From 44e70718df4fc2fadf1665eb9374df71aeda1f03 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Mon, 10 Feb 2025 22:54:02 +0000 Subject: [PATCH 01/16] KVM: SVM: Ensure PSP module is initialized if KVM module is built-in The kernel's initcall infrastructure lacks the ability to express dependencies between initcalls, whereas the modules infrastructure automatically handles dependencies via symbol loading. Ensure the PSP SEV driver is initialized before proceeding in sev_hardware_setup() if KVM is built-in as the dependency isn't handled by the initcall infrastructure. Signed-off-by: Sean Christopherson Reviewed-by: Tom Lendacky Signed-off-by: Ashish Kalra Message-ID: Signed-off-by: Paolo Bonzini --- arch/x86/kvm/svm/sev.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c index a2a794c32050..0dbb25442ec1 100644 --- a/arch/x86/kvm/svm/sev.c +++ b/arch/x86/kvm/svm/sev.c @@ -2972,6 +2972,16 @@ void __init sev_hardware_setup(void) WARN_ON_ONCE(!boot_cpu_has(X86_FEATURE_FLUSHBYASID))) goto out; + /* + * The kernel's initcall infrastructure lacks the ability to express + * dependencies between initcalls, whereas the modules infrastructure + * automatically handles dependencies via symbol loading. Ensure the + * PSP SEV driver is initialized before proceeding if KVM is built-in, + * as the dependency isn't handled by the initcall infrastructure. + */ + if (IS_BUILTIN(CONFIG_KVM_AMD) && sev_module_init()) + goto out; + /* Retrieve SEV CPUID information */ cpuid(0x8000001f, &eax, &ebx, &ecx, &edx); -- 2.51.0 From 409f45387c937145adeeeebc6d6032c2ec232b35 Mon Sep 17 00:00:00 2001 From: Ashish Kalra Date: Mon, 10 Feb 2025 22:54:18 +0000 Subject: [PATCH 02/16] x86/sev: Fix broken SNP support with KVM module built-in Fix issues with enabling SNP host support and effectively SNP support which is broken with respect to the KVM module being built-in. SNP host support is enabled in snp_rmptable_init() which is invoked as device_initcall(). SNP check on IOMMU is done during IOMMU PCI init (IOMMU_PCI_INIT stage). And for that reason snp_rmptable_init() is currently invoked via device_initcall() and cannot be invoked via subsys_initcall() as core IOMMU subsystem gets initialized via subsys_initcall(). Now, if kvm_amd module is built-in, it gets initialized before SNP host support is enabled in snp_rmptable_init() : [ 10.131811] kvm_amd: TSC scaling supported [ 10.136384] kvm_amd: Nested Virtualization enabled [ 10.141734] kvm_amd: Nested Paging enabled [ 10.146304] kvm_amd: LBR virtualization supported [ 10.151557] kvm_amd: SEV enabled (ASIDs 100 - 509) [ 10.156905] kvm_amd: SEV-ES enabled (ASIDs 1 - 99) [ 10.162256] kvm_amd: SEV-SNP enabled (ASIDs 1 - 99) [ 10.171508] kvm_amd: Virtual VMLOAD VMSAVE supported [ 10.177052] kvm_amd: Virtual GIF supported ... ... [ 10.201648] kvm_amd: in svm_enable_virtualization_cpu And then svm_x86_ops->enable_virtualization_cpu() (svm_enable_virtualization_cpu) programs MSR_VM_HSAVE_PA as following: wrmsrl(MSR_VM_HSAVE_PA, sd->save_area_pa); So VM_HSAVE_PA is non-zero before SNP support is enabled on all CPUs. snp_rmptable_init() gets invoked after svm_enable_virtualization_cpu() as following : ... [ 11.256138] kvm_amd: in svm_enable_virtualization_cpu ... [ 11.264918] SEV-SNP: in snp_rmptable_init This triggers a #GP exception in snp_rmptable_init() when snp_enable() is invoked to set SNP_EN in SYSCFG MSR: [ 11.294289] unchecked MSR access error: WRMSR to 0xc0010010 (tried to write 0x0000000003fc0000) at rIP: 0xffffffffaf5d5c28 (native_write_msr+0x8/0x30) ... [ 11.294404] Call Trace: [ 11.294482] [ 11.294513] ? show_stack_regs+0x26/0x30 [ 11.294522] ? ex_handler_msr+0x10f/0x180 [ 11.294529] ? search_extable+0x2b/0x40 [ 11.294538] ? fixup_exception+0x2dd/0x340 [ 11.294542] ? exc_general_protection+0x14f/0x440 [ 11.294550] ? asm_exc_general_protection+0x2b/0x30 [ 11.294557] ? __pfx_snp_enable+0x10/0x10 [ 11.294567] ? native_write_msr+0x8/0x30 [ 11.294570] ? __snp_enable+0x5d/0x70 [ 11.294575] snp_enable+0x19/0x20 [ 11.294578] __flush_smp_call_function_queue+0x9c/0x3a0 [ 11.294586] generic_smp_call_function_single_interrupt+0x17/0x20 [ 11.294589] __sysvec_call_function+0x20/0x90 [ 11.294596] sysvec_call_function+0x80/0xb0 [ 11.294601] [ 11.294603] [ 11.294605] asm_sysvec_call_function+0x1f/0x30 ... [ 11.294631] arch_cpu_idle+0xd/0x20 [ 11.294633] default_idle_call+0x34/0xd0 [ 11.294636] do_idle+0x1f1/0x230 [ 11.294643] ? complete+0x71/0x80 [ 11.294649] cpu_startup_entry+0x30/0x40 [ 11.294652] start_secondary+0x12d/0x160 [ 11.294655] common_startup_64+0x13e/0x141 [ 11.294662] This #GP exception is getting triggered due to the following errata for AMD family 19h Models 10h-1Fh Processors: Processor may generate spurious #GP(0) Exception on WRMSR instruction: Description: The Processor will generate a spurious #GP(0) Exception on a WRMSR instruction if the following conditions are all met: - the target of the WRMSR is a SYSCFG register. - the write changes the value of SYSCFG.SNPEn from 0 to 1. - One of the threads that share the physical core has a non-zero value in the VM_HSAVE_PA MSR. The document being referred to above: https://www.amd.com/content/dam/amd/en/documents/processor-tech-docs/revision-guides/57095-PUB_1_01.pdf To summarize, with kvm_amd module being built-in, KVM/SVM initialization happens before host SNP is enabled and this SVM initialization sets VM_HSAVE_PA to non-zero, which then triggers a #GP when SYSCFG.SNPEn is being set and this will subsequently cause SNP_INIT(_EX) to fail with INVALID_CONFIG error as SYSCFG[SnpEn] is not set on all CPUs. Essentially SNP host enabling code should be invoked before KVM initialization, which is currently not the case when KVM is built-in. Add fix to call snp_rmptable_init() early from iommu_snp_enable() directly and not invoked via device_initcall() which enables SNP host support before KVM initialization with kvm_amd module built-in. Add additional handling for `iommu=off` or `amd_iommu=off` options. Note that IOMMUs need to be enabled for SNP initialization, therefore, if host SNP support is enabled but late IOMMU initialization fails then that will cause PSP driver's SNP_INIT to fail as IOMMU SNP sanity checks in SNP firmware will fail with invalid configuration error as below: [ 9.723114] ccp 0000:23:00.1: sev enabled [ 9.727602] ccp 0000:23:00.1: psp enabled [ 9.732527] ccp 0000:a2:00.1: enabling device (0000 -> 0002) [ 9.739098] ccp 0000:a2:00.1: no command queues available [ 9.745167] ccp 0000:a2:00.1: psp enabled [ 9.805337] ccp 0000:23:00.1: SEV-SNP: failed to INIT rc -5, error 0x3 [ 9.866426] ccp 0000:23:00.1: SEV API:1.53 build:5 Fixes: c3b86e61b756 ("x86/cpufeatures: Enable/unmask SEV-SNP CPU feature") Co-developed-by: Sean Christopherson Signed-off-by: Sean Christopherson Co-developed-by: Vasant Hegde Signed-off-by: Vasant Hegde