Merge tag 'x86-cleanups-2022-06-05' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

+49 -53

arch/x86/Kconfig

··· 41 41 depends on FUNCTION_TRACER 42 42 select DYNAMIC_FTRACE 43 43 help 44 - We keep the static function tracing (!DYNAMIC_FTRACE) around 45 - in order to test the non static function tracing in the 46 - generic code, as other architectures still use it. But we 47 - only need to keep it around for x86_64. No need to keep it 48 - for x86_32. For x86_32, force DYNAMIC_FTRACE. 44 + We keep the static function tracing (!DYNAMIC_FTRACE) around 45 + in order to test the non static function tracing in the 46 + generic code, as other architectures still use it. But we 47 + only need to keep it around for x86_64. No need to keep it 48 + for x86_32. For x86_32, force DYNAMIC_FTRACE. 49 49 # 50 50 # Arch settings 51 51 # ··· 394 394 default $(success,$(srctree)/scripts/gcc-x86_64-has-stack-protector.sh $(CC)) if 64BIT 395 395 default $(success,$(srctree)/scripts/gcc-x86_32-has-stack-protector.sh $(CC)) 396 396 help 397 - We have to make sure stack protector is unconditionally disabled if 398 - the compiler produces broken code or if it does not let us control 399 - the segment on 32-bit kernels. 397 + We have to make sure stack protector is unconditionally disabled if 398 + the compiler produces broken code or if it does not let us control 399 + the segment on 32-bit kernels. 400 400 401 401 menu "Processor type and features" 402 402 ··· 532 532 533 533 If you have one of these systems, or if you want to build a 534 534 generic distribution kernel, say Y here - otherwise say N. 535 - endif 535 + endif # X86_32 536 536 537 537 if X86_64 538 538 config X86_EXTENDED_PLATFORM ··· 551 551 552 552 If you have one of these systems, or if you want to build a 553 553 generic distribution kernel, say Y here - otherwise say N. 554 - endif 554 + endif # X86_64 555 555 # This is an alphabetically sorted list of 64 bit extended platforms 556 556 # Please maintain the alphabetic order if and when there are additions 557 557 config X86_NUMACHIP ··· 599 599 bool "Goldfish (Virtual Platform)" 600 600 depends on X86_EXTENDED_PLATFORM 601 601 help 602 - Enable support for the Goldfish virtual platform used primarily 603 - for Android development. Unless you are building for the Android 604 - Goldfish emulator say N here. 602 + Enable support for the Goldfish virtual platform used primarily 603 + for Android development. Unless you are building for the Android 604 + Goldfish emulator say N here. 605 605 606 606 config X86_INTEL_CE 607 607 bool "CE4100 TV platform" ··· 900 900 memory contents and CPU state. TDX guests are protected from 901 901 some attacks from the VMM. 902 902 903 - endif #HYPERVISOR_GUEST 903 + endif # HYPERVISOR_GUEST 904 904 905 905 source "arch/x86/Kconfig.cpu" 906 906 ··· 1167 1167 prompt "Intel MCE features" 1168 1168 depends on X86_MCE && X86_LOCAL_APIC 1169 1169 help 1170 - Additional support for intel specific MCE features such as 1171 - the thermal monitor. 1170 + Additional support for intel specific MCE features such as 1171 + the thermal monitor. 1172 1172 1173 1173 config X86_MCE_AMD 1174 1174 def_bool y 1175 1175 prompt "AMD MCE features" 1176 1176 depends on X86_MCE && X86_LOCAL_APIC && AMD_NB 1177 1177 help 1178 - Additional support for AMD specific MCE features such as 1179 - the DRAM Error Threshold. 1178 + Additional support for AMD specific MCE features such as 1179 + the DRAM Error Threshold. 1180 1180 1181 1181 config X86_ANCIENT_MCE 1182 1182 bool "Support for old Pentium 5 / WinChip machine checks" ··· 1254 1254 default y 1255 1255 depends on X86_64 1256 1256 help 1257 - This enables emulation of the legacy vsyscall page. Disabling 1258 - it is roughly equivalent to booting with vsyscall=none, except 1259 - that it will also disable the helpful warning if a program 1260 - tries to use a vsyscall. With this option set to N, offending 1261 - programs will just segfault, citing addresses of the form 1262 - 0xffffffffff600?00. 1257 + This enables emulation of the legacy vsyscall page. Disabling 1258 + it is roughly equivalent to booting with vsyscall=none, except 1259 + that it will also disable the helpful warning if a program 1260 + tries to use a vsyscall. With this option set to N, offending 1261 + programs will just segfault, citing addresses of the form 1262 + 0xffffffffff600?00. 1263 1263 1264 - This option is required by many programs built before 2013, and 1265 - care should be used even with newer programs if set to N. 1264 + This option is required by many programs built before 2013, and 1265 + care should be used even with newer programs if set to N. 1266 1266 1267 - Disabling this option saves about 7K of kernel size and 1268 - possibly 4K of additional runtime pagetable memory. 1267 + Disabling this option saves about 7K of kernel size and 1268 + possibly 4K of additional runtime pagetable memory. 1269 1269 1270 1270 config X86_IOPL_IOPERM 1271 1271 bool "IOPERM and IOPL Emulation" ··· 2002 2002 bool "EFI mixed-mode support" 2003 2003 depends on EFI_STUB && X86_64 2004 2004 help 2005 - Enabling this feature allows a 64-bit kernel to be booted 2006 - on a 32-bit firmware, provided that your CPU supports 64-bit 2007 - mode. 2005 + Enabling this feature allows a 64-bit kernel to be booted 2006 + on a 32-bit firmware, provided that your CPU supports 64-bit 2007 + mode. 2008 2008 2009 - Note that it is not possible to boot a mixed-mode enabled 2010 - kernel via the EFI boot stub - a bootloader that supports 2011 - the EFI handover protocol must be used. 2009 + Note that it is not possible to boot a mixed-mode enabled 2010 + kernel via the EFI boot stub - a bootloader that supports 2011 + the EFI handover protocol must be used. 2012 2012 2013 - If unsure, say N. 2013 + If unsure, say N. 2014 2014 2015 2015 source "kernel/Kconfig.hz" 2016 2016 ··· 2235 2235 select DYNAMIC_MEMORY_LAYOUT 2236 2236 default RANDOMIZE_BASE 2237 2237 help 2238 - Randomizes the base virtual address of kernel memory sections 2239 - (physical memory mapping, vmalloc & vmemmap). This security feature 2240 - makes exploits relying on predictable memory locations less reliable. 2238 + Randomizes the base virtual address of kernel memory sections 2239 + (physical memory mapping, vmalloc & vmemmap). This security feature 2240 + makes exploits relying on predictable memory locations less reliable. 2241 2241 2242 - The order of allocations remains unchanged. Entropy is generated in 2243 - the same way as RANDOMIZE_BASE. Current implementation in the optimal 2244 - configuration have in average 30,000 different possible virtual 2245 - addresses for each memory section. 2242 + The order of allocations remains unchanged. Entropy is generated in 2243 + the same way as RANDOMIZE_BASE. Current implementation in the optimal 2244 + configuration have in average 30,000 different possible virtual 2245 + addresses for each memory section. 2246 2246 2247 - If unsure, say Y. 2247 + If unsure, say Y. 2248 2248 2249 2249 config RANDOMIZE_MEMORY_PHYSICAL_PADDING 2250 2250 hex "Physical memory mapping padding" if EXPERT ··· 2254 2254 range 0x1 0x40 if MEMORY_HOTPLUG 2255 2255 range 0x0 0x40 2256 2256 help 2257 - Define the padding in terabytes added to the existing physical 2258 - memory size during kernel memory randomization. It is useful 2259 - for memory hotplug support but reduces the entropy available for 2260 - address randomization. 2257 + Define the padding in terabytes added to the existing physical 2258 + memory size during kernel memory randomization. It is useful 2259 + for memory hotplug support but reduces the entropy available for 2260 + address randomization. 2261 2261 2262 - If unsure, leave at the default value. 2262 + If unsure, leave at the default value. 2263 2263 2264 2264 config HOTPLUG_CPU 2265 2265 def_bool y ··· 2606 2606 2607 2607 endmenu 2608 2608 2609 - 2610 2609 menu "Bus options (PCI etc.)" 2611 2610 2612 2611 choice ··· 2829 2830 2830 2831 endmenu 2831 2832 2832 - 2833 2833 menu "Binary Emulations" 2834 2834 2835 2835 config IA32_EMULATION ··· 2866 2868 def_bool y 2867 2869 depends on IA32_EMULATION || X86_X32_ABI 2868 2870 2869 - if COMPAT 2870 2871 config COMPAT_FOR_U64_ALIGNMENT 2871 2872 def_bool y 2872 - endif 2873 + depends on COMPAT 2873 2874 2874 2875 endmenu 2875 - 2876 2876 2877 2877 config HAVE_ATOMIC_IOMAP 2878 2878 def_bool y

