == SWP_HWPOISON; } -static inline void num_poisoned_pages_inc(void) -{ - atomic_long_inc(&num_poisoned_pages); -} - -static inline void num_poisoned_pages_sub(long i) -{ - atomic_long_sub(i, &num_poisoned_pages); -} - -#else /* CONFIG_MEMORY_FAILURE */ +#else static inline swp_entry_t make_hwpoison_entry(struct page *page) { @@ -618,15 +606,7 @@ static inline int is_hwpoison_entry(swp_entry_t swp) { return 0; } - -static inline void num_poisoned_pages_inc(void) -{ -} - -static inline void num_poisoned_pages_sub(long i) -{ -} -#endif /* CONFIG_MEMORY_FAILURE */ +#endif static inline int non_swap_entry(swp_entry_t entry) { diff --git a/mm/memory-failure.c b/mm/memory-failure.c index 5942e1c0407e..aa6ce685b863 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -74,6 +74,16 @@ atomic_long_t num_poisoned_pages __read_mostly = ATOMIC_LONG_INIT(0); static bool hw_memory_failure __read_mostly = false; +static inline void num_poisoned_pages_inc(void) +{ + atomic_long_inc(&num_poisoned_pages); +} + +static inline void num_poisoned_pages_sub(long i) +{ + atomic_long_sub(i, &num_poisoned_pages); +} + /* * Return values: * 1: the page is dissolved (if needed) and taken off from buddy, </cut>

2 years, 2 months

[TCWG CI] Failure after v6.0-rc3-29-gc0357a73fa4a: arm64/sysreg: Align field names in ID_AA64DFR0_EL1 with architecture