Cc: Signed-off-by: Ashish Kalra Acked-by: Joerg Roedel Message-ID: <138b520fb83964782303b43ade4369cd181fdd9c.1739226950.git.ashish.kalra@amd.com> Signed-off-by: Paolo Bonzini --- arch/x86/include/asm/sev.h | 2 ++ arch/x86/virt/svm/sev.c | 23 +++++++---------------- drivers/iommu/amd/init.c | 34 ++++++++++++++++++++++++++++++---- 3 files changed, 39 insertions(+), 20 deletions(-) diff --git a/arch/x86/include/asm/sev.h b/arch/x86/include/asm/sev.h index 5d9685f92e5c..1581246491b5 100644 --- a/arch/x86/include/asm/sev.h +++ b/arch/x86/include/asm/sev.h @@ -531,6 +531,7 @@ static inline void __init snp_secure_tsc_init(void) { } #ifdef CONFIG_KVM_AMD_SEV bool snp_probe_rmptable_info(void); +int snp_rmptable_init(void); int snp_lookup_rmpentry(u64 pfn, bool *assigned, int *level); void snp_dump_hva_rmpentry(unsigned long address); int psmash(u64 pfn); @@ -541,6 +542,7 @@ void kdump_sev_callback(void); void snp_fixup_e820_tables(void); #else static inline bool snp_probe_rmptable_info(void) { return false; } +static inline int snp_rmptable_init(void) { return -ENOSYS; } static inline int snp_lookup_rmpentry(u64 pfn, bool *assigned, int *level) { return -ENODEV; } static inline void snp_dump_hva_rmpentry(unsigned long address) {} static inline int psmash(u64 pfn) { return -ENODEV; } diff --git a/arch/x86/virt/svm/sev.c b/arch/x86/virt/svm/sev.c index 1dcc027ec77e..42e74a5a7d78 100644 --- a/arch/x86/virt/svm/sev.c +++ b/arch/x86/virt/svm/sev.c @@ -505,19 +505,19 @@ static bool __init setup_rmptable(void) * described in the SNP_INIT_EX firmware command description in the SNP * firmware ABI spec. */ -static int __init snp_rmptable_init(void) +int __init snp_rmptable_init(void) { unsigned int i; u64 val; - if (!cc_platform_has(CC_ATTR_HOST_SEV_SNP)) - return 0; + if (WARN_ON_ONCE(!cc_platform_has(CC_ATTR_HOST_SEV_SNP))) + return -ENOSYS; - if (!amd_iommu_snp_en) - goto nosnp; + if (WARN_ON_ONCE(!amd_iommu_snp_en)) + return -ENOSYS; if (!setup_rmptable()) - goto nosnp; + return -ENOSYS; /* * Check if SEV-SNP is already enabled, this can happen in case of @@ -530,7 +530,7 @@ static int __init snp_rmptable_init(void) /* Zero out the RMP bookkeeping area */ if (!clear_rmptable_bookkeeping()) { free_rmp_segment_table(); - goto nosnp; + return -ENOSYS; } /* Zero out the RMP entries */ @@ -562,17 +562,8 @@ skip_enable: crash_kexec_post_notifiers = true; return 0; - -nosnp: - cc_platform_clear(CC_ATTR_HOST_SEV_SNP); - return -ENOSYS; } -/* - * This must be called after the IOMMU has been initialized. - */ -device_initcall(snp_rmptable_init); - static void set_rmp_segment_info(unsigned int segment_shift) { rmp_segment_shift = segment_shift; diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c index c5cd92edada0..2fecfed75e54 100644 --- a/drivers/iommu/amd/init.c +++ b/drivers/iommu/amd/init.c @@ -3194,7 +3194,7 @@ out: return true; } -static void iommu_snp_enable(void) +static __init void iommu_snp_enable(void) { #ifdef CONFIG_KVM_AMD_SEV if (!cc_platform_has(CC_ATTR_HOST_SEV_SNP)) @@ -3219,6 +3219,14 @@ static void iommu_snp_enable(void) goto disable_snp; } + /* + * Enable host SNP support once SNP support is checked on IOMMU. + */ + if (snp_rmptable_init()) { + pr_warn("SNP: RMP initialization failed, SNP cannot be supported.\n"); + goto disable_snp; + } + pr_info("IOMMU SNP support enabled.\n"); return; @@ -3318,6 +3326,19 @@ static int __init iommu_go_to_state(enum iommu_init_state state) ret = state_next(); } + /* + * SNP platform initilazation requires IOMMUs to be fully configured. + * If the SNP support on IOMMUs has NOT been checked, simply mark SNP + * as unsupported. If the SNP support on IOMMUs has been checked and + * host SNP support enabled but RMP enforcement has not been enabled + * in IOMMUs, then the system is in a half-baked state, but can limp + * along as all memory should be Hypervisor-Owned in the RMP. WARN, + * but leave SNP as "supported" to avoid confusing the kernel. + */ + if (ret && cc_platform_has(CC_ATTR_HOST_SEV_SNP) && + !WARN_ON_ONCE(amd_iommu_snp_en)) + cc_platform_clear(CC_ATTR_HOST_SEV_SNP); + return ret; } @@ -3426,18 +3447,23 @@ void __init amd_iommu_detect(void) int ret; if (no_iommu || (iommu_detected && !gart_iommu_aperture)) - return; + goto disable_snp; if (!amd_iommu_sme_check()) - return; + goto disable_snp; ret = iommu_go_to_state(IOMMU_IVRS_DETECTED); if (ret) - return; + goto disable_snp; amd_iommu_detected = true; iommu_detected = 1; x86_init.iommu.iommu_init = amd_iommu_init; + return; + +disable_snp: + if (cc_platform_has(CC_ATTR_HOST_SEV_SNP)) + cc_platform_clear(CC_ATTR_HOST_SEV_SNP); } /**************************************************************************** -- 2.51.0 From d1d0963121769d8d16150b913fe886e48efefa51 Mon Sep 17 00:00:00 2001 From: Masahiro Yamada Date: Tue, 11 Feb 2025 09:29:06 +0900 Subject: [PATCH 03/16] tools: fix annoying "mkdir -p ..." logs when building tools in parallel When CONFIG_OBJTOOL=y or CONFIG_DEBUG_INFO_BTF=y, parallel builds show awkward "mkdir -p ..." logs. $ make -j16 [ snip ] mkdir -p /home/masahiro/ref/linux/tools/objtool && make O=/home/masahiro/ref/linux subdir=tools/objtool --no-print-directory -C objtool mkdir -p /home/masahiro/ref/linux/tools/bpf/resolve_btfids && make O=/home/masahiro/ref/linux subdir=tools/bpf/resolve_btfids --no-print-directory -C bpf/resolve_btfids Defining MAKEFLAGS= on the command line wipes out command line switches from the resultant MAKEFLAGS definition, even though the command line switches are active. [1] MAKEFLAGS puts all single-letter options into the first word, and that word will be empty if no single-letter options were given. [2] However, this breaks if MAKEFLAGS= is given on the command line. The tools/ and tools/% targets set MAKEFLAGS= on the command line, which breaks the following code in tools/scripts/Makefile.include: short-opts := $(firstword -$(MAKEFLAGS)) If MAKEFLAGS really needs modification, it should be done through the environment variable, as follows: MAKEFLAGS= $(MAKE) ... That said, I question whether modifying MAKEFLAGS is necessary here. The only flag we might want to exclude is --no-print-directory, as the tools build system changes the working directory. However, people might find the "Entering/Leaving directory" logs annoying. I simply removed the offending MAKEFLAGS=. [1]: https://savannah.gnu.org/bugs/?62469 [2]: https://www.gnu.org/software/make/manual/make.html#Testing-Flags Fixes: ea01fa9f63ae ("tools: Connect to the kernel build system") Fixes: a50e43332756 ("perf tools: Honor parallel jobs") Signed-off-by: Masahiro Yamada Tested-by: Daniel Xu --- Makefile | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) diff --git a/Makefile b/Makefile index 89628e354ca7..52207bcb1a9d 100644 --- a/Makefile +++ b/Makefile @@ -1421,18 +1421,13 @@ ifneq ($(wildcard $(resolve_btfids_O)),) $(Q)$(MAKE) -sC $(srctree)/tools/bpf/resolve_btfids O=$(resolve_btfids_O) clean endif -# Clear a bunch of variables before executing the submake -ifeq ($(quiet),silent_) -tools_silent=s -endif - tools/: FORCE $(Q)mkdir -p $(objtree)/tools - $(Q)$(MAKE) LDFLAGS= MAKEFLAGS="$(tools_silent) $(filter --j% -j,$(MAKEFLAGS))" O=$(abspath $(objtree)) subdir=tools -C $(srctree)/tools/ + $(Q)$(MAKE) LDFLAGS= O=$(abspath $(objtree)) subdir=tools -C $(srctree)/tools/ tools/%: FORCE $(Q)mkdir -p $(objtree)/tools - $(Q)$(MAKE) LDFLAGS= MAKEFLAGS="$(tools_silent) $(filter --j% -j,$(MAKEFLAGS))" O=$(abspath $(objtree)) subdir=tools -C $(srctree)/tools/ $* + $(Q)$(MAKE) LDFLAGS= O=$(abspath $(objtree)) subdir=tools -C $(srctree)/tools/ $* # --------------------------------------------------------------------------- # Kernel selftest -- 2.51.0 From 140332b6ed727a4ec2e5722a1ccda28b52d45771 Mon Sep 17 00:00:00 2001 From: Masahiro Yamada Date: Thu, 13 Feb 2025 15:26:44 +0900 Subject: [PATCH 04/16] kbuild: fix linux-headers package build when $(CC) cannot link userspace Since commit 5f73e7d0386d ("kbuild: refactor cross-compiling linux-headers package"), the linux-headers Debian package fails to build when $(CC) cannot build userspace applications, for example, when using toolchains installed by the 0day bot. The host programs in the linux-headers package should be rebuilt using the disto's cross-compiler, ${DEB_HOST_GNU_TYPE}-gcc instead of $(CC). Hence, the variable 'CC' must be expanded in this shell script instead of in the top-level Makefile. Commit f354fc88a72a ("kbuild: install-extmod-build: add missing quotation marks for CC variable") was not a correct fix because CC="ccache gcc" should be unrelated when rebuilding userspace tools. Fixes: 5f73e7d0386d ("kbuild: refactor cross-compiling linux-headers package") Reported-by: Jeff Johnson Closes: https://lore.kernel.org/linux-kbuild/CAK7LNARb3xO3ptBWOMpwKcyf3=zkfhMey5H2KnB1dOmUwM79dA@mail.gmail.com/T/#t Signed-off-by: Masahiro Yamada Tested-by: Jeff Johnson --- scripts/package/install-extmod-build | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/scripts/package/install-extmod-build b/scripts/package/install-extmod-build index b724626ea0ca..2966473b4660 100755 --- a/scripts/package/install-extmod-build +++ b/scripts/package/install-extmod-build @@ -62,8 +62,8 @@ if [ "${CC}" != "${HOSTCC}" ]; then # # Clear VPATH and srcroot because the source files reside in the output # directory. - # shellcheck disable=SC2016 # $(MAKE), $(CC), and $(build) will be expanded by Make - "${MAKE}" run-command KBUILD_RUN_COMMAND='+$(MAKE) HOSTCC="$(CC)" VPATH= srcroot=. $(build)='"${destdir}"/scripts + # shellcheck disable=SC2016 # $(MAKE) and $(build) will be expanded by Make + "${MAKE}" run-command KBUILD_RUN_COMMAND='+$(MAKE) HOSTCC='"${CC}"' VPATH= srcroot=. $(build)='"${destdir}"/scripts rm -f "${destdir}/scripts/Kbuild" fi -- 2.51.0 From d440148418f4816b4973ec6723bf63821793a0a7 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sat, 15 Feb 2025 09:28:55 -0800 Subject: [PATCH 05/16] tegra210-adma: fix 32-bit x86 build The Tegra210 Audio DMA controller driver did a plain divide: page_no = (res_page->start - res_base->start) / cdata->ch_base_offset; which causes problems on 32-bit x86 configurations that have 64-bit resource sizes: x86_64-linux-ld: drivers/dma/tegra210-adma.o: in function `tegra_adma_probe': tegra210-adma.c:(.text+0x1322): undefined reference to `__udivdi3' because gcc doesn't generate the trivial code for a 64-by-32 divide, turning it into a function call to do a full 64-by-64 divide. And the kernel intentionally doesn't provide that helper function, because 99% of the time all you want is the narrower version. Of course, tegra210 is a 64-bit architecture and the 32-bit x86 build is purely for build testing, so this really is just about build coverage failure. But build coverage is good. Side note: div_u64() would be suboptimal if you actually have a 32-bit resource_t, so our "helper" for divides are admittedly making it harder than it should be to generate good code for all the possible cases. At some point, I'll consider 32-bit x86 so entirely legacy that I can't find it in myself to care any more, and we'll just add the __udivdi3 library function. But for now, the right thing to do is to use "div_u64()" to show that you know that you are doing the simpler divide with a 32-bit number. And the build error enforces that. While fixing the build issue, also check for division-by-zero, and for overflow. Which hopefully cannot happen on real production hardware, but the value of 'ch_base_offset' can definitely be zero in other places. Reported-by: Guenter Roeck Signed-off-by: Linus Torvalds --- drivers/dma/tegra210-adma.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/drivers/dma/tegra210-adma.c b/drivers/dma/tegra210-adma.c index 6896da8ac7ef..5c6a5b358987 100644 --- a/drivers/dma/tegra210-adma.c +++ b/drivers/dma/tegra210-adma.c @@ -887,7 +887,7 @@ static int tegra_adma_probe(struct platform_device *pdev) const struct tegra_adma_chip_data *cdata; struct tegra_adma *tdma; struct resource *res_page, *res_base; - int ret, i, page_no; + int ret, i; cdata = of_device_get_match_data(&pdev->dev); if (!cdata) { @@ -914,9 +914,20 @@ static int tegra_adma_probe(struct platform_device *pdev) res_base = platform_get_resource_byname(pdev, IORESOURCE_MEM, "global"); if (res_base) { - page_no = (res_page->start - res_base->start) / cdata->ch_base_offset; - if (page_no <= 0) + resource_size_t page_offset, page_no; + unsigned int ch_base_offset; + + if (res_page->start < res_base->start) + return -EINVAL; + page_offset = res_page->start - res_base->start; + ch_base_offset = cdata->ch_base_offset; + if (!ch_base_offset) return -EINVAL; + + page_no = div_u64(page_offset, ch_base_offset); + if (!page_no || page_no > INT_MAX) + return -EINVAL; + tdma->ch_page_no = page_no - 1; tdma->base_addr = devm_ioremap_resource(&pdev->dev, res_base); if (IS_ERR(tdma->base_addr)) -- 2.51.0 From 1b71c2fb04e7a713abc6edde4a412416ff3158f2 Mon Sep 17 00:00:00 2001 From: =?utf8?q?Thomas=20Wei=C3=9Fschuh?= Date: Thu, 13 Feb 2025 15:55:17 +0100 Subject: [PATCH 06/16] kbuild: userprogs: fix bitsize and target detection on clang MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit scripts/Makefile.clang was changed in the linked commit to move --target from KBUILD_CFLAGS to KBUILD_CPPFLAGS, as that generally has a broader scope. However that variable is not inspected by the userprogs logic, breaking cross compilation on clang. Use both variables to detect bitsize and target arguments for userprogs. Fixes: feb843a469fb ("kbuild: add $(CLANG_FLAGS) to KBUILD_CPPFLAGS") Cc: stable@vger.kernel.org Signed-off-by: Thomas Weißschuh Reviewed-by: Nathan Chancellor Signed-off-by: Masahiro Yamada --- Makefile | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Makefile b/Makefile index 52207bcb1a9d..272db408be5c 100644 --- a/Makefile +++ b/Makefile @@ -1120,8 +1120,8 @@ LDFLAGS_vmlinux += --orphan-handling=$(CONFIG_LD_ORPHAN_WARN_LEVEL) endif # Align the bit size of userspace programs with the kernel -KBUILD_USERCFLAGS += $(filter -m32 -m64 --target=%, $(KBUILD_CFLAGS)) -KBUILD_USERLDFLAGS += $(filter -m32 -m64 --target=%, $(KBUILD_CFLAGS)) +KBUILD_USERCFLAGS += $(filter -m32 -m64 --target=%, $(KBUILD_CPPFLAGS) $(KBUILD_CFLAGS)) +KBUILD_USERLDFLAGS += $(filter -m32 -m64 --target=%, $(KBUILD_CPPFLAGS) $(KBUILD_CFLAGS)) # make the checker run with the right architecture CHECKFLAGS += --arch=$(ARCH) -- 2.51.0 From b28fb1f2ef45eeef1cd2c23149b50d184d545a3e Mon Sep 17 00:00:00 2001 From: =?utf8?q?Uwe=20Kleine-K=C3=B6nig?= Date: Thu, 13 Feb 2025 17:04:29 +0100 Subject: [PATCH 07/16] modpost: Fix a few typos in a comment MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit Namely: s/becasue/because/ and s/wiht/with/ plus an added article. Signed-off-by: Uwe Kleine-König Signed-off-by: Masahiro Yamada --- scripts/mod/modpost.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c index 36b28987a2f0..c35d22607978 100644 --- a/scripts/mod/modpost.c +++ b/scripts/mod/modpost.c @@ -190,8 +190,8 @@ static struct module *new_module(const char *name, size_t namelen) /* * Set mod->is_gpl_compatible to true by default. If MODULE_LICENSE() - * is missing, do not check the use for EXPORT_SYMBOL_GPL() becasue - * modpost will exit wiht error anyway. + * is missing, do not check the use for EXPORT_SYMBOL_GPL() because + * modpost will exit with an error anyway. */ mod->is_gpl_compatible = true; -- 2.51.0 From 129fe718819cc5e24ea2f489db9ccd4371f0c6f6 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 14 Feb 2025 11:55:47 -0500 Subject: [PATCH 08/16] tracing: Do not allow mmap() of persistent ring buffer When trying to mmap a trace instance buffer that is attached to reserve_mem, it would crash: BUG: unable to handle page fault for address: ffffe97bd00025c8 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 2862f3067 P4D 2862f3067 PUD 0 Oops: Oops: 0000 [#1] PREEMPT_RT SMP PTI CPU: 4 UID: 0 PID: 981 Comm: mmap-rb Not tainted 6.14.0-rc2-test-00003-g7f1a5e3fbf9e-dirty #233 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2 04/01/2014 RIP: 0010:validate_page_before_insert+0x5/0xb0 Code: e2 01 89 d0 c3 cc cc cc cc 66 66 2e 0f 1f 84 00 00 00 00 00 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 0f 1f 44 00 00 <48> 8b 46 08 a8 01 75 67 66 90 48 89 f0 8b 50 34 85 d2 74 76 48 89 RSP: 0018:ffffb148c2f3f968 EFLAGS: 00010246 RAX: ffff9fa5d3322000 RBX: ffff9fa5ccff9c08 RCX: 00000000b879ed29 RDX: ffffe97bd00025c0 RSI: ffffe97bd00025c0 RDI: ffff9fa5ccff9c08 RBP: ffffb148c2f3f9f0 R08: 0000000000000004 R09: 0000000000000004 R10: 0000000000000000 R11: 0000000000000200 R12: 0000000000000000 R13: 00007f16a18d5000 R14: ffff9fa5c48db6a8 R15: 0000000000000000 FS: 00007f16a1b54740(0000) GS:ffff9fa73df00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffe97bd00025c8 CR3: 00000001048c6006 CR4: 0000000000172ef0 Call Trace: ? __die_body.cold+0x19/0x1f ? __die+0x2e/0x40 ? page_fault_oops+0x157/0x2b0 ? search_module_extables+0x53/0x80 ? validate_page_before_insert+0x5/0xb0 ? kernelmode_fixup_or_oops.isra.0+0x5f/0x70 ? __bad_area_nosemaphore+0x16e/0x1b0 ? bad_area_nosemaphore+0x16/0x20 ? do_kern_addr_fault+0x77/0x90 ? exc_page_fault+0x22b/0x230 ? asm_exc_page_fault+0x2b/0x30 ? validate_page_before_insert+0x5/0xb0 ? vm_insert_pages+0x151/0x400 __rb_map_vma+0x21f/0x3f0 ring_buffer_map+0x21b/0x2f0 tracing_buffers_mmap+0x70/0xd0 __mmap_region+0x6f0/0xbd0 mmap_region+0x7f/0x130 do_mmap+0x475/0x610 vm_mmap_pgoff+0xf2/0x1d0 ksys_mmap_pgoff+0x166/0x200 __x64_sys_mmap+0x37/0x50 x64_sys_call+0x1670/0x1d70 do_syscall_64+0xbb/0x1d0 entry_SYSCALL_64_after_hwframe+0x77/0x7f The reason was that the code that maps the ring buffer pages to user space has: page = virt_to_page((void *)cpu_buffer->subbuf_ids[s]); And uses that in: vm_insert_pages(vma, vma->vm_start, pages, &nr_pages); But virt_to_page() does not work with vmap()'d memory which is what the persistent ring buffer has. It is rather trivial to allow this, but for now just disable mmap() of instances that have their ring buffer from the reserve_mem option. If an mmap() is performed on a persistent buffer it will return -ENODEV just like it would if the .mmap field wasn't defined in the file_operations structure. Cc: stable@vger.kernel.org Cc: Masami Hiramatsu Cc: Mathieu Desnoyers Cc: Vincent Donnefort Link: https://lore.kernel.org/20250214115547.0d7287d3@gandalf.local.home Fixes: 9b7bdf6f6ece6 ("tracing: Have trace_printk not use binary prints if boot buffer") Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 25ff37aab00f..0e6d517e74e0 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -8279,6 +8279,10 @@ static int tracing_buffers_mmap(struct file *filp, struct vm_area_struct *vma) struct trace_iterator *iter = &info->iter; int ret = 0; + /* Currently the boot mapped buffer is not supported for mmap */ + if (iter->tr->flags & TRACE_ARRAY_FL_BOOT) + return -ENODEV; + ret = get_snapshot_map(iter->tr); if (ret) return ret; -- 2.51.0 From 97937834ae876f29565415ab15f1284666dc6be3 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 14 Feb 2025 12:35:12 -0500 Subject: [PATCH 09/16] ring-buffer: Update pages_touched to reflect persistent buffer content The pages_touched field represents the number of subbuffers in the ring buffer that have content that can be read. This is used in accounting of "dirty_pages" and "buffer_percent" to allow the user to wait for the buffer to be filled to a certain amount before it reads the buffer in blocking mode. The persistent buffer never updated this value so it was set to zero, and this accounting would take it as it had no content. This would cause user space to wait for content even though there's enough content in the ring buffer that satisfies the buffer_percent. Cc: stable@vger.kernel.org Cc: Masami Hiramatsu Cc: Mathieu Desnoyers Cc: Vincent Donnefort Link: https://lore.kernel.org/20250214123512.0631436e@gandalf.local.home Fixes: 5f3b6e839f3ce ("ring-buffer: Validate boot range memory events") Signed-off-by: Steven Rostedt (Google) --- kernel/trace/ring_buffer.