by ci_notify＠linaro.org

Failure after v6.0-rc3-29-gc0357a73fa4a: arm64/sysreg: Align field names in ID_AA64DFR0_EL1 with architecture: Results changed to -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 21252 # First few build errors in logs: # 00:36:33 drivers/perf/arm_spe_pmu.c:677:14: error: ‘ID_AA64DFR0_PMSVER_8_2’ undeclared (first use in this function); did you mean ‘ID_AA64DFR0_PMSVer_8_2’? # 00:36:33 drivers/perf/arm_spe_pmu.c:679:14: error: ‘ID_AA64DFR0_PMSVER_8_3’ undeclared (first use in this function); did you mean ‘ID_AA64DFR0_PMSVer_8_3’? # 00:36:34 drivers/perf/arm_spe_pmu.c:961:52: error: ‘ID_AA64DFR0_PMSVER_SHIFT’ undeclared (first use in this function); did you mean ‘ID_AA64DFR0_PMSVer_SHIFT’? # 00:36:34 make[2]: *** [scripts/Makefile.build:249: drivers/perf/arm_spe_pmu.o] Error 1 # 00:36:43 make[1]: *** [scripts/Makefile.build:465: drivers/perf] Error 2 # 00:36:54 drivers/hwtracing/coresight/coresight-etm4x-core.c:969:26: error: ‘ID_AA64DFR0_TRACEVER_SHIFT’ undeclared (first use in this function); did you mean ‘ID_AA64DFR0_TraceVer_SHIFT’? # 00:36:54 drivers/hwtracing/coresight/coresight-etm4x-core.c:1057:57: error: ‘ID_AA64DFR0_TRACE_FILT_SHIFT’ undeclared (first use in this function); did you mean ‘ID_AA64DFR0_TraceFilt_SHIFT’? # 00:36:54 drivers/hwtracing/coresight/coresight-etm4x-core.c:970:1: error: control reaches end of non-void function [-Werror=return-type] # 00:36:54 make[2]: *** [scripts/Makefile.build:249: drivers/hwtracing/coresight/coresight-etm4x-core.o] Error 1 # 00:37:14 drivers/hwtracing/coresight/coresight-trbe.h:23:76: error: ‘ID_AA64DFR0_TRBE_SHIFT’ undeclared (first use in this function); did you mean ‘ID_AA64DFR0_WRPs_SHIFT’? from -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 21330 # linux build successful: all THIS IS THE END OF INTERESTING STUFF. BELOW ARE LINKS TO BUILDS, REPRODUCTION INSTRUCTIONS, AND THE RAW COMMIT. For latest status see comments in https://linaro.atlassian.net/browse/GNU-681 . Status of v6.0-rc3-29-gc0357a73fa4a commit for tcwg_kernel: commit c0357a73fa4a96d8ed9ee46e9927d9fcbc9d0828 Author: Mark Brown <broonie(a)kernel.org> Date: Sat Sep 10 17:33:49 2022 +0100 arm64/sysreg: Align field names in ID_AA64DFR0_EL1 with architecture The naming scheme the architecture uses for the fields in ID_AA64DFR0_EL1 does not align well with kernel conventions, using as it does a lot of MixedCase in various arrangements. In preparation for automatically generating the defines for this register rename the defines used to match what is in the architecture. Signed-off-by: Mark Brown <broonie(a)kernel.org> Link: https://lore.kernel.org/r/20220910163354.860255-2-broonie@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas(a)arm.com> * gnu-release-aarch64-next-allyesconfig ** Failure after v6.0-rc3-29-gc0357a73fa4a: arm64/sysreg: Align field names in ID_AA64DFR0_EL1 with architecture: ** https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-al… Bad build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-al… Good build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-al… Reproduce current build: <cut> mkdir -p investigate-linux-c0357a73fa4a96d8ed9ee46e9927d9fcbc9d0828 cd investigate-linux-c0357a73fa4a96d8ed9ee46e9927d9fcbc9d0828 # Fetch scripts git clone https://git.linaro.org/toolchain/jenkins-scripts # Fetch manifests for bad and good builds mkdir -p bad/artifacts good/artifacts curl -o bad/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-al… --fail curl -o good/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-aarch64-next-al… --fail # Reproduce bad build (cd bad; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) # Reproduce good build (cd good; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) </cut> Full commit (up to 1000 lines): <cut> commit c0357a73fa4a96d8ed9ee46e9927d9fcbc9d0828 Author: Mark Brown <broonie(a)kernel.org> Date: Sat Sep 10 17:33:49 2022 +0100 arm64/sysreg: Align field names in ID_AA64DFR0_EL1 with architecture The naming scheme the architecture uses for the fields in ID_AA64DFR0_EL1 does not align well with kernel conventions, using as it does a lot of MixedCase in various arrangements. In preparation for automatically generating the defines for this register rename the defines used to match what is in the architecture. Signed-off-by: Mark Brown <broonie(a)kernel.org> Link: https://lore.kernel.org/r/20220910163354.860255-2-broonie@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas(a)arm.com> --- arch/arm64/include/asm/assembler.h | 2 +- arch/arm64/include/asm/cpufeature.h | 2 +- arch/arm64/include/asm/el2_setup.h | 8 +++---- arch/arm64/include/asm/hw_breakpoint.h | 4 ++-- arch/arm64/include/asm/sysreg.h | 40 +++++++++++++++++----------------- arch/arm64/kernel/cpufeature.c | 14 ++++++------ arch/arm64/kernel/debug-monitors.c | 2 +- arch/arm64/kernel/perf_event.c | 8 +++---- arch/arm64/kvm/debug.c | 4 ++-- arch/arm64/kvm/hyp/nvhe/pkvm.c | 12 +++++----- arch/arm64/kvm/pmu-emul.c | 16 +++++++------- arch/arm64/kvm/sys_regs.c | 16 +++++++------- 12 files changed, 64 insertions(+), 64 deletions(-) diff --git a/arch/arm64/include/asm/assembler.h b/arch/arm64/include/asm/assembler.h index c1fc5f7bb978..71222f2b8a5a 100644 --- a/arch/arm64/include/asm/assembler.h +++ b/arch/arm64/include/asm/assembler.h @@ -512,7 +512,7 @@ alternative_endif */ .macro reset_pmuserenr_el0, tmpreg mrs \tmpreg, id_aa64dfr0_el1 - sbfx \tmpreg, \tmpreg, #ID_AA64DFR0_PMUVER_SHIFT, #4 + sbfx \tmpreg, \tmpreg, #ID_AA64DFR0_PMUVer_SHIFT, #4 cmp \tmpreg, #1 // Skip if no PMU present b.lt 9000f msr pmuserenr_el0, xzr // Disable PMU access from EL0 diff --git a/arch/arm64/include/asm/cpufeature.h b/arch/arm64/include/asm/cpufeature.h index 79bb9e58d9c6..e3b63967c8a9 100644 --- a/arch/arm64/include/asm/cpufeature.h +++ b/arch/arm64/include/asm/cpufeature.h @@ -553,7 +553,7 @@ cpuid_feature_cap_perfmon_field(u64 features, int field, u64 cap) u64 mask = GENMASK_ULL(field + 3, field); /* Treat IMPLEMENTATION DEFINED functionality as unimplemented */ - if (val == ID_AA64DFR0_PMUVER_IMP_DEF) + if (val == ID_AA64DFR0_PMUVer_IMP_DEF) val = 0; if (val > cap) { diff --git a/arch/arm64/include/asm/el2_setup.h b/arch/arm64/include/asm/el2_setup.h index b6e9bea7c9ec..03af4278bc23 100644 --- a/arch/arm64/include/asm/el2_setup.h +++ b/arch/arm64/include/asm/el2_setup.h @@ -40,7 +40,7 @@ .macro __init_el2_debug mrs x1, id_aa64dfr0_el1 - sbfx x0, x1, #ID_AA64DFR0_PMUVER_SHIFT, #4 + sbfx x0, x1, #ID_AA64DFR0_PMUVer_SHIFT, #4 cmp x0, #1 b.lt .Lskip_pmu_\@ // Skip if no PMU present mrs x0, pmcr_el0 // Disable debug access traps @@ -49,7 +49,7 @@ csel x2, xzr, x0, lt // all PMU counters from EL1 /* Statistical profiling */ - ubfx x0, x1, #ID_AA64DFR0_PMSVER_SHIFT, #4 + ubfx x0, x1, #ID_AA64DFR0_PMSVer_SHIFT, #4 cbz x0, .Lskip_spe_\@ // Skip if SPE not present mrs_s x0, SYS_PMBIDR_EL1 // If SPE available at EL2, @@ -65,7 +65,7 @@ .Lskip_spe_\@: /* Trace buffer */ - ubfx x0, x1, #ID_AA64DFR0_TRBE_SHIFT, #4 + ubfx x0, x1, #ID_AA64DFR0_TraceBuffer_SHIFT, #4 cbz x0, .Lskip_trace_\@ // Skip if TraceBuffer is not present mrs_s x0, SYS_TRBIDR_EL1 @@ -137,7 +137,7 @@ mov x0, xzr mrs x1, id_aa64dfr0_el1 - ubfx x1, x1, #ID_AA64DFR0_PMSVER_SHIFT, #4 + ubfx x1, x1, #ID_AA64DFR0_PMSVer_SHIFT, #4 cmp x1, #3 b.lt .Lset_debug_fgt_\@ /* Disable PMSNEVFR_EL1 read and write traps */ diff --git a/arch/arm64/include/asm/hw_breakpoint.h b/arch/arm64/include/asm/hw_breakpoint.h index bc7aaed4b34e..d667c03d5f5e 100644 --- a/arch/arm64/include/asm/hw_breakpoint.h +++ b/arch/arm64/include/asm/hw_breakpoint.h @@ -142,7 +142,7 @@ static inline int get_num_brps(void) u64 dfr0 = read_sanitised_ftr_reg(SYS_ID_AA64DFR0_EL1); return 1 + cpuid_feature_extract_unsigned_field(dfr0, - ID_AA64DFR0_BRPS_SHIFT); + ID_AA64DFR0_BRPs_SHIFT); } /* Determine number of WRP registers available. */ @@ -151,7 +151,7 @@ static inline int get_num_wrps(void) u64 dfr0 = read_sanitised_ftr_reg(SYS_ID_AA64DFR0_EL1); return 1 + cpuid_feature_extract_unsigned_field(dfr0, - ID_AA64DFR0_WRPS_SHIFT); + ID_AA64DFR0_WRPs_SHIFT); } #endif /* __ASM_BREAKPOINT_H */ diff --git a/arch/arm64/include/asm/sysreg.h b/arch/arm64/include/asm/sysreg.h index c7876363c6e5..b1e9e4d3d964 100644 --- a/arch/arm64/include/asm/sysreg.h +++ b/arch/arm64/include/asm/sysreg.h @@ -700,26 +700,26 @@ /* id_aa64dfr0 */ #define ID_AA64DFR0_MTPMU_SHIFT 48 -#define ID_AA64DFR0_TRBE_SHIFT 44 -#define ID_AA64DFR0_TRACE_FILT_SHIFT 40 -#define ID_AA64DFR0_DOUBLELOCK_SHIFT 36 -#define ID_AA64DFR0_PMSVER_SHIFT 32 -#define ID_AA64DFR0_CTX_CMPS_SHIFT 28 -#define ID_AA64DFR0_WRPS_SHIFT 20 -#define ID_AA64DFR0_BRPS_SHIFT 12 -#define ID_AA64DFR0_PMUVER_SHIFT 8 -#define ID_AA64DFR0_TRACEVER_SHIFT 4 -#define ID_AA64DFR0_DEBUGVER_SHIFT 0 - -#define ID_AA64DFR0_PMUVER_8_0 0x1 -#define ID_AA64DFR0_PMUVER_8_1 0x4 -#define ID_AA64DFR0_PMUVER_8_4 0x5 -#define ID_AA64DFR0_PMUVER_8_5 0x6 -#define ID_AA64DFR0_PMUVER_8_7 0x7 -#define ID_AA64DFR0_PMUVER_IMP_DEF 0xf - -#define ID_AA64DFR0_PMSVER_8_2 0x1 -#define ID_AA64DFR0_PMSVER_8_3 0x2 +#define ID_AA64DFR0_TraceBuffer_SHIFT 44 +#define ID_AA64DFR0_TraceFilt_SHIFT 40 +#define ID_AA64DFR0_DoubleLock_SHIFT 36 +#define ID_AA64DFR0_PMSVer_SHIFT 32 +#define ID_AA64DFR0_CTX_CMPs_SHIFT 28 +#define ID_AA64DFR0_WRPs_SHIFT 20 +#define ID_AA64DFR0_BRPs_SHIFT 12 +#define ID_AA64DFR0_PMUVer_SHIFT 8 +#define ID_AA64DFR0_TraceVer_SHIFT 4 +#define ID_AA64DFR0_DebugVer_SHIFT 0 + +#define ID_AA64DFR0_PMUVer_8_0 0x1 +#define ID_AA64DFR0_PMUVer_8_1 0x4 +#define ID_AA64DFR0_PMUVer_8_4 0x5 +#define ID_AA64DFR0_PMUVer_8_5 0x6 +#define ID_AA64DFR0_PMUVer_8_7 0x7 +#define ID_AA64DFR0_PMUVer_IMP_DEF 0xf + +#define ID_AA64DFR0_PMSVer_8_2 0x1 +#define ID_AA64DFR0_PMSVer_8_3 0x2 #define ID_DFR0_PERFMON_SHIFT 24 diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c index c2e42feb3e1a..c694d6cbf82d 100644 --- a/arch/arm64/kernel/cpufeature.c +++ b/arch/arm64/kernel/cpufeature.c @@ -434,17 +434,17 @@ static const struct arm64_ftr_bits ftr_id_mmfr0[] = { }; static const struct arm64_ftr_bits ftr_id_aa64dfr0[] = { - S_ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64DFR0_DOUBLELOCK_SHIFT, 4, 0), - ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64DFR0_PMSVER_SHIFT, 4, 0), - ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64DFR0_CTX_CMPS_SHIFT, 4, 0), - ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64DFR0_WRPS_SHIFT, 4, 0), - ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64DFR0_BRPS_SHIFT, 4, 0), + S_ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64DFR0_DoubleLock_SHIFT, 4, 0), + ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64DFR0_PMSVer_SHIFT, 4, 0), + ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64DFR0_CTX_CMPs_SHIFT, 4, 0), + ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64DFR0_WRPs_SHIFT, 4, 0), + ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64DFR0_BRPs_SHIFT, 4, 0), /* * We can instantiate multiple PMU instances with different levels * of support. */ - S_ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_EXACT, ID_AA64DFR0_PMUVER_SHIFT, 4, 0), - ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_EXACT, ID_AA64DFR0_DEBUGVER_SHIFT, 4, 0x6), + S_ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_EXACT, ID_AA64DFR0_PMUVer_SHIFT, 4, 0), + ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_EXACT, ID_AA64DFR0_DebugVer_SHIFT, 4, 0x6), ARM64_FTR_END, }; diff --git a/arch/arm64/kernel/debug-monitors.c b/arch/arm64/kernel/debug-monitors.c index bf9fe71589bc..572c8ac2873c 100644 --- a/arch/arm64/kernel/debug-monitors.c +++ b/arch/arm64/kernel/debug-monitors.c @@ -28,7 +28,7 @@ u8 debug_monitors_arch(void) { return cpuid_feature_extract_unsigned_field(read_sanitised_ftr_reg(SYS_ID_AA64DFR0_EL1), - ID_AA64DFR0_DEBUGVER_SHIFT); + ID_AA64DFR0_DebugVer_SHIFT); } /* diff --git a/arch/arm64/kernel/perf_event.c b/arch/arm64/kernel/perf_event.c index cb69ff1e6138..83f7a8980623 100644 --- a/arch/arm64/kernel/perf_event.c +++ b/arch/arm64/kernel/perf_event.c @@ -390,7 +390,7 @@ static const struct attribute_group armv8_pmuv3_caps_attr_group = { */ static bool armv8pmu_has_long_event(struct arm_pmu *cpu_pmu) { - return (cpu_pmu->pmuver >= ID_AA64DFR0_PMUVER_8_5); + return (cpu_pmu->pmuver >= ID_AA64DFR0_PMUVer_8_5); } static inline bool armv8pmu_event_has_user_read(struct perf_event *event) @@ -1145,8 +1145,8 @@ static void __armv8pmu_probe_pmu(void *info) dfr0 = read_sysreg(id_aa64dfr0_el1); pmuver = cpuid_feature_extract_unsigned_field(dfr0, - ID_AA64DFR0_PMUVER_SHIFT); - if (pmuver == ID_AA64DFR0_PMUVER_IMP_DEF || pmuver == 0) + ID_AA64DFR0_PMUVer_SHIFT); + if (pmuver == ID_AA64DFR0_PMUVer_IMP_DEF || pmuver == 0) return; cpu_pmu->pmuver = pmuver; @@ -1172,7 +1172,7 @@ static void __armv8pmu_probe_pmu(void *info) pmceid, ARMV8_PMUV3_MAX_COMMON_EVENTS); /* store PMMIR_EL1 register for sysfs */ - if (pmuver >= ID_AA64DFR0_PMUVER_8_4 && (pmceid_raw[1] & BIT(31))) + if (pmuver >= ID_AA64DFR0_PMUVer_8_4 && (pmceid_raw[1] & BIT(31))) cpu_pmu->reg_pmmir = read_cpuid(PMMIR_EL1); else cpu_pmu->reg_pmmir = 0; diff --git a/arch/arm64/kvm/debug.c b/arch/arm64/kvm/debug.c index 0b28d7db7c76..14878cd55c53 100644 --- a/arch/arm64/kvm/debug.c +++ b/arch/arm64/kvm/debug.c @@ -295,12 +295,12 @@ void kvm_arch_vcpu_load_debug_state_flags(struct kvm_vcpu *vcpu) * If SPE is present on this CPU and is available at current EL, * we may need to check if the host state needs to be saved. */ - if (cpuid_feature_extract_unsigned_field(dfr0, ID_AA64DFR0_PMSVER_SHIFT) && + if (cpuid_feature_extract_unsigned_field(dfr0, ID_AA64DFR0_PMSVer_SHIFT) && !(read_sysreg_s(SYS_PMBIDR_EL1) & BIT(SYS_PMBIDR_EL1_P_SHIFT))) vcpu_set_flag(vcpu, DEBUG_STATE_SAVE_SPE); /* Check if we have TRBE implemented and available at the host */ - if (cpuid_feature_extract_unsigned_field(dfr0, ID_AA64DFR0_TRBE_SHIFT) && + if (cpuid_feature_extract_unsigned_field(dfr0, ID_AA64DFR0_TraceBuffer_SHIFT) && !(read_sysreg_s(SYS_TRBIDR_EL1) & TRBIDR_PROG)) vcpu_set_flag(vcpu, DEBUG_STATE_SAVE_TRBE); } diff --git a/arch/arm64/kvm/hyp/nvhe/pkvm.c b/arch/arm64/kvm/hyp/nvhe/pkvm.c index fc3e32709ba2..ba1327fac03f 100644 --- a/arch/arm64/kvm/hyp/nvhe/pkvm.c +++ b/arch/arm64/kvm/hyp/nvhe/pkvm.c @@ -86,32 +86,32 @@ static void pvm_init_traps_aa64dfr0(struct kvm_vcpu *vcpu) u64 cptr_set = 0; /* Trap/constrain PMU */ - if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_PMUVER), feature_ids)) { + if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_PMUVer), feature_ids)) { mdcr_set |= MDCR_EL2_TPM | MDCR_EL2_TPMCR; mdcr_clear |= MDCR_EL2_HPME | MDCR_EL2_MTPME | MDCR_EL2_HPMN_MASK; } /* Trap Debug */ - if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_DEBUGVER), feature_ids)) + if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_DebugVer), feature_ids)) mdcr_set |= MDCR_EL2_TDRA | MDCR_EL2_TDA | MDCR_EL2_TDE; /* Trap OS Double Lock */ - if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_DOUBLELOCK), feature_ids)) + if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_DoubleLock), feature_ids)) mdcr_set |= MDCR_EL2_TDOSA; /* Trap SPE */ - if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_PMSVER), feature_ids)) { + if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_PMSVer), feature_ids)) { mdcr_set |= MDCR_EL2_TPMS; mdcr_clear |= MDCR_EL2_E2PB_MASK << MDCR_EL2_E2PB_SHIFT; } /* Trap Trace Filter */ - if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_TRACE_FILT), feature_ids)) + if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_TraceFilt), feature_ids)) mdcr_set |= MDCR_EL2_TTRF; /* Trap Trace */ - if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_TRACEVER), feature_ids)) + if (!FIELD_GET(ARM64_FEATURE_MASK(ID_AA64DFR0_TraceVer), feature_ids)) cptr_set |= CPTR_EL2_TTA; vcpu->arch.mdcr_el2 |= mdcr_set; diff --git a/arch/arm64/kvm/pmu-emul.c b/arch/arm64/kvm/pmu-emul.c index 11c43bed5f97..331478c67dca 100644 --- a/arch/arm64/kvm/pmu-emul.c +++ b/arch/arm64/kvm/pmu-emul.c @@ -33,12 +33,12 @@ static u32 kvm_pmu_event_mask(struct kvm *kvm) pmuver = kvm->arch.arm_pmu->pmuver; switch (pmuver) { - case ID_AA64DFR0_PMUVER_8_0: + case ID_AA64DFR0_PMUVer_8_0: return GENMASK(9, 0); - case ID_AA64DFR0_PMUVER_8_1: - case ID_AA64DFR0_PMUVER_8_4: - case ID_AA64DFR0_PMUVER_8_5: - case ID_AA64DFR0_PMUVER_8_7: + case ID_AA64DFR0_PMUVer_8_1: + case ID_AA64DFR0_PMUVer_8_4: + case ID_AA64DFR0_PMUVer_8_5: + case ID_AA64DFR0_PMUVer_8_7: return GENMASK(15, 0); default: /* Shouldn't be here, just for sanity */ WARN_ONCE(1, "Unknown PMU version %d\n", pmuver); @@ -774,7 +774,7 @@ void kvm_host_pmu_init(struct arm_pmu *pmu) { struct arm_pmu_entry *entry; - if (pmu->pmuver == 0 || pmu->pmuver == ID_AA64DFR0_PMUVER_IMP_DEF) + if (pmu->pmuver == 0 || pmu->pmuver == ID_AA64DFR0_PMUVer_IMP_DEF) return; mutex_lock(&arm_pmus_lock); @@ -828,7 +828,7 @@ static struct arm_pmu *kvm_pmu_probe_armpmu(void) if (event->pmu) { pmu = to_arm_pmu(event->pmu); if (pmu->pmuver == 0 || - pmu->pmuver == ID_AA64DFR0_PMUVER_IMP_DEF) + pmu->pmuver == ID_AA64DFR0_PMUVer_IMP_DEF) pmu = NULL; } @@ -856,7 +856,7 @@ u64 kvm_pmu_get_pmceid(struct kvm_vcpu *vcpu, bool pmceid1) * Don't advertise STALL_SLOT, as PMMIR_EL0 is handled * as RAZ */ - if (vcpu->kvm->arch.arm_pmu->pmuver >= ID_AA64DFR0_PMUVER_8_4) + if (vcpu->kvm->arch.arm_pmu->pmuver >= ID_AA64DFR0_PMUVer_8_4) val &= ~BIT_ULL(ARMV8_PMUV3_PERFCTR_STALL_SLOT - 32); base = 32; } diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index fa61793467a7..626d7769e23d 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -1110,14 +1110,14 @@ static u64 read_id_reg(const struct kvm_vcpu *vcpu, break; case SYS_ID_AA64DFR0_EL1: /* Limit debug to ARMv8.0 */ - val &= ~ARM64_FEATURE_MASK(ID_AA64DFR0_DEBUGVER); - val |= FIELD_PREP(ARM64_FEATURE_MASK(ID_AA64DFR0_DEBUGVER), 6); + val &= ~ARM64_FEATURE_MASK(ID_AA64DFR0_DebugVer); + val |= FIELD_PREP(ARM64_FEATURE_MASK(ID_AA64DFR0_DebugVer), 6); /* Limit guests to PMUv3 for ARMv8.4 */ val = cpuid_feature_cap_perfmon_field(val, - ID_AA64DFR0_PMUVER_SHIFT, - kvm_vcpu_has_pmu(vcpu) ? ID_AA64DFR0_PMUVER_8_4 : 0); + ID_AA64DFR0_PMUVer_SHIFT, + kvm_vcpu_has_pmu(vcpu) ? ID_AA64DFR0_PMUVer_8_4 : 0); /* Hide SPE from guests */ - val &= ~ARM64_FEATURE_MASK(ID_AA64DFR0_PMSVER); + val &= ~ARM64_FEATURE_MASK(ID_AA64DFR0_PMSVer); break; case SYS_ID_DFR0_EL1: /* Limit guests to PMUv3 for ARMv8.4 */ @@ -1827,9 +1827,9 @@ static bool trap_dbgdidr(struct kvm_vcpu *vcpu, u64 pfr = read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1); u32 el3 = !!cpuid_feature_extract_unsigned_field(pfr, ID_AA64PFR0_EL1_EL3_SHIFT); - p->regval = ((((dfr >> ID_AA64DFR0_WRPS_SHIFT) & 0xf) << 28) | - (((dfr >> ID_AA64DFR0_BRPS_SHIFT) & 0xf) << 24) | - (((dfr >> ID_AA64DFR0_CTX_CMPS_SHIFT) & 0xf) << 20) + p->regval = ((((dfr >> ID_AA64DFR0_WRPs_SHIFT) & 0xf) << 28) | + (((dfr >> ID_AA64DFR0_BRPs_SHIFT) & 0xf) << 24) | + (((dfr >> ID_AA64DFR0_CTX_CMPs_SHIFT) & 0xf) << 20) | (6 << 16) | (1 << 15) | (el3 << 14) | (el3 << 12)); return true; } </cut>