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 0419d41a2060..bb6089c2951e 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1850,6 +1850,11 @@ static void rb_meta_validate_events(struct ring_buffer_per_cpu *cpu_buffer) cpu_buffer->cpu); goto invalid; } + + /* If the buffer has content, update pages_touched */ + if (ret) + local_inc(&cpu_buffer->pages_touched); + entries += ret; entry_bytes += local_read(&head_page->page->commit); local_set(&cpu_buffer->head_page->entries, ret); -- 2.51.0 From 0ad2507d5d93f39619fc42372c347d6006b64319 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 16 Feb 2025 14:02:44 -0800 Subject: [PATCH 10/16] Linux 6.14-rc3 --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 272db408be5c..96407c1d6be1 100644 --- a/Makefile +++ b/Makefile @@ -2,7 +2,7 @@ VERSION = 6 PATCHLEVEL = 14 SUBLEVEL = 0 -EXTRAVERSION = -rc2 +EXTRAVERSION = -rc3 NAME = Baby Opossum Posse # *DOCUMENTATION* -- 2.51.0 From d42044aad6528e0c9533dbaf836d1b0fbb19fe2d Mon Sep 17 00:00:00 2001 From: David Reaver Date: Sun, 12 Jan 2025 07:26:55 -0800 Subject: [PATCH 11/16] PM: hibernate: Replace deprecated kmap_atomic() with kmap_local_page() kmap_atomic() is deprecated and should be replaced with kmap_local_page() [1][2]. kmap_local_page() is faster in kernels with HIGHMEM enabled, can take page faults, and allows preemption. According to [2], this replacement is safe as long as the code between kmap_atomic() and kunmap_atomic() does not implicitly depend on disabling page faults or preemption. In all of the call sites in this patch, the only thing happening between mapping and unmapping pages is copy_page() calls, and I don't suspect they depend on disabling page faults or preemption. Link: https://lwn.net/Articles/836144/ [1] Link: https://docs.kernel.org/mm/highmem.html#temporary-virtual-mappings [2] Signed-off-by: David Reaver Link: https://patch.msgid.link/20250112152658.20132-1-me@davidreaver.com Signed-off-by: Rafael J. Wysocki --- kernel/power/snapshot.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index c9fb559a6399..4e6e24e8b854 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -2270,9 +2270,9 @@ int snapshot_read_next(struct snapshot_handle *handle) */ void *kaddr; - kaddr = kmap_atomic(page); + kaddr = kmap_local_page(page); copy_page(buffer, kaddr); - kunmap_atomic(kaddr); + kunmap_local(kaddr); handle->buffer = buffer; } else { handle->buffer = page_address(page); @@ -2561,9 +2561,9 @@ static void copy_last_highmem_page(void) if (last_highmem_page) { void *dst; - dst = kmap_atomic(last_highmem_page); + dst = kmap_local_page(last_highmem_page); copy_page(dst, buffer); - kunmap_atomic(dst); + kunmap_local(dst); last_highmem_page = NULL; } } @@ -2881,13 +2881,13 @@ static inline void swap_two_pages_data(struct page *p1, struct page *p2, { void *kaddr1, *kaddr2; - kaddr1 = kmap_atomic(p1); - kaddr2 = kmap_atomic(p2); + kaddr1 = kmap_local_page(p1); + kaddr2 = kmap_local_page(p2); copy_page(buf, kaddr1); copy_page(kaddr1, kaddr2); copy_page(kaddr2, buf); - kunmap_atomic(kaddr2); - kunmap_atomic(kaddr1); + kunmap_local(kaddr2); + kunmap_local(kaddr1); } /** -- 2.51.0 From 258e231dc29fbd72bc82c16859a8304f71780ba2 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Mon, 17 Feb 2025 21:03:01 +0100 Subject: [PATCH 12/16] PM: Rearrange documentation related to __pm_runtime_disable() There are only two callers of __pm_runtime_disable(), one of which is device_suspend_late() and the other is pm_runtime_disable() that has its own kerneldoc comment and there are no plans to add any more of them. Since they use different values of the __pm_runtime_disable() second parameter, the actual code behavior is different in each case, but it is all documented in the __pm_runtime_disable() kerneldoc comment which is not particularly straightforward. For this reason, move the information from the __pm_runtime_disable() kerneldoc comment to the pm_runtime_disable() one and into a separate comment in device_suspend_late() and remove the __pm_runtime_disable() kerneldoc comment altogether. No functional impact. Signed-off-by: Rafael J. Wysocki Reviewed-by: Ulf Hansson Link: https://patch.msgid.link/12617588.O9o76ZdvQC@rjwysocki.net --- drivers/base/power/main.c | 4 ++++ drivers/base/power/runtime.c | 14 -------------- include/linux/pm_runtime.h | 17 ++++++++++++----- 3 files changed, 16 insertions(+), 19 deletions(-) diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index 40e1d8d8a589..dffa2aa1ba7d 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -1404,6 +1404,10 @@ static int device_suspend_late(struct device *dev, pm_message_t state, bool asyn TRACE_DEVICE(dev); TRACE_SUSPEND(0); + /* + * Disable runtime PM for the device without checking if there is a + * pending resume request for it. + */ __pm_runtime_disable(dev, false); dpm_wait_for_subordinate(dev, async); diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c index 2ee45841486b..a5aed89e1a6b 100644 --- a/drivers/base/power/runtime.c +++ b/drivers/base/power/runtime.c @@ -1460,20 +1460,6 @@ int pm_runtime_barrier(struct device *dev) } EXPORT_SYMBOL_GPL(pm_runtime_barrier); -/** - * __pm_runtime_disable - Disable runtime PM of a device. - * @dev: Device to handle. - * @check_resume: If set, check if there's a resume request for the device. - * - * Increment power.disable_depth for the device and if it was zero previously, - * cancel all pending runtime PM requests for the device and wait for all - * operations in progress to complete. The device can be either active or - * suspended after its runtime PM has been disabled. - * - * If @check_resume is set and there's a resume request pending when - * __pm_runtime_disable() is called and power.disable_depth is zero, the - * function will wake up the device before disabling its runtime PM. - */ void __pm_runtime_disable(struct device *dev, bool check_resume) { spin_lock_irq(&dev->power.lock); diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index d39dc863f612..72c62e1171ca 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -556,11 +556,18 @@ static inline int pm_runtime_set_suspended(struct device *dev) * pm_runtime_disable - Disable runtime PM for a device. * @dev: Target device. * - * Prevent the runtime PM framework from working with @dev (by incrementing its - * "blocking" counter). - * - * For each invocation of this function for @dev there must be a matching - * pm_runtime_enable() call in order for runtime PM to be enabled for it. + * Prevent the runtime PM framework from working with @dev by incrementing its + * "disable" counter. + * + * If the counter is zero when this function runs and there is a pending runtime + * resume request for @dev, it will be resumed. If the counter is still zero at + * that point, all of the pending runtime PM requests for @dev will be canceled + * and all runtime PM operations in progress involving it will be waited for to + * complete. + * + * For each invocation of this function for @dev, there must be a matching + * pm_runtime_enable() call, so that runtime PM is eventually enabled for it + * again. */ static inline void pm_runtime_disable(struct device *dev) { -- 2.51.0 From 3e5eee147b7b0f5a93f56beffe34e81fdd00fa0d Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Tue, 18 Feb 2025 21:11:42 +0100 Subject: [PATCH 13/16] PM: Block enabling