2 years, 3 months

[TCWG CI] Failure after v6.0-rc1-21-g3f74cd046fbe: ata: libahci_platform: Parse ports-implemented property in resources getter

by ci_notify＠linaro.org

Failure after v6.0-rc1-21-g3f74cd046fbe: ata: libahci_platform: Parse ports-implemented property in resources getter: Results changed to -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 23039 # First few build errors in logs: # 00:35:13 drivers/ata/ahci_st.c:147:24: error: unused variable ‘dev’ [-Werror=unused-variable] # 00:35:13 make[2]: *** [scripts/Makefile.build:249: drivers/ata/ahci_st.o] Error 1 # 00:36:37 make[1]: *** [scripts/Makefile.build:465: drivers/ata] Error 2 # 00:38:02 make: *** [Makefile:1855: drivers] Error 2 from -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 31637 # linux build successful: all THIS IS THE END OF INTERESTING STUFF. BELOW ARE LINKS TO BUILDS, REPRODUCTION INSTRUCTIONS, AND THE RAW COMMIT. For latest status see comments in https://linaro.atlassian.net/browse/GNU-681 . Status of v6.0-rc1-21-g3f74cd046fbe commit for tcwg_kernel: commit 3f74cd046fbed349be977606f938e6429155e7b5 Author: Serge Semin <Sergey.Semin(a)baikalelectronics.ru> Date: Fri Sep 9 22:36:07 2022 +0300 ata: libahci_platform: Parse ports-implemented property in resources getter The ports-implemented property is mainly used on the OF-based platforms with no ports mapping initialized by a bootloader/BIOS firmware. Seeing the same of_property_read_u32()-based pattern has already been implemented in the generic AHCI LLDD (glue) driver and in the Mediatek, St AHCI drivers let's move the property read procedure to the generic ahci_platform_get_resources() method. Thus we'll have the forced ports mapping feature supported for each OF-based platform which requires that, and stop re-implementing the same pattern in there a bit simplifying the code. Signed-off-by: Serge Semin <Sergey.Semin(a)baikalelectronics.ru> Signed-off-by: Damien Le Moal <damien.lemoal(a)opensource.wdc.com> * gnu-release-arm-next-allmodconfig ** Failure after v6.0-rc1-21-g3f74cd046fbe: ata: libahci_platform: Parse ports-implemented property in resources getter: ** https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… Bad build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… Good build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… Reproduce current build: <cut> mkdir -p investigate-linux-3f74cd046fbed349be977606f938e6429155e7b5 cd investigate-linux-3f74cd046fbed349be977606f938e6429155e7b5 # Fetch scripts git clone https://git.linaro.org/toolchain/jenkins-scripts # Fetch manifests for bad and good builds mkdir -p bad/artifacts good/artifacts curl -o bad/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… --fail curl -o good/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… --fail # Reproduce bad build (cd bad; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) # Reproduce good build (cd good; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) </cut> Full commit (up to 1000 lines): <cut> commit 3f74cd046fbed349be977606f938e6429155e7b5 Author: Serge Semin <Sergey.Semin(a)baikalelectronics.ru> Date: Fri Sep 9 22:36:07 2022 +0300 ata: libahci_platform: Parse ports-implemented property in resources getter The ports-implemented property is mainly used on the OF-based platforms with no ports mapping initialized by a bootloader/BIOS firmware. Seeing the same of_property_read_u32()-based pattern has already been implemented in the generic AHCI LLDD (glue) driver and in the Mediatek, St AHCI drivers let's move the property read procedure to the generic ahci_platform_get_resources() method. Thus we'll have the forced ports mapping feature supported for each OF-based platform which requires that, and stop re-implementing the same pattern in there a bit simplifying the code. Signed-off-by: Serge Semin <Sergey.Semin(a)baikalelectronics.ru> Signed-off-by: Damien Le Moal <damien.lemoal(a)opensource.wdc.com> --- drivers/ata/ahci_mtk.c | 2 -- drivers/ata/ahci_platform.c | 3 --- drivers/ata/ahci_st.c | 3 --- drivers/ata/libahci_platform.c | 3 +++ 4 files changed, 3 insertions(+), 8 deletions(-) diff --git a/drivers/ata/ahci_mtk.c b/drivers/ata/ahci_mtk.c index 1f6c85fde983..c056378e3e72 100644 --- a/drivers/ata/ahci_mtk.c +++ b/drivers/ata/ahci_mtk.c @@ -118,8 +118,6 @@ static int mtk_ahci_parse_property(struct ahci_host_priv *hpriv, SYS_CFG_SATA_EN); } - of_property_read_u32(np, "ports-implemented", &hpriv->force_port_map); - return 0; } diff --git a/drivers/ata/ahci_platform.c b/drivers/ata/ahci_platform.c index 28a8de5b48b9..9b56490ecbc3 100644 --- a/drivers/ata/ahci_platform.c +++ b/drivers/ata/ahci_platform.c @@ -56,9 +56,6 @@ static int ahci_probe(struct platform_device *pdev) if (rc) return rc; - of_property_read_u32(dev->of_node, - "ports-implemented", &hpriv->force_port_map); - if (of_device_is_compatible(dev->of_node, "hisilicon,hisi-ahci")) hpriv->flags |= AHCI_HFLAG_NO_FBS | AHCI_HFLAG_NO_NCQ; diff --git a/drivers/ata/ahci_st.c b/drivers/ata/ahci_st.c index 7526653c843b..068621099c00 100644 --- a/drivers/ata/ahci_st.c +++ b/drivers/ata/ahci_st.c @@ -168,9 +168,6 @@ static int st_ahci_probe(struct platform_device *pdev) st_ahci_configure_oob(hpriv->mmio); - of_property_read_u32(dev->of_node, - "ports-implemented", &hpriv->force_port_map); - err = ahci_platform_init_host(pdev, hpriv, &st_ahci_port_info, &ahci_platform_sht); if (err) { diff --git a/drivers/ata/libahci_platform.c b/drivers/ata/libahci_platform.c index bacb974c1b16..085f99b2eb5a 100644 --- a/drivers/ata/libahci_platform.c +++ b/drivers/ata/libahci_platform.c @@ -485,6 +485,9 @@ struct ahci_host_priv *ahci_platform_get_resources(struct platform_device *pdev, goto err_out; } + of_property_read_u32(dev->of_node, + "ports-implemented", &hpriv->force_port_map); + if (child_nodes) { for_each_child_of_node(dev->of_node, child) { u32 port; </cut>

2 years, 3 months

[TCWG CI] Failure after v6.0-rc1-35-g108713a713c7: crypto: aspeed - Add HACE hash driver

by ci_notify＠linaro.org

Failure after v6.0-rc1-35-g108713a713c7: crypto: aspeed - Add HACE hash driver: Results changed to -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 23063 # First few build errors in logs: # 00:31:16 make[1]: *** [scripts/Makefile.modpost:128: modules-only.symvers] Error 1 # 00:31:16 make: *** [Makefile:1769: modules] Error 2 from -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 31638 # linux build successful: all THIS IS THE END OF INTERESTING STUFF. BELOW ARE LINKS TO BUILDS, REPRODUCTION INSTRUCTIONS, AND THE RAW COMMIT. For latest status see comments in https://linaro.atlassian.net/browse/GNU-681 . Status of v6.0-rc1-35-g108713a713c7 commit for tcwg_kernel: commit 108713a713c7e4b7d07e6cd9b808503d5bb7089b Author: Neal Liu <neal_liu(a)aspeedtech.com> Date: Thu Aug 18 11:59:52 2022 +0800 crypto: aspeed - Add HACE hash driver Hash and Crypto Engine (HACE) is designed to accelerate the throughput of hash data digest, encryption, and decryption. Basically, HACE can be divided into two independently engines - Hash Engine and Crypto Engine. This patch aims to add HACE hash engine driver for hash accelerator. Signed-off-by: Neal Liu <neal_liu(a)aspeedtech.com> Signed-off-by: Johnny Huang <johnny_huang(a)aspeedtech.com> Reviewed-by: Dhananjay Phadke <dphadke(a)linux.microsoft.com> Signed-off-by: Herbert Xu <herbert(a)gondor.apana.org.au> * gnu-release-arm-next-allmodconfig ** Failure after v6.0-rc1-35-g108713a713c7: crypto: aspeed - Add HACE hash driver: ** https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… Bad build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… Good build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… Reproduce current build: <cut> mkdir -p investigate-linux-108713a713c7e4b7d07e6cd9b808503d5bb7089b cd investigate-linux-108713a713c7e4b7d07e6cd9b808503d5bb7089b # Fetch scripts git clone https://git.linaro.org/toolchain/jenkins-scripts # Fetch manifests for bad and good builds mkdir -p bad/artifacts good/artifacts curl -o bad/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… --fail curl -o good/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-next-allmod… --fail # Reproduce bad build (cd bad; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) # Reproduce good build (cd good; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) </cut> Full commit (up to 1000 lines): <cut> commit 108713a713c7e4b7d07e6cd9b808503d5bb7089b Author: Neal Liu <neal_liu(a)aspeedtech.com> Date: Thu Aug 18 11:59:52 2022 +0800 crypto: aspeed - Add HACE hash driver Hash and Crypto Engine (HACE) is designed to accelerate the throughput of hash data digest, encryption, and decryption. Basically, HACE can be divided into two independently engines - Hash Engine and Crypto Engine. This patch aims to add HACE hash engine driver for hash accelerator. Signed-off-by: Neal Liu <neal_liu(a)aspeedtech.com> Signed-off-by: Johnny Huang <johnny_huang(a)aspeedtech.com> Reviewed-by: Dhananjay Phadke <dphadke(a)linux.microsoft.com> Signed-off-by: Herbert Xu <herbert(a)gondor.apana.org.au> --- MAINTAINERS | 7 + drivers/crypto/Kconfig | 1 + drivers/crypto/Makefile | 1 + drivers/crypto/aspeed/Kconfig | 32 + drivers/crypto/aspeed/Makefile | 6 + drivers/crypto/aspeed/aspeed-hace-hash.c | 1389 ++++++++++++++++++++++++++++++ drivers/crypto/aspeed/aspeed-hace.c | 206 +++++ drivers/crypto/aspeed/aspeed-hace.h | 186 ++++ 8 files changed, 1828 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index 8a5012ba6ff9..164f67e59e5f 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3210,6 +3210,13 @@ S: Maintained F: Documentation/devicetree/bindings/usb/aspeed,ast2600-udc.yaml F: drivers/usb/gadget/udc/aspeed_udc.c +ASPEED CRYPTO DRIVER +M: Neal Liu <neal_liu(a)aspeedtech.com> +L: linux-aspeed(a)lists.ozlabs.org (moderated for non-subscribers) +S: Maintained +F: Documentation/devicetree/bindings/crypto/aspeed,ast2500-hace.yaml +F: drivers/crypto/aspeed/ + ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS M: Corentin Chary <corentin.chary(a)gmail.com> L: acpi4asus-user(a)lists.sourceforge.net diff --git a/drivers/crypto/Kconfig b/drivers/crypto/Kconfig index 3e6aa319920b..f6fe1945edd6 100644 --- a/drivers/crypto/Kconfig +++ b/drivers/crypto/Kconfig @@ -818,5 +818,6 @@ config CRYPTO_DEV_SA2UL acceleration for cryptographic algorithms on these devices. source "drivers/crypto/keembay/Kconfig" +source "drivers/crypto/aspeed/Kconfig" endif # CRYPTO_HW diff --git a/drivers/crypto/Makefile b/drivers/crypto/Makefile index f81703a86b98..116de173a66c 100644 --- a/drivers/crypto/Makefile +++ b/drivers/crypto/Makefile @@ -1,5 +1,6 @@ # SPDX-License-Identifier: GPL-2.0 obj-$(CONFIG_CRYPTO_DEV_ALLWINNER) += allwinner/ +obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed/ obj-$(CONFIG_CRYPTO_DEV_ATMEL_AES) += atmel-aes.o obj-$(CONFIG_CRYPTO_DEV_ATMEL_SHA) += atmel-sha.o obj-$(CONFIG_CRYPTO_DEV_ATMEL_TDES) += atmel-tdes.o diff --git a/drivers/crypto/aspeed/Kconfig b/drivers/crypto/aspeed/Kconfig new file mode 100644 index 000000000000..03e3fc61498e --- /dev/null +++ b/drivers/crypto/aspeed/Kconfig @@ -0,0 +1,32 @@ +config CRYPTO_DEV_ASPEED + tristate "Support for Aspeed cryptographic engine driver" + depends on ARCH_ASPEED + help + Hash and Crypto Engine (HACE) is designed to accelerate the + throughput of hash data digest, encryption and decryption. + + Select y here to have support for the cryptographic driver + available on Aspeed SoC. + +config CRYPTO_DEV_ASPEED_DEBUG + bool "Enable Aspeed crypto debug messages" + depends on CRYPTO_DEV_ASPEED + help + Print Aspeed crypto debugging messages if you use this + option to ask for those messages. + Avoid enabling this option for production build to + minimize driver timing. + +config CRYPTO_DEV_ASPEED_HACE_HASH + bool "Enable Aspeed Hash & Crypto Engine (HACE) hash" + depends on CRYPTO_DEV_ASPEED + select CRYPTO_ENGINE + select CRYPTO_SHA1 + select CRYPTO_SHA256 + select CRYPTO_SHA512 + select CRYPTO_HMAC + help + Select here to enable Aspeed Hash & Crypto Engine (HACE) + hash driver. + Supports multiple message digest standards, including + SHA-1, SHA-224, SHA-256, SHA-384, SHA-512, and so on. diff --git a/drivers/crypto/aspeed/Makefile b/drivers/crypto/aspeed/Makefile new file mode 100644 index 000000000000..8bc8d4fed5a9 --- /dev/null +++ b/drivers/crypto/aspeed/Makefile @@ -0,0 +1,6 @@ +obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed_crypto.o +aspeed_crypto-objs := aspeed-hace.o \ + $(hace-hash-y) + +obj-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) += aspeed-hace-hash.o +hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) := aspeed-hace-hash.o diff --git a/drivers/crypto/aspeed/aspeed-hace-hash.c b/drivers/crypto/aspeed/aspeed-hace-hash.c new file mode 100644 index 000000000000..0a44ffc0e13b --- /dev/null +++ b/drivers/crypto/aspeed/aspeed-hace-hash.c @@ -0,0 +1,1389 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * Copyright (c) 2021 Aspeed Technology Inc. + */ + +#include "aspeed-hace.h" + +#ifdef CONFIG_CRYPTO_DEV_ASPEED_DEBUG +#define AHASH_DBG(h, fmt, ...) \ + dev_info((h)->dev, "%s() " fmt, __func__, ##__VA_ARGS__) +#else +#define AHASH_DBG(h, fmt, ...) \ + dev_dbg((h)->dev, "%s() " fmt, __func__, ##__VA_ARGS__) +#endif + +/* Initialization Vectors for SHA-family */ +static const __be32 sha1_iv[8] = { + cpu_to_be32(SHA1_H0), cpu_to_be32(SHA1_H1), + cpu_to_be32(SHA1_H2), cpu_to_be32(SHA1_H3), + cpu_to_be32(SHA1_H4), 0, 0, 0 +}; + +static const __be32 sha224_iv[8] = { + cpu_to_be32(SHA224_H0), cpu_to_be32(SHA224_H1), + cpu_to_be32(SHA224_H2), cpu_to_be32(SHA224_H3), + cpu_to_be32(SHA224_H4), cpu_to_be32(SHA224_H5), + cpu_to_be32(SHA224_H6), cpu_to_be32(SHA224_H7), +}; + +static const __be32 sha256_iv[8] = { + cpu_to_be32(SHA256_H0), cpu_to_be32(SHA256_H1), + cpu_to_be32(SHA256_H2), cpu_to_be32(SHA256_H3), + cpu_to_be32(SHA256_H4), cpu_to_be32(SHA256_H5), + cpu_to_be32(SHA256_H6), cpu_to_be32(SHA256_H7), +}; + +static const __be64 sha384_iv[8] = { + cpu_to_be64(SHA384_H0), cpu_to_be64(SHA384_H1), + cpu_to_be64(SHA384_H2), cpu_to_be64(SHA384_H3), + cpu_to_be64(SHA384_H4), cpu_to_be64(SHA384_H5), + cpu_to_be64(SHA384_H6), cpu_to_be64(SHA384_H7) +}; + +static const __be64 sha512_iv[8] = { + cpu_to_be64(SHA512_H0), cpu_to_be64(SHA512_H1), + cpu_to_be64(SHA512_H2), cpu_to_be64(SHA512_H3), + cpu_to_be64(SHA512_H4), cpu_to_be64(SHA512_H5), + cpu_to_be64(SHA512_H6), cpu_to_be64(SHA512_H7) +}; + +static const __be32 sha512_224_iv[16] = { + cpu_to_be32(0xC8373D8CUL), cpu_to_be32(0xA24D5419UL), + cpu_to_be32(0x6699E173UL), cpu_to_be32(0xD6D4DC89UL), + cpu_to_be32(0xAEB7FA1DUL), cpu_to_be32(0x829CFF32UL), + cpu_to_be32(0x14D59D67UL), cpu_to_be32(0xCF9F2F58UL), + cpu_to_be32(0x692B6D0FUL), cpu_to_be32(0xA84DD47BUL), + cpu_to_be32(0x736FE377UL), cpu_to_be32(0x4289C404UL), + cpu_to_be32(0xA8859D3FUL), cpu_to_be32(0xC8361D6AUL), + cpu_to_be32(0xADE61211UL), cpu_to_be32(0xA192D691UL) +}; + +static const __be32 sha512_256_iv[16] = { + cpu_to_be32(0x94213122UL), cpu_to_be32(0x2CF72BFCUL), + cpu_to_be32(0xA35F559FUL), cpu_to_be32(0xC2644CC8UL), + cpu_to_be32(0x6BB89323UL), cpu_to_be32(0x51B1536FUL), + cpu_to_be32(0x19773896UL), cpu_to_be32(0xBDEA4059UL), + cpu_to_be32(0xE23E2896UL), cpu_to_be32(0xE3FF8EA8UL), + cpu_to_be32(0x251E5EBEUL), cpu_to_be32(0x92398653UL), + cpu_to_be32(0xFC99012BUL), cpu_to_be32(0xAAB8852CUL), + cpu_to_be32(0xDC2DB70EUL), cpu_to_be32(0xA22CC581UL) +}; + +/* The purpose of this padding is to ensure that the padded message is a + * multiple of 512 bits (SHA1/SHA224/SHA256) or 1024 bits (SHA384/SHA512). + * The bit "1" is appended at the end of the message followed by + * "padlen-1" zero bits. Then a 64 bits block (SHA1/SHA224/SHA256) or + * 128 bits block (SHA384/SHA512) equals to the message length in bits + * is appended. + * + * For SHA1/SHA224/SHA256, padlen is calculated as followed: + * - if message length < 56 bytes then padlen = 56 - message length + * - else padlen = 64 + 56 - message length + * + * For SHA384/SHA512, padlen is calculated as followed: + * - if message length < 112 bytes then padlen = 112 - message length + * - else padlen = 128 + 112 - message length + */ +static void aspeed_ahash_fill_padding(struct aspeed_hace_dev *hace_dev, + struct aspeed_sham_reqctx *rctx) +{ + unsigned int index, padlen; + __be64 bits[2]; + + AHASH_DBG(hace_dev, "rctx flags:0x%x\n", (u32)rctx->flags); + + switch (rctx->flags & SHA_FLAGS_MASK) { + case SHA_FLAGS_SHA1: + case SHA_FLAGS_SHA224: + case SHA_FLAGS_SHA256: + bits[0] = cpu_to_be64(rctx->digcnt[0] << 3); + index = rctx->bufcnt & 0x3f; + padlen = (index < 