of runtime PM during system suspend If device_prepare() runs on a device that has never had runtime PM enabled so far, it may reasonably assume that runtime PM will not be enabled for that device during the system suspend-resume cycle currently in progress, but this has never been guaranteed. To verify this assumption, make device_prepare() arrange for triggering a device warning accompanied by a call trace dump if runtime PM is enabled for such a device after it has returned. Signed-off-by: Rafael J. Wysocki Reviewed-by: Ulf Hansson Link: https://patch.msgid.link/6131109.lOV4Wx5bFT@rjwysocki.net --- drivers/base/power/main.c | 9 +++++++++ drivers/base/power/runtime.c | 24 ++++++++++++++++++++++++ include/linux/pm.h | 1 + include/linux/pm_runtime.h | 4 ++++ 4 files changed, 38 insertions(+) diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index dffa2aa1ba7d..acabd9f3e60f 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -1109,6 +1109,8 @@ static void device_complete(struct device *dev, pm_message_t state) device_unlock(dev); out: + /* If enabling runtime PM for the device is blocked, unblock it. */ + pm_runtime_unblock(dev); pm_runtime_put(dev); } @@ -1815,6 +1817,13 @@ static int device_prepare(struct device *dev, pm_message_t state) * it again during the complete phase. */ pm_runtime_get_noresume(dev); + /* + * If runtime PM is disabled for the device at this point and it has + * never been enabled so far, it should not be enabled until this system + * suspend-resume cycle is complete, so prepare to trigger a warning on + * subsequent attempts to enable it. + */ + pm_runtime_block_if_disabled(dev); if (dev->power.syscore) return 0; diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c index a5aed89e1a6b..797ea38ceba7 100644 --- a/drivers/base/power/runtime.c +++ b/drivers/base/power/runtime.c @@ -1460,6 +1460,26 @@ int pm_runtime_barrier(struct device *dev) } EXPORT_SYMBOL_GPL(pm_runtime_barrier); +void pm_runtime_block_if_disabled(struct device *dev) +{ + spin_lock_irq(&dev->power.lock); + + if (dev->power.disable_depth && dev->power.last_status == RPM_INVALID) + dev->power.last_status = RPM_BLOCKED; + + spin_unlock_irq(&dev->power.lock); +} + +void pm_runtime_unblock(struct device *dev) +{ + spin_lock_irq(&dev->power.lock); + + if (dev->power.last_status == RPM_BLOCKED) + dev->power.last_status = RPM_INVALID; + + spin_unlock_irq(&dev->power.lock); +} + void __pm_runtime_disable(struct device *dev, bool check_resume) { spin_lock_irq(&dev->power.lock); @@ -1518,6 +1538,10 @@ void pm_runtime_enable(struct device *dev) if (--dev->power.disable_depth > 0) goto out; + if (dev->power.last_status == RPM_BLOCKED) { + dev_warn(dev, "Attempt to enable runtime PM when it is blocked\n"); + dump_stack(); + } dev->power.last_status = RPM_INVALID; dev->power.accounting_timestamp = ktime_get_mono_fast_ns(); diff --git a/include/linux/pm.h b/include/linux/pm.h index 78855d794342..6ca6f34c58c3 100644 --- a/include/linux/pm.h +++ b/include/linux/pm.h @@ -597,6 +597,7 @@ enum rpm_status { RPM_RESUMING, RPM_SUSPENDED, RPM_SUSPENDING, + RPM_BLOCKED, }; /* diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index 72c62e1171ca..10769119867b 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -77,6 +77,8 @@ extern int pm_runtime_get_if_in_use(struct device *dev); extern int pm_schedule_suspend(struct device *dev, unsigned int delay); extern int __pm_runtime_set_status(struct device *dev, unsigned int status); extern int pm_runtime_barrier(struct device *dev); +extern void pm_runtime_block_if_disabled(struct device *dev); +extern void pm_runtime_unblock(struct device *dev); extern void pm_runtime_enable(struct device *dev); extern void __pm_runtime_disable(struct device *dev, bool check_resume); extern void pm_runtime_allow(struct device *dev); @@ -271,6 +273,8 @@ static inline int pm_runtime_get_if_active(struct device *dev) static inline int __pm_runtime_set_status(struct device *dev, unsigned int status) { return 0; } static inline int pm_runtime_barrier(struct device *dev) { return 0; } +static inline void pm_runtime_block_if_disabled(struct device *dev) {} +static inline void pm_runtime_unblock(struct device *dev) {} static inline void pm_runtime_enable(struct device *dev) {} static inline void __pm_runtime_disable(struct device *dev, bool c) {} static inline void pm_runtime_allow(struct device *dev) {} -- 2.51.0 From 758cc55ce3d5d79e8f98adbd03ad2cd29133af33 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Tue, 18 Feb 2025 21:13:09 +0100 Subject: [PATCH 14/16] PM: runtime: Introduce pm_runtime_blocked() Introduce a new helper function called pm_runtime_blocked() for checking the power.last_status value indicating whether or not enabling runtime PM for the given device has been blocked (which happens in the "prepare" phase of system-wide suspend if runtime PM is disabled for the given device at that point). Signed-off-by: Rafael J. Wysocki Reviewed-by: Ulf Hansson Link: https://patch.msgid.link/4632087.LvFx2qVVIh@rjwysocki.net --- drivers/base/power/runtime.c | 17 +++++++++++++++++ include/linux/pm_runtime.h | 2 ++ 2 files changed, 19 insertions(+) diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c index 797ea38ceba7..c0f5a9f89299 100644 --- a/drivers/base/power/runtime.c +++ b/drivers/base/power/runtime.c @@ -1555,6 +1555,23 @@ out: } EXPORT_SYMBOL_GPL(pm_runtime_enable); +bool pm_runtime_blocked(struct device *dev) +{ + bool ret; + + /* + * dev->power.last_status is a bit field, so in case it is updated via + * RMW, read it under the spin lock. + */ + spin_lock_irq(&dev->power.lock); + + ret = dev->power.last_status == RPM_BLOCKED; + + spin_unlock_irq(&dev->power.lock); + + return ret; +} + static void pm_runtime_disable_action(void *data) { pm_runtime_dont_use_autosuspend(data); diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index 10769119867b..aea0395c10a1 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -81,6 +81,7 @@ extern void pm_runtime_block_if_disabled(struct device *dev); extern void pm_runtime_unblock(struct device *dev); extern void pm_runtime_enable(struct device *dev); extern void __pm_runtime_disable(struct device *dev, bool check_resume); +extern bool pm_runtime_blocked(struct device *dev); extern void pm_runtime_allow(struct device *dev); extern void pm_runtime_forbid(struct device *dev); extern void pm_runtime_no_callbacks(struct device *dev); @@ -277,6 +278,7 @@ static inline void pm_runtime_block_if_disabled(struct device *dev) {} static inline void pm_runtime_unblock(struct device *dev) {} static inline void pm_runtime_enable(struct device *dev) {} static inline void __pm_runtime_disable(struct device *dev, bool c) {} +static inline bool pm_runtime_blocked(struct device *dev) { return true; } static inline void pm_runtime_allow(struct device *dev) {} static inline