56) ? (56 - index) : ((64 + 56) - index); + *(rctx->buffer + rctx->bufcnt) = 0x80; + memset(rctx->buffer + rctx->bufcnt + 1, 0, padlen - 1); + memcpy(rctx->buffer + rctx->bufcnt + padlen, bits, 8); + rctx->bufcnt += padlen + 8; + break; + default: + bits[1] = cpu_to_be64(rctx->digcnt[0] << 3); + bits[0] = cpu_to_be64(rctx->digcnt[1] << 3 | + rctx->digcnt[0] >> 61); + index = rctx->bufcnt & 0x7f; + padlen = (index < 112) ? (112 - index) : ((128 + 112) - index); + *(rctx->buffer + rctx->bufcnt) = 0x80; + memset(rctx->buffer + rctx->bufcnt + 1, 0, padlen - 1); + memcpy(rctx->buffer + rctx->bufcnt + padlen, bits, 16); + rctx->bufcnt += padlen + 16; + break; + } +} + +/* + * Prepare DMA buffer before hardware engine + * processing. + */ +static int aspeed_ahash_dma_prepare(struct aspeed_hace_dev *hace_dev) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + int length, remain; + + length = rctx->total + rctx->bufcnt; + remain = length % rctx->block_size; + + AHASH_DBG(hace_dev, "length:0x%x, remain:0x%x\n", length, remain); + + if (rctx->bufcnt) + memcpy(hash_engine->ahash_src_addr, rctx->buffer, rctx->bufcnt); + + if (rctx->total + rctx->bufcnt < ASPEED_CRYPTO_SRC_DMA_BUF_LEN) { + scatterwalk_map_and_copy(hash_engine->ahash_src_addr + + rctx->bufcnt, rctx->src_sg, + rctx->offset, rctx->total - remain, 0); + rctx->offset += rctx->total - remain; + + } else { + dev_warn(hace_dev->dev, "Hash data length is too large\n"); + return -EINVAL; + } + + scatterwalk_map_and_copy(rctx->buffer, rctx->src_sg, + rctx->offset, remain, 0); + + rctx->bufcnt = remain; + rctx->digest_dma_addr = dma_map_single(hace_dev->dev, rctx->digest, + SHA512_DIGEST_SIZE, + DMA_BIDIRECTIONAL); + if (dma_mapping_error(hace_dev->dev, rctx->digest_dma_addr)) { + dev_warn(hace_dev->dev, "dma_map() rctx digest error\n"); + return -ENOMEM; + } + + hash_engine->src_length = length - remain; + hash_engine->src_dma = hash_engine->ahash_src_dma_addr; + hash_engine->digest_dma = rctx->digest_dma_addr; + + return 0; +} + +/* + * Prepare DMA buffer as SG list buffer before + * hardware engine processing. + */ +static int aspeed_ahash_dma_prepare_sg(struct aspeed_hace_dev *hace_dev) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + struct aspeed_sg_list *src_list; + struct scatterlist *s; + int length, remain, sg_len, i; + int rc = 0; + + remain = (rctx->total + rctx->bufcnt) % rctx->block_size; + length = rctx->total + rctx->bufcnt - remain; + + AHASH_DBG(hace_dev, "%s:0x%x, %s:0x%x, %s:0x%x, %s:0x%x\n", + "rctx total", rctx->total, "bufcnt", rctx->bufcnt, + "length", length, "remain", remain); + + sg_len = dma_map_sg(hace_dev->dev, rctx->src_sg, rctx->src_nents, + DMA_TO_DEVICE); + if (!sg_len) { + dev_warn(hace_dev->dev, "dma_map_sg() src error\n"); + rc = -ENOMEM; + goto end; + } + + src_list = (struct aspeed_sg_list *)hash_engine->ahash_src_addr; + rctx->digest_dma_addr = dma_map_single(hace_dev->dev, rctx->digest, + SHA512_DIGEST_SIZE, + DMA_BIDIRECTIONAL); + if (dma_mapping_error(hace_dev->dev, rctx->digest_dma_addr)) { + dev_warn(hace_dev->dev, "dma_map() rctx digest error\n"); + rc = -ENOMEM; + goto free_src_sg; + } + + if (rctx->bufcnt != 0) { + rctx->buffer_dma_addr = dma_map_single(hace_dev->dev, + rctx->buffer, + rctx->block_size * 2, + DMA_TO_DEVICE); + if (dma_mapping_error(hace_dev->dev, rctx->buffer_dma_addr)) { + dev_warn(hace_dev->dev, "dma_map() rctx buffer error\n"); + rc = -ENOMEM; + goto free_rctx_digest; + } + + src_list[0].phy_addr = rctx->buffer_dma_addr; + src_list[0].len = rctx->bufcnt; + length -= src_list[0].len; + + /* Last sg list */ + if (length == 0) + src_list[0].len |= HASH_SG_LAST_LIST; + + src_list[0].phy_addr = cpu_to_le32(src_list[0].phy_addr); + src_list[0].len = cpu_to_le32(src_list[0].len); + src_list++; + } + + if (length != 0) { + for_each_sg(rctx->src_sg, s, sg_len, i) { + src_list[i].phy_addr = sg_dma_address(s); + + if (length > sg_dma_len(s)) { + src_list[i].len = sg_dma_len(s); + length -= sg_dma_len(s); + + } else { + /* Last sg list */ + src_list[i].len = length; + src_list[i].len |= HASH_SG_LAST_LIST; + length = 0; + } + + src_list[i].phy_addr = cpu_to_le32(src_list[i].phy_addr); + src_list[i].len = cpu_to_le32(src_list[i].len); + } + } + + if (length != 0) { + rc = -EINVAL; + goto free_rctx_buffer; + } + + rctx->offset = rctx->total - remain; + hash_engine->src_length = rctx->total + rctx->bufcnt - remain; + hash_engine->src_dma = hash_engine->ahash_src_dma_addr; + hash_engine->digest_dma = rctx->digest_dma_addr; + + return 0; + +free_rctx_buffer: + if (rctx->bufcnt != 0) + dma_unmap_single(hace_dev->dev, rctx->buffer_dma_addr, + rctx->block_size * 2, DMA_TO_DEVICE); +free_rctx_digest: + dma_unmap_single(hace_dev->dev, rctx->digest_dma_addr, + SHA512_DIGEST_SIZE, DMA_BIDIRECTIONAL); +free_src_sg: + dma_unmap_sg(hace_dev->dev, rctx->src_sg, rctx->src_nents, + DMA_TO_DEVICE); +end: + return rc; +} + +static int aspeed_ahash_complete(struct aspeed_hace_dev *hace_dev) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + + AHASH_DBG(hace_dev, "\n"); + + hash_engine->flags &= ~CRYPTO_FLAGS_BUSY; + + crypto_finalize_hash_request(hace_dev->crypt_engine_hash, req, 0); + + return 0; +} + +/* + * Copy digest to the corresponding request result. + * This function will be called at final() stage. + */ +static int aspeed_ahash_transfer(struct aspeed_hace_dev *hace_dev) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + + AHASH_DBG(hace_dev, "\n"); + + dma_unmap_single(hace_dev->dev, rctx->digest_dma_addr, + SHA512_DIGEST_SIZE, DMA_BIDIRECTIONAL); + + dma_unmap_single(hace_dev->dev, rctx->buffer_dma_addr, + rctx->block_size * 2, DMA_TO_DEVICE); + + memcpy(req->result, rctx->digest, rctx->digsize); + + return aspeed_ahash_complete(hace_dev); +} + +/* + * Trigger hardware engines to do the math. + */ +static int aspeed_hace_ahash_trigger(struct aspeed_hace_dev *hace_dev, + aspeed_hace_fn_t resume) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + + AHASH_DBG(hace_dev, "src_dma:0x%x, digest_dma:0x%x, length:0x%x\n", + hash_engine->src_dma, hash_engine->digest_dma, + hash_engine->src_length); + + rctx->cmd |= HASH_CMD_INT_ENABLE; + hash_engine->resume = resume; + + ast_hace_write(hace_dev, hash_engine->src_dma, ASPEED_HACE_HASH_SRC); + ast_hace_write(hace_dev, hash_engine->digest_dma, + ASPEED_HACE_HASH_DIGEST_BUFF); + ast_hace_write(hace_dev, hash_engine->digest_dma, + ASPEED_HACE_HASH_KEY_BUFF); + ast_hace_write(hace_dev, hash_engine->src_length, + ASPEED_HACE_HASH_DATA_LEN); + + /* Memory barrier to ensure all data setup before engine starts */ + mb(); + + ast_hace_write(hace_dev, rctx->cmd, ASPEED_HACE_HASH_CMD); + + return -EINPROGRESS; +} + +/* + * HMAC resume aims to do the second pass produces + * the final HMAC code derived from the inner hash + * result and the outer key. + */ +static int aspeed_ahash_hmac_resume(struct aspeed_hace_dev *hace_dev) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + struct crypto_ahash *tfm = crypto_ahash_reqtfm(req); + struct aspeed_sham_ctx *tctx = crypto_ahash_ctx(tfm); + struct aspeed_sha_hmac_ctx *bctx = tctx->base; + int rc = 0; + + AHASH_DBG(hace_dev, "\n"); + + dma_unmap_single(hace_dev->dev, rctx->digest_dma_addr, + SHA512_DIGEST_SIZE, DMA_BIDIRECTIONAL); + + dma_unmap_single(hace_dev->dev, rctx->buffer_dma_addr, + rctx->block_size * 2, DMA_TO_DEVICE); + + /* o key pad + hash sum 1 */ + memcpy(rctx->buffer, bctx->opad, rctx->block_size); + memcpy(rctx->buffer + rctx->block_size, rctx->digest, rctx->digsize); + + rctx->bufcnt = rctx->block_size + rctx->digsize; + rctx->digcnt[0] = rctx->block_size + rctx->digsize; + + aspeed_ahash_fill_padding(hace_dev, rctx); + memcpy(rctx->digest, rctx->sha_iv, rctx->ivsize); + + rctx->digest_dma_addr = dma_map_single(hace_dev->dev, rctx->digest, + SHA512_DIGEST_SIZE, + DMA_BIDIRECTIONAL); + if (dma_mapping_error(hace_dev->dev, rctx->digest_dma_addr)) { + dev_warn(hace_dev->dev, "dma_map() rctx digest error\n"); + rc = -ENOMEM; + goto end; + } + + rctx->buffer_dma_addr = dma_map_single(hace_dev->dev, rctx->buffer, + rctx->block_size * 2, + DMA_TO_DEVICE); + if (dma_mapping_error(hace_dev->dev, rctx->buffer_dma_addr)) { + dev_warn(hace_dev->dev, "dma_map() rctx buffer error\n"); + rc = -ENOMEM; + goto free_rctx_digest; + } + + hash_engine->src_dma = rctx->buffer_dma_addr; + hash_engine->src_length = rctx->bufcnt; + hash_engine->digest_dma = rctx->digest_dma_addr; + + return aspeed_hace_ahash_trigger(hace_dev, aspeed_ahash_transfer); + +free_rctx_digest: + dma_unmap_single(hace_dev->dev, rctx->digest_dma_addr, + SHA512_DIGEST_SIZE, DMA_BIDIRECTIONAL); +end: + return rc; +} + +static int aspeed_ahash_req_final(struct aspeed_hace_dev *hace_dev) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + int rc = 0; + + AHASH_DBG(hace_dev, "\n"); + + aspeed_ahash_fill_padding(hace_dev, rctx); + + rctx->digest_dma_addr = dma_map_single(hace_dev->dev, + rctx->digest, + SHA512_DIGEST_SIZE, + DMA_BIDIRECTIONAL); + if (dma_mapping_error(hace_dev->dev, rctx->digest_dma_addr)) { + dev_warn(hace_dev->dev, "dma_map() rctx digest error\n"); + rc = -ENOMEM; + goto end; + } + + rctx->buffer_dma_addr = dma_map_single(hace_dev->dev, + rctx->buffer, + rctx->block_size * 2, + DMA_TO_DEVICE); + if (dma_mapping_error(hace_dev->dev, rctx->buffer_dma_addr)) { + dev_warn(hace_dev->dev, "dma_map() rctx buffer error\n"); + rc = -ENOMEM; + goto free_rctx_digest; + } + + hash_engine->src_dma = rctx->buffer_dma_addr; + hash_engine->src_length = rctx->bufcnt; + hash_engine->digest_dma = rctx->digest_dma_addr; + + if (rctx->flags & SHA_FLAGS_HMAC) + return aspeed_hace_ahash_trigger(hace_dev, + aspeed_ahash_hmac_resume); + + return aspeed_hace_ahash_trigger(hace_dev, aspeed_ahash_transfer); + +free_rctx_digest: + dma_unmap_single(hace_dev->dev, rctx->digest_dma_addr, + SHA512_DIGEST_SIZE, DMA_BIDIRECTIONAL); +end: + return rc; +} + +static int aspeed_ahash_update_resume_sg(struct aspeed_hace_dev *hace_dev) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + + AHASH_DBG(hace_dev, "\n"); + + dma_unmap_sg(hace_dev->dev, rctx->src_sg, rctx->src_nents, + DMA_TO_DEVICE); + + if (rctx->bufcnt != 0) + dma_unmap_single(hace_dev->dev, rctx->buffer_dma_addr, + rctx->block_size * 2, + DMA_TO_DEVICE); + + dma_unmap_single(hace_dev->dev, rctx->digest_dma_addr, + SHA512_DIGEST_SIZE, DMA_BIDIRECTIONAL); + + scatterwalk_map_and_copy(rctx->buffer, rctx->src_sg, rctx->offset, + rctx->total - rctx->offset, 0); + + rctx->bufcnt = rctx->total - rctx->offset; + rctx->cmd &= ~HASH_CMD_HASH_SRC_SG_CTRL; + + if (rctx->flags & SHA_FLAGS_FINUP) + return aspeed_ahash_req_final(hace_dev); + + return aspeed_ahash_complete(hace_dev); +} + +static int aspeed_ahash_update_resume(struct aspeed_hace_dev *hace_dev) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + + AHASH_DBG(hace_dev, "\n"); + + dma_unmap_single(hace_dev->dev, rctx->digest_dma_addr, + SHA512_DIGEST_SIZE, DMA_BIDIRECTIONAL); + + if (rctx->flags & SHA_FLAGS_FINUP) + return aspeed_ahash_req_final(hace_dev); + + return aspeed_ahash_complete(hace_dev); +} + +static int aspeed_ahash_req_update(struct aspeed_hace_dev *hace_dev) +{ + struct aspeed_engine_hash *hash_engine = &hace_dev->hash_engine; + struct ahash_request *req = hash_engine->req; + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + aspeed_hace_fn_t resume; + int ret; + + AHASH_DBG(hace_dev, "\n"); + + if (hace_dev->version == AST2600_VERSION) { + rctx->cmd |= HASH_CMD_HASH_SRC_SG_CTRL; + resume = aspeed_ahash_update_resume_sg; + + } else { + resume = aspeed_ahash_update_resume; + } + + ret = hash_engine->dma_prepare(hace_dev); + if (ret) + return ret; + + return aspeed_hace_ahash_trigger(hace_dev, resume); +} + +static int aspeed_hace_hash_handle_queue(struct aspeed_hace_dev *hace_dev, + struct ahash_request *req) +{ + return crypto_transfer_hash_request_to_engine( + hace_dev->crypt_engine_hash, req); +} + +static int aspeed_ahash_do_request(struct crypto_engine *engine, void *areq) +{ + struct ahash_request *req = ahash_request_cast(areq); + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + struct crypto_ahash *tfm = crypto_ahash_reqtfm(req); + struct aspeed_sham_ctx *tctx = crypto_ahash_ctx(tfm); + struct aspeed_hace_dev *hace_dev = tctx->hace_dev; + struct aspeed_engine_hash *hash_engine; + int ret = 0; + + hash_engine = &hace_dev->hash_engine; + hash_engine->flags |= CRYPTO_FLAGS_BUSY; + + if (rctx->op == SHA_OP_UPDATE) + ret = aspeed_ahash_req_update(hace_dev); + else if (rctx->op == SHA_OP_FINAL) + ret = aspeed_ahash_req_final(hace_dev); + + if (ret != -EINPROGRESS) + return ret; + + return 0; +} + +static int aspeed_ahash_prepare_request(struct crypto_engine *engine, + void *areq) +{ + struct ahash_request *req = ahash_request_cast(areq); + struct crypto_ahash *tfm = crypto_ahash_reqtfm(req); + struct aspeed_sham_ctx *tctx = crypto_ahash_ctx(tfm); + struct aspeed_hace_dev *hace_dev = tctx->hace_dev; + struct aspeed_engine_hash *hash_engine; + + hash_engine = &hace_dev->hash_engine; + hash_engine->req = req; + + if (hace_dev->version == AST2600_VERSION) + hash_engine->dma_prepare = aspeed_ahash_dma_prepare_sg; + else + hash_engine->dma_prepare = aspeed_ahash_dma_prepare; + + return 0; +} + +static int aspeed_sham_update(struct ahash_request *req) +{ + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + struct crypto_ahash *tfm = crypto_ahash_reqtfm(req); + struct aspeed_sham_ctx *tctx = crypto_ahash_ctx(tfm); + struct aspeed_hace_dev *hace_dev = tctx->hace_dev; + + AHASH_DBG(hace_dev, "req->nbytes: %d\n", req->nbytes); + + rctx->total = req->nbytes; + rctx->src_sg = req->src; + rctx->offset = 0; + rctx->src_nents = sg_nents(req->src); + rctx->op = SHA_OP_UPDATE; + + rctx->digcnt[0] += rctx->total; + if (rctx->digcnt[0] < rctx->total) + rctx->digcnt[1]++; + + if (rctx->bufcnt + rctx->total < rctx->block_size) { + scatterwalk_map_and_copy(rctx->buffer + rctx->bufcnt, + rctx->src_sg, rctx->offset, + rctx->total, 0); + rctx->bufcnt += rctx->total; + + return 0; + } + + return aspeed_hace_hash_handle_queue(hace_dev, req); +} + +static int aspeed_sham_shash_digest(struct crypto_shash *tfm, u32 flags, + const u8 *data, unsigned int len, u8 *out) +{ + SHASH_DESC_ON_STACK(shash, tfm); + + shash->tfm = tfm; + + return crypto_shash_digest(shash, data, len, out); +} + +static int aspeed_sham_final(struct ahash_request *req) +{ + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + struct crypto_ahash *tfm = crypto_ahash_reqtfm(req); + struct aspeed_sham_ctx *tctx = crypto_ahash_ctx(tfm); + struct aspeed_hace_dev *hace_dev = tctx->hace_dev; + + AHASH_DBG(hace_dev, "req->nbytes:%d, rctx->total:%d\n", + req->nbytes, rctx->total); + rctx->op = SHA_OP_FINAL; + + return aspeed_hace_hash_handle_queue(hace_dev, req); +} + +static int aspeed_sham_finup(struct ahash_request *req) +{ + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + struct crypto_ahash *tfm = crypto_ahash_reqtfm(req); + struct aspeed_sham_ctx *tctx = crypto_ahash_ctx(tfm); + struct aspeed_hace_dev *hace_dev = tctx->hace_dev; + int rc1, rc2; + + AHASH_DBG(hace_dev, "req->nbytes: %d\n", req->nbytes); + + rctx->flags |= SHA_FLAGS_FINUP; + + rc1 = aspeed_sham_update(req); + if (rc1 == -EINPROGRESS || rc1 == -EBUSY) + return rc1; + + /* + * final() has to be always called to cleanup resources + * even if update() failed, except EINPROGRESS + */ + rc2 = aspeed_sham_final(req); + + return rc1 ? : rc2; +} + +static int aspeed_sham_init(struct ahash_request *req) +{ + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + struct crypto_ahash *tfm = crypto_ahash_reqtfm(req); + struct aspeed_sham_ctx *tctx = crypto_ahash_ctx(tfm); + struct aspeed_hace_dev *hace_dev = tctx->hace_dev; + struct aspeed_sha_hmac_ctx *bctx = tctx->base; + + AHASH_DBG(hace_dev, "%s: digest size:%d\n", + crypto_tfm_alg_name(&tfm->base), + crypto_ahash_digestsize(tfm)); + + rctx->cmd = HASH_CMD_ACC_MODE; + rctx->flags = 0; + + switch (crypto_ahash_digestsize(tfm)) { + case SHA1_DIGEST_SIZE: + rctx->cmd |= HASH_CMD_SHA1 | HASH_CMD_SHA_SWAP; + rctx->flags |= SHA_FLAGS_SHA1; + rctx->digsize = SHA1_DIGEST_SIZE; + rctx->block_size = SHA1_BLOCK_SIZE; + rctx->sha_iv = sha1_iv; + rctx->ivsize = 32; + memcpy(rctx->digest, sha1_iv, rctx->ivsize); + break; + case SHA224_DIGEST_SIZE: + rctx->cmd |= HASH_CMD_SHA224 | HASH_CMD_SHA_SWAP; + rctx->flags |= SHA_FLAGS_SHA224; + rctx->digsize = SHA224_DIGEST_SIZE; + rctx->block_size = SHA224_BLOCK_SIZE; + rctx->sha_iv = sha224_iv; + rctx->ivsize = 32; + memcpy(rctx->digest, sha224_iv, rctx->ivsize); + break; + case SHA256_DIGEST_SIZE: + rctx->cmd |= HASH_CMD_SHA256 | HASH_CMD_SHA_SWAP; + rctx->flags |= SHA_FLAGS_SHA256; + rctx->digsize = SHA256_DIGEST_SIZE; + rctx->block_size = SHA256_BLOCK_SIZE; + rctx->sha_iv = sha256_iv; + rctx->ivsize = 32; + memcpy(rctx->digest, sha256_iv, rctx->ivsize); + break; + case SHA384_DIGEST_SIZE: + rctx->cmd |= HASH_CMD_SHA512_SER | HASH_CMD_SHA384 | + HASH_CMD_SHA_SWAP; + rctx->flags |= SHA_FLAGS_SHA384; + rctx->digsize = SHA384_DIGEST_SIZE; + rctx->block_size = SHA384_BLOCK_SIZE; + rctx->sha_iv = (const __be32 *)sha384_iv; + rctx->ivsize = 64; + memcpy(rctx->digest, sha384_iv, rctx->ivsize); + break; + case SHA512_DIGEST_SIZE: + rctx->cmd |= HASH_CMD_SHA512_SER | HASH_CMD_SHA512 | + HASH_CMD_SHA_SWAP; + rctx->flags |= SHA_FLAGS_SHA512; + rctx->digsize = SHA512_DIGEST_SIZE; + rctx->block_size = SHA512_BLOCK_SIZE; + rctx->sha_iv = (const __be32 *)sha512_iv; + rctx->ivsize = 64; + memcpy(rctx->digest, sha512_iv, rctx->ivsize); + break; + default: + dev_warn(tctx->hace_dev->dev, "digest size %d not support\n", + crypto_ahash_digestsize(tfm)); + return -EINVAL; + } + + rctx->bufcnt = 0; + rctx->total = 0; + rctx->digcnt[0] = 0; + rctx->digcnt[1] = 0; + + /* HMAC init */ + if (tctx->flags & SHA_FLAGS_HMAC) { + rctx->digcnt[0] = rctx->block_size; + rctx->bufcnt = rctx->block_size; + memcpy(rctx->buffer, bctx->ipad, rctx->block_size); + rctx->flags |= SHA_FLAGS_HMAC; + } + + return 0; +} + +static int aspeed_sha512s_init(struct ahash_request *req) +{ + struct aspeed_sham_reqctx *rctx = ahash_request_ctx(req); + struct crypto_ahash *tfm = crypto_ahash_reqtfm(req); + struct aspeed_sham_ctx *tctx = crypto_ahash_ctx(tfm); + struct aspeed_hace_dev *hace_dev = tctx->hace_dev; + struct aspeed_sha_hmac_ctx *bctx = tctx->base; + + AHASH_DBG(hace_dev, "digest size: %d\n", crypto_ahash_digestsize(tfm)); + + rctx->cmd = HASH_CMD_ACC_MODE; + rctx->flags = 0; + + switch (crypto_ahash_digestsize(tfm)) { + case SHA224_DIGEST_SIZE: + rctx->cmd |= HASH_CMD_SHA512_SER | HASH_CMD_SHA512_224 | + HASH_CMD_SHA_SWAP; + rctx->flags |= SHA_FLAGS_SHA512_224; + rctx->digsize = SHA224_DIGEST_SIZE; + rctx->block_size = SHA512_BLOCK_SIZE; + rctx->sha_iv = sha512_224_iv; + rctx->ivsize = 64; + memcpy(rctx->digest, sha512_224_iv, rctx->ivsize); + break; + case SHA256_DIGEST_SIZE: + rctx->cmd |= HASH_CMD_SHA512_SER | HASH_CMD_SHA512_256 | + HASH_CMD_SHA_SWAP; + rctx->flags |= SHA_FLAGS_SHA512_256; + rctx->digsize = SHA256_DIGEST_SIZE; + rctx->block_size = SHA512_BLOCK_SIZE; + rctx->sha_iv = sha512_256_iv; + rctx->ivsize = 64; + memcpy(rctx->digest, sha512_256_iv, rctx->ivsize); + break; + default: + dev_warn(tctx->hace_dev->dev, "digest size %d not support\n", + crypto_ahash_digestsize(tfm)); + return -EINVAL; + } + + rctx->bufcnt = 0; + rctx->total = 0; + rctx->digcnt[0] = 0; + rctx->digcnt[1] = 0; + + /* HMAC init */ + if (tctx->flags & SHA_FLAGS_HMAC) { + rctx->digcnt[0] = rctx->block_size; + rctx->bufcnt = rctx->block_size; + memcpy(rctx->buffer, bctx->ipad, rctx->block_size); + rctx->flags |= SHA_FLAGS_HMAC; + } + + return 0; +} + +static int aspeed_sham_digest(struct ahash_request *req) +{ + return aspeed_sham_init(req) ? : aspeed_sham_finup(req); +} + +static int aspeed_sham_setkey(struct crypto_ahash *tfm, const u8 *key, + unsigned int keylen) +{ + struct aspeed_sham_ctx *tctx = crypto_ahash_ctx(tfm); + struct aspeed_hace_dev *hace_dev = tctx->hace_dev; + struct aspeed_sha_hmac_ctx *bctx = tctx->base; + int ds = crypto_shash_digestsize(bctx->shash); + int bs = crypto_shash_blocksize(bctx->shash); + int err = 0; + int i; + + AHASH_DBG(hace_dev, "%s: keylen:%d\n", crypto_tfm_alg_name(&tfm->base), + keylen); + + if (keylen > bs) { + err = aspeed_sham_shash_digest(bctx->shash, + crypto_shash_get_flags(bctx->shash), + key, keylen, bctx->ipad); + if (err) + return err; + keylen = ds; + + } else { + memcpy(bctx->ipad, key, keylen); + } + + memset(bctx->ipad + keylen, 0, bs - keylen); + memcpy(bctx->opad, bctx->ipad, bs); + + for (i = 0; i < bs; i++) { + bctx->ipad[i] ^= HMAC_IPAD_VALUE; + bctx->opad[i] ^= HMAC_OPAD_VALUE; + } + + return err; +} + +static int aspeed_sham_cra_init(struct crypto_tfm *tfm) +{ + struct ahash_alg *alg = __crypto_ahash_alg(tfm->__crt_alg); + struct aspeed_sham_ctx *tctx = crypto_tfm_ctx(tfm); + struct aspeed_hace_alg *ast_alg; + + ast_alg = container_of(alg, struct aspeed_hace_alg, alg.ahash); + tctx->hace_dev = ast_alg->hace_dev; + tctx->flags = 0; + + crypto_ahash_set_reqsize(__crypto_ahash_cast(tfm), + sizeof(struct aspeed_sham_reqctx)); + + if (ast_alg->alg_base) { + /* hmac related */ + struct aspeed_sha_hmac_ctx *bctx = tctx->base; + + tctx->flags |= SHA_FLAGS_HMAC; + bctx->shash = crypto_alloc_shash(ast_alg->alg_base, 0, + CRYPTO_ALG_NEED_FALLBACK); + if (IS_ERR(bctx->shash)) { + dev_warn(ast_alg->hace_dev->dev, + "base driver '%s' could not be loaded.\n", + ast_alg->alg_base); + return PTR_ERR(bctx->shash); + } + } + </cut>

2 years, 3 months

[TCWG CI] Failure after v5.19-69-gcae15c2ed8e6: vdpa/mlx5: Implement susupend virtqueue callback

by ci_notify＠linaro.org

Failure after v5.19-69-gcae15c2ed8e6: vdpa/mlx5: Implement susupend virtqueue callback: Results changed to -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 20673 # First few build errors in logs: # 00:04:45 drivers/vdpa/mlx5/net/mlx5_vnet.c:2829:10: error: ‘const struct vdpa_config_ops’ has no member named ‘suspend’ # 00:04:45 drivers/vdpa/mlx5/net/mlx5_vnet.c:2829:20: error: positional initialization of field in ‘struct’ declared with ‘designated_init’ attribute [-Werror=designated-init] # 00:04:45 drivers/vdpa/mlx5/net/mlx5_vnet.c:2829:20: error: initialization of ‘int (*)(struct vdpa_device *, u16, u64, u64, u64)’ {aka ‘int (*)(struct vdpa_device *, short unsigned int, long long unsigned int, long long unsigned int, long long unsigned int)’} from incompatible pointer type ‘int (*)(struct vdpa_device *)’ [-Werror=incompatible-pointer-types] # 00:04:45 make[3]: *** [drivers/vdpa/mlx5/net/mlx5_vnet.o] Error 1 # 00:04:51 make[2]: *** [drivers/vdpa/mlx5] Error 2 # 00:04:57 make[1]: *** [drivers/vdpa] Error 2 # 00:32:05 make: *** [drivers] Error 2 from -10 # build_abe binutils: -9 # build_abe stage1: -5 # build_abe qemu: -2 # linux_n_obj: 20764 # linux build successful: all THIS IS THE END OF INTERESTING STUFF. BELOW ARE LINKS TO BUILDS, REPRODUCTION INSTRUCTIONS, AND THE RAW COMMIT. For latest status see comments in https://linaro.atlassian.net/browse/GNU-681 . Status of v5.19-69-gcae15c2ed8e6 commit for tcwg_kernel: commit cae15c2ed8e6e058bd5e32de292ab7982640161e Author: Eli Cohen <elic(a)nvidia.com> Date: Thu Jul 14 14:39:26 2022 +0300 vdpa/mlx5: Implement susupend virtqueue callback Implement the suspend callback allowing to suspend the virtqueues so they stop processing descriptors. This is required to allow to query a consistent state of the virtqueue while live migration is taking place. Signed-off-by: Eli Cohen <elic(a)nvidia.com> Message-Id: <20220714113927.85729-2-elic(a)nvidia.com> Signed-off-by: Michael S. Tsirkin <mst(a)redhat.com> * gnu-release-arm-norov-allyesconfig ** Failure after v5.19-69-gcae15c2ed8e6: vdpa/mlx5: Implement susupend virtqueue callback: ** https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-norov-allye… Bad build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-norov-allye… Good build: https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-norov-allye… Reproduce current build: <cut> mkdir -p investigate-linux-cae15c2ed8e6e058bd5e32de292ab7982640161e cd investigate-linux-cae15c2ed8e6e058bd5e32de292ab7982640161e # Fetch scripts git clone https://git.linaro.org/toolchain/jenkins-scripts # Fetch manifests for bad and good builds mkdir -p bad/artifacts good/artifacts curl -o bad/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-norov-allye… --fail curl -o good/artifacts/manifest.sh https://ci.linaro.org/job/tcwg_kernel-gnu-build-gnu-release-arm-norov-allye… --fail # Reproduce bad build (cd bad; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) # Reproduce good build (cd good; ../jenkins-scripts/tcwg_kernel-build.sh ^^ true %%rr[top_artifacts] artifacts) </cut> Full commit (up to 1000 lines): <cut> commit cae15c2ed8e6e058bd5e32de292ab7982640161e Author: Eli Cohen <elic(a)nvidia.com> Date: Thu Jul 14 14:39:26 2022 +0300 vdpa/mlx5: Implement susupend virtqueue callback Implement the suspend callback allowing to suspend the virtqueues so they stop processing descriptors. This is required to allow to query a consistent state of the virtqueue while live migration is taking place. Signed-off-by: Eli Cohen <elic(a)nvidia.com> Message-Id: <20220714113927.85729-2-elic(a)nvidia.com> Signed-off-by: Michael S. Tsirkin <mst(a)redhat.com> --- drivers/vdpa/mlx5/net/mlx5_vnet.c | 83 ++++++++++++++++++++++++++++++++++++-- include/linux/mlx5/mlx5_ifc_vdpa.h | 8 ++++ 2 files changed, 88 insertions(+), 3 deletions(-) diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c index 99bbbf38c8b1..a476e06476f6 100644 --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c @@ -164,6 +164,7 @@ struct mlx5_vdpa_net { bool setup; u32 cur_num_vqs; u32 rqt_size; + bool nb_registered; struct notifier_block nb; struct vdpa_callback config_cb; struct mlx5_vdpa_wq_ent cvq_ent; @@ -895,6 +896,7 @@ static int create_virtqueue(struct mlx5_vdpa_net *ndev, struct mlx5_vdpa_virtque if (err) goto err_cmd; + mvq->fw_state = MLX5_VIRTIO_NET_Q_OBJECT_STATE_INIT; kfree(in); mvq->virtq_id = MLX5_GET(general_obj_out_cmd_hdr, out, obj_id); @@ -922,6 +924,7 @@ static void destroy_virtqueue(struct mlx5_vdpa_net *ndev, struct mlx5_vdpa_virtq mlx5_vdpa_warn(&ndev->mvdev, "destroy virtqueue 0x%x\n", mvq->virtq_id); return; } + mvq->fw_state = MLX5_VIRTIO_NET_Q_OBJECT_NONE; umems_destroy(ndev, mvq); } @@ -1121,6 +1124,20 @@ static int query_virtqueue(struct mlx5_vdpa_net *ndev, struct mlx5_vdpa_virtqueu return err; } +static bool is_valid_state_change(int oldstate, int newstate) +{ + switch (oldstate) { + case MLX5_VIRTIO_NET_Q_OBJECT_STATE_INIT: + return newstate == MLX5_VIRTIO_NET_Q_OBJECT_STATE_RDY; + case MLX5_VIRTIO_NET_Q_OBJECT_STATE_RDY: + return newstate == MLX5_VIRTIO_NET_Q_OBJECT_STATE_SUSPEND; + case MLX5_VIRTIO_NET_Q_OBJECT_STATE_SUSPEND: + case MLX5_VIRTIO_NET_Q_OBJECT_STATE_ERR: + default: + return false; + } +} + static int modify_virtqueue(struct mlx5_vdpa_net *ndev, struct mlx5_vdpa_virtqueue *mvq, int state) { int inlen = MLX5_ST_SZ_BYTES(modify_virtio_net_q_in); @@ -1130,6 +1147,12 @@ static int modify_virtqueue(struct mlx5_vdpa_net *ndev, struct mlx5_vdpa_virtque void *in; int err; + if (mvq->fw_state == MLX5_VIRTIO_NET_Q_OBJECT_NONE) + return 0; + + if (!is_valid_state_change(mvq->fw_state, state)) + return -EINVAL; + in = kzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -1992,6 +2015,7 @@ static void mlx5_vdpa_set_vq_ready(struct vdpa_device *vdev, u16 idx, bool ready struct mlx5_vdpa_dev *mvdev = to_mvdev(vdev); struct mlx5_vdpa_net *ndev = to_mlx5_vdpa_ndev(mvdev); struct mlx5_vdpa_virtqueue *mvq; + int err; if (!mvdev->actual_features) return; @@ -2005,8 +2029,16 @@ static void mlx5_vdpa_set_vq_ready(struct vdpa_device *vdev, u16 idx, bool ready } mvq = &ndev->vqs[idx]; - if (!ready) + if (!ready) { suspend_vq(ndev, mvq); + } else { + err = modify_virtqueue(ndev, mvq, MLX5_VIRTIO_NET_Q_OBJECT_STATE_RDY); + if (err) { + mlx5_vdpa_warn(mvdev, "modify VQ %d to ready failed (%d)\n", idx, err); + ready = false; + } + } + mvq->ready = ready; } @@ -2733,6 +2765,37 @@ static int mlx5_vdpa_get_vendor_vq_stats(struct vdpa_device *vdev, u16 idx, return err; } +static void mlx5_vdpa_cvq_suspend(struct mlx5_vdpa_dev *mvdev) +{ + struct mlx5_control_vq *cvq; + + if (!(mvdev->actual_features & BIT_ULL(VIRTIO_NET_F_CTRL_VQ))) + return; + + cvq = &mvdev->cvq; + cvq->ready = false; +} + +static int mlx5_vdpa_suspend(struct vdpa_device *vdev) +{ + struct mlx5_vdpa_dev *mvdev = to_mvdev(vdev); + struct mlx5_vdpa_net *ndev = to_mlx5_vdpa_ndev(mvdev); + struct mlx5_vdpa_virtqueue *mvq; + int i; + + down_write(&ndev->reslock); + mlx5_notifier_unregister(mvdev->mdev, &ndev->nb); + ndev->nb_registered = false; + flush_workqueue(ndev->mvdev.wq); + for (i = 0; i < ndev->cur_num_vqs; i++) { + mvq = &ndev->vqs[i]; + suspend_vq(ndev, mvq); + } + mlx5_vdpa_cvq_suspend(mvdev); + up_write(&ndev->reslock); + return 0; +} + static const struct vdpa_config_ops mlx5_vdpa_ops = { .set_vq_address = mlx5_vdpa_set_vq_address, .set_vq_num = mlx5_vdpa_set_vq_num, @@ -2763,6 +2826,7 @@ static const struct vdpa_config_ops mlx5_vdpa_ops = { .get_generation = mlx5_vdpa_get_generation, .set_map = mlx5_vdpa_set_map, .free = mlx5_vdpa_free, + .suspend = mlx5_vdpa_suspend, }; static int query_mtu(struct mlx5_core_dev *mdev, u16 *mtu) @@ -2828,6 +2892,7 @@ static void init_mvqs(struct mlx5_vdpa_net *ndev) mvq->index = i; mvq->ndev = ndev; mvq->fwqp.fw = true; + mvq->fw_state = MLX5_VIRTIO_NET_Q_OBJECT_NONE; } for (; i < ndev->mvdev.max_vqs; i++) { mvq = &ndev->vqs[i]; @@ -2902,13 +2967,21 @@ static int event_handler(struct notifier_block *nb, unsigned long event, void *p switch (eqe->sub_type) { case MLX5_PORT_CHANGE_SUBTYPE_DOWN: case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE: + down_read(&ndev->reslock); + if (!ndev->nb_registered) { + up_read(&ndev->reslock); + return NOTIFY_DONE; + } wqent = kzalloc(sizeof(*wqent), GFP_ATOMIC); - if (!wqent) + if (!wqent) { + up_read(&ndev->reslock); return NOTIFY_DONE; + } wqent->mvdev = &ndev->mvdev; INIT_WORK(&wqent->work, update_carrier); queue_work(ndev->mvdev.wq, &wqent->work); + up_read(&ndev->reslock); ret = NOTIFY_OK; break; default: @@ -3062,6 +3135,7 @@ static int mlx5_vdpa_dev_add(struct vdpa_mgmt_dev *v_mdev, const char *name, ndev->nb.notifier_call = event_handler; mlx5_notifier_register(mdev, &ndev->nb); + ndev->nb_registered = true; mvdev->vdev.mdev = &mgtdev->mgtdev; err = _vdpa_register_device(&mvdev->vdev, max_vqs + 1); if (err) @@ -3093,7 +3167,10 @@ static void mlx5_vdpa_dev_del(struct vdpa_mgmt_dev *v_mdev, struct vdpa_device * struct mlx5_vdpa_net *ndev = to_mlx5_vdpa_ndev(mvdev); struct workqueue_struct *wq; - mlx5_notifier_unregister(mvdev->mdev, &ndev->nb); + if (ndev->nb_registered) { + mlx5_notifier_unregister(mvdev->mdev, &ndev->nb); + ndev->nb_registered = false; + } wq = mvdev->wq; mvdev->wq = NULL; destroy_workqueue(wq); diff --git a/include/linux/mlx5/mlx5_ifc_vdpa.h b/include/linux/mlx5/mlx5_ifc_vdpa.h index 4414ed5b6ed2..9becdc3fa503 100644 --- a/include/linux/mlx5/mlx5_ifc_vdpa.h +++ b/include/linux/mlx5/mlx5_ifc_vdpa.h @@ -150,6 +150,14 @@ enum { MLX5_VIRTIO_NET_Q_OBJECT_STATE_ERR = 0x3, }; +/* This indicates that the object was not created or has already + * been desroyed. It is very safe to assume that this object will never + * have so many states + */ +enum { + MLX5_VIRTIO_NET_Q_OBJECT_NONE = 0xffffffff +}; + enum { MLX5_RQTC_LIST_Q_TYPE_RQ = 0x0, MLX5_RQTC_LIST_Q_TYPE_VIRTIO_NET_Q = 0x1, </cut>

2 years, 3 months

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

linaro-kernel September 2022