void pm_runtime_forbid(struct device *dev) {} -- 2.51.0 From bca84a7b93fdc744d79d94423c2cb905b1832310 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Tue, 18 Feb 2025 21:16:48 +0100 Subject: [PATCH 15/16] PM: sleep: Use DPM_FLAG_SMART_SUSPEND conditionally A recent discussion has revealed that using DPM_FLAG_SMART_SUSPEND unconditionally is generally problematic because it may lead to situations in which the device's runtime PM information is internally inconsistent or does not reflect its real state [1]. For this reason, change the handling of DPM_FLAG_SMART_SUSPEND so that it is only taken into account if it is consistently set by the drivers of all devices having any PM callbacks throughout dependency graphs in accordance with the following rules: - The "smart suspend" feature is only enabled for devices whose drivers ask for it (that is, set DPM_FLAG_SMART_SUSPEND) and for devices without PM callbacks unless they have never had runtime PM enabled. - The "smart suspend" feature is not enabled for a device if it has not been enabled for the device's parent unless the parent does not take children into account or it has never had runtime PM enabled. - The "smart suspend" feature is not enabled for a device if it has not been enabled for one of the device's suppliers taking runtime PM into account unless that supplier has never had runtime PM enabled. Namely, introduce a new device PM flag called smart_suspend that is only set if the above conditions are met and update all DPM_FLAG_SMART_SUSPEND users to check power.smart_suspend instead of directly checking the latter. At the same time, drop the power.set_active flage introduced recently in commit 3775fc538f53 ("PM: sleep: core: Synchronize runtime PM status of parents and children") because it is now sufficient to check power.smart_suspend along with the dev_pm_skip_resume() return value to decide whether or not pm_runtime_set_active() needs to be called for the device. Link: https://lore.kernel.org/linux-pm/CAPDyKFroyU3YDSfw_Y6k3giVfajg3NQGwNWeteJWqpW29BojhQ@mail.gmail.com/ [1] Fixes: 7585946243d6 ("PM: sleep: core: Restrict power.set_active propagation") Signed-off-by: Rafael J. Wysocki Reviewed-by: Ulf Hansson Acked-by: Bjorn Helgaas # drivers/pci Link: https://patch.msgid.link/1914558.tdWV9SEqCh@rjwysocki.net --- drivers/acpi/device_pm.c | 4 +-- drivers/base/power/main.c | 63 ++++++++++++++++++++++++++++++--------- drivers/mfd/intel-lpss.c | 2 +- drivers/pci/pci-driver.c | 6 ++-- include/linux/device.h | 9 ++++++ include/linux/pm.h | 2 +- 6 files changed, 64 insertions(+), 22 deletions(-) diff --git a/drivers/acpi/device_pm.c b/drivers/acpi/device_pm.c index 3b4d048c4941..dbd4446025ec 100644 --- a/drivers/acpi/device_pm.c +++ b/drivers/acpi/device_pm.c @@ -1161,7 +1161,7 @@ EXPORT_SYMBOL_GPL(acpi_subsys_complete); */ int acpi_subsys_suspend(struct device *dev) { - if (!dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND) || + if (!dev_pm_smart_suspend(dev) || acpi_dev_needs_resume(dev, ACPI_COMPANION(dev))) pm_runtime_resume(dev); @@ -1320,7 +1320,7 @@ EXPORT_SYMBOL_GPL(acpi_subsys_restore_early); */ int acpi_subsys_poweroff(struct device *dev) { - if (!dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND) || + if (!dev_pm_smart_suspend(dev) || acpi_dev_needs_resume(dev, ACPI_COMPANION(dev))) pm_runtime_resume(dev); diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index acabd9f3e60f..cc9903065900 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -656,15 +656,13 @@ static void device_resume_noirq(struct device *dev, pm_message_t state, bool asy * so change its status accordingly. * * Otherwise, the device is going to be resumed, so set its PM-runtime - * status to "active" unless its power.set_active flag is clear, in + * status to "active" unless its power.smart_suspend flag is clear, in * which case it is not necessary to update its PM-runtime status. */ - if (skip_resume) { + if (skip_resume) pm_runtime_set_suspended(dev); - } else if (dev->power.set_active) { + else if (dev_pm_smart_suspend(dev)) pm_runtime_set_active(dev); - dev->power.set_active = false; - } if (dev->pm_domain) { info = "noirq power domain "; @@ -1282,14 +1280,8 @@ Skip: dev->power.may_skip_resume)) dev->power.must_resume = true; - if (dev->power.must_resume) { - if (dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND)) { - dev->power.set_active = true; - if (dev->parent && !dev->parent->power.ignore_children) - dev->parent->power.set_active = true; - } + if (dev->power.must_resume) dpm_superior_set_must_resume(dev); - } Complete: complete_all(&dev->power.completion); @@ -1797,6 +1789,49 @@ int dpm_suspend(pm_message_t state) return error; } +static void device_prepare_smart_suspend(struct device *dev) +{ + struct device_link *link; + int idx; + + /* + * The "smart suspend" feature is enabled for devices whose drivers ask + * for it and for devices without PM callbacks unless runtime PM is + * disabled and enabling it is blocked for them. + * + * However, if "smart suspend" is not enabled for the device's parent + * or any of its suppliers that take runtime PM into account, it cannot + * be enabled for the device either. + */ + dev->power.smart_suspend = (dev->power.no_pm_callbacks || + dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND)) && + !pm_runtime_blocked(dev); + + if (!dev_pm_smart_suspend(dev)) + return; + + if (dev->parent && !dev_pm_smart_suspend(dev->parent) && + !dev->parent->power.ignore_children && !pm_runtime_blocked(dev->parent)) { + dev->power.smart_suspend = false; + return; + } + + idx = device_links_read_lock(); + + list_for_each_entry_rcu_locked(link, &dev->links.suppliers, c_node) { + if (!(link->flags | DL_FLAG_PM_RUNTIME)) + continue; + + if (!dev_pm_smart_suspend(link->supplier) && + !pm_runtime_blocked(link->supplier)) { + dev->power.smart_suspend = false; + break; + } + } + + device_links_read_unlock(idx); +} + /** * device_prepare - Prepare a device for system power transition. * @dev: Device to handle. @@ -1858,6 +1893,7 @@ unlock: pm_runtime_put(dev); return ret; } + device_prepare_smart_suspend(dev); /* * A positive return value from ->prepare() means "this device appears * to be runtime-suspended and its state is fine, so if it really is @@ -2033,6 +2069,5 @@ void device_pm_check_callbacks(struct device *dev) bool dev_pm_skip_suspend(struct device *dev) { - return dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND) && - pm_runtime_status_suspended(dev); + return dev_pm_smart_suspend(dev) && pm_runtime_status_suspended(dev); } diff --git a/drivers/mfd/intel-lpss.c b/drivers/mfd/intel-lpss.c index 3ba05ebb9035..63d6694f7145 100644 --- a/drivers/mfd/intel-lpss.c +++ b/drivers/mfd/intel-lpss.c @@ -480,7 +480,7 @@ EXPORT_SYMBOL_NS_GPL(intel_lpss_remove, "INTEL_LPSS"); static int resume_lpss_device(struct device *dev, void *data) { - if (!dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND)) + if (!dev_pm_smart_suspend(dev)) pm_runtime_resume(dev); return 0; diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c index f57ea36d125d..02726f36beb5 100644 --- a/drivers/pci/pci-driver.c +++ b/drivers/pci/pci-driver.c @@ -812,8 +812,7 @@ static int pci_pm_suspend(struct device *dev) * suspend callbacks can cope with runtime-suspended devices, it is * better to resume the device from runtime suspend here. */ - if (!dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND) || - pci_dev_need_resume(pci_dev)) { + if (!dev_pm_smart_suspend(dev) || pci_dev_need_resume(pci_dev)) { pm_runtime_resume(dev); pci_dev->state_saved = false; } else { @@ -1151,8 +1150,7 @@ static int pci_pm_poweroff(struct device *dev) } /* The reason to do that is the same as in pci_pm_suspend(). */ - if (!dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND) || - pci_dev_need_resume(pci_dev)) { + if (!dev_pm_smart_suspend(dev) || pci_dev_need_resume(pci_dev)) { pm_runtime_resume(dev); pci_dev->state_saved = false; } else { diff --git a/include/linux/device.h b/include/linux/device.h index 80a5b3268986..615282365052 100644 --- a/include/linux/device.h +++ b/include/linux/device.h @@ -1025,6 +1025,15 @@ static inline bool dev_pm_test_driver_flags(struct device *dev, u32 flags) return !!(dev->power.driver_flags & flags); } +static inline bool dev_pm_smart_suspend(struct device *dev) +{ +#ifdef CONFIG_PM_SLEEP + return dev->power.smart_suspend; +#else + return false; +#endif +} + static inline void device_lock(struct device *dev) { mutex_lock(&dev->mutex); diff --git a/include/linux/pm.h b/include/linux/pm.h index 6ca6f34c58c3..24647108f0ad 100644 --- a/include/linux/pm.h +++ b/include/linux/pm.h @@ -680,8 +680,8 @@ struct dev_pm_info { bool syscore:1; bool no_pm_callbacks:1; /* Owned by the PM core */ bool async_in_progress:1; /* Owned by the PM core */ + bool smart_suspend:1; /* Owned by the PM core */ bool must_resume:1; /* Owned by the PM core */ - bool set_active:1; /* Owned by the PM core */ bool may_skip_resume:1; /* Set by subsystems */ #else bool should_wakeup:1; -- 2.51.0 From 520a552f19d55825108ab83da093084c9afb50e9 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Tue, 18 Feb 2025 21:20:46 +0100 Subject: [PATCH 16/16] PM: sleep: Avoid unnecessary checks in device_prepare_smart_suspend() Add an optimization (on top of previous changes) to avoid calling pm_runtime_blocked(), which involves acquiring the device's PM spinlock, for devices with no PM callbacks and runtime PM "blocked". Signed-off-by: Rafael J. Wysocki Reviewed-by: Ulf Hansson Link: https://patch.msgid.link/2978873.e9J7NaK4W3@rjwysocki.net --- drivers/base/power/main.c | 16 +++++++++------- drivers/base/power/runtime.c | 9 +++++++-- include/linux/pm_runtime.h | 4 ++-- 3 files changed, 18 insertions(+), 11 deletions(-) diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index cc9903065900..a06ef91fbdb9 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -1796,16 +1796,14 @@ static void device_prepare_smart_suspend(struct device *dev) /* * The "smart suspend" feature is enabled for devices whose drivers ask - * for it and for devices without PM callbacks unless runtime PM is - * disabled and enabling it is blocked for them. + * for it and for devices without PM callbacks. * * However, if "smart suspend" is not enabled for the device's parent * or any of its suppliers that take runtime PM into account, it cannot * be enabled for the device either. */ - dev->power.smart_suspend = (dev->power.no_pm_callbacks || - dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND)) && - !pm_runtime_blocked(dev); + dev->power.smart_suspend = dev->power.no_pm_callbacks || + dev_pm_test_driver_flags(dev, DPM_FLAG_SMART_SUSPEND); if (!dev_pm_smart_suspend(dev)) return; @@ -1843,6 +1841,7 @@ static void device_prepare_smart_suspend(struct device *dev) static int device_prepare(struct device *dev, pm_message_t state) { int (*callback)(struct device *) = NULL; + bool no_runtime_pm; int ret = 0; /* @@ -1858,7 +1857,7 @@ static int device_prepare(struct device *dev, pm_message_t state) * suspend-resume cycle is complete, so prepare to trigger a warning on * subsequent attempts to enable it. */ - pm_runtime_block_if_disabled(dev); + no_runtime_pm = pm_runtime_block_if_disabled(dev); if (dev->power.syscore) return 0; @@ -1893,7 +1892,10 @@ unlock: pm_runtime_put(dev); return ret; } - device_prepare_smart_suspend(dev); + /* Do not enable "smart suspend" for devices without runtime PM. */ + if (!no_runtime_pm) + device_prepare_smart_suspend(dev); + /* * A positive return value from ->prepare() means "this device appears * to be runtime-suspended and its state is fine, so if it really is diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c index c0f5a9f89299..e772e45d30f3 100644 --- a/drivers/base/power/runtime.c +++ b/drivers/base/power/runtime.c @@ -1460,14 +1460,19 @@ int pm_runtime_barrier(struct device *dev) } EXPORT_SYMBOL_GPL(pm_runtime_barrier); -void pm_runtime_block_if_disabled(struct device *dev) +bool pm_runtime_block_if_disabled(struct device *dev) { + bool ret; + spin_lock_irq(&dev->power.lock); - if (dev->power.disable_depth && dev->power.last_status == RPM_INVALID) + ret = dev->power.disable_depth && dev->power.last_status == RPM_INVALID; + if (ret) dev->power.last_status = RPM_BLOCKED; spin_unlock_irq(&dev->power.lock); + + return ret; } void pm_runtime_unblock(struct device *dev) diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index aea0395c10a1..01ead602aedd 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -77,7 +77,7 @@ extern int pm_runtime_get_if_in_use(struct device *dev); extern int pm_schedule_suspend(struct device *dev, unsigned int delay); extern int __pm_runtime_set_status(struct device *dev, unsigned int status); extern int pm_runtime_barrier(struct device *dev); -extern void pm_runtime_block_if_disabled(struct device *dev); +extern bool pm_runtime_block_if_disabled(struct device *dev); extern void pm_runtime_unblock(struct device *dev); extern void pm_runtime_enable(struct device *dev); extern void __pm_runtime_disable(struct device *dev, bool check_resume); @@ -274,7 +274,7 @@ static inline int pm_runtime_get_if_active(struct device *dev) static inline int __pm_runtime_set_status(struct device *dev, unsigned int status) { return 0; } static inline int pm_runtime_barrier(struct device *dev) { return 0; } -static inline void pm_runtime_block_if_disabled(struct device *dev) {} +static inline bool pm_runtime_block_if_disabled(struct device *dev) { return true; } static inline void pm_runtime_unblock(struct device *dev) {} static inline void pm_runtime_enable(struct device *dev) {} static inline void __pm_runtime_disable(struct device *dev, bool c) {} -- 2.51.0