+14 -15

arch/x86/Kconfig.debug

··· 73 73 bool "Set upper limit of TLB entries to flush one-by-one" 74 74 depends on DEBUG_KERNEL 75 75 help 76 + X86-only for now. 76 77 77 - X86-only for now. 78 + This option allows the user to tune the amount of TLB entries the 79 + kernel flushes one-by-one instead of doing a full TLB flush. In 80 + certain situations, the former is cheaper. This is controlled by the 81 + tlb_flushall_shift knob under /sys/kernel/debug/x86. If you set it 82 + to -1, the code flushes the whole TLB unconditionally. Otherwise, 83 + for positive values of it, the kernel will use single TLB entry 84 + invalidating instructions according to the following formula: 78 85 79 - This option allows the user to tune the amount of TLB entries the 80 - kernel flushes one-by-one instead of doing a full TLB flush. In 81 - certain situations, the former is cheaper. This is controlled by the 82 - tlb_flushall_shift knob under /sys/kernel/debug/x86. If you set it 83 - to -1, the code flushes the whole TLB unconditionally. Otherwise, 84 - for positive values of it, the kernel will use single TLB entry 85 - invalidating instructions according to the following formula: 86 + flush_entries <= active_tlb_entries / 2^tlb_flushall_shift 86 87 87 - flush_entries <= active_tlb_entries / 2^tlb_flushall_shift 88 - 89 - If in doubt, say "N". 88 + If in doubt, say "N". 90 89 91 90 config IOMMU_DEBUG 92 91 bool "Enable IOMMU debugging" ··· 118 119 depends on DEBUG_KERNEL && INSTRUCTION_DECODER 119 120 depends on !COMPILE_TEST 120 121 help 121 - Perform x86 instruction decoder selftests at build time. 122 - This option is useful for checking the sanity of x86 instruction 123 - decoder code. 124 - If unsure, say "N". 122 + Perform x86 instruction decoder selftests at build time. 123 + This option is useful for checking the sanity of x86 instruction 124 + decoder code. 125 + If unsure, say "N". 125 126 126 127 choice 127 128 prompt "IO delay type"

+1 -1

arch/x86/kvm/vmx/vmx.c

··· 6219 6219 int size = PAGE_SIZE << L1D_CACHE_ORDER; 6220 6220 6221 6221 /* 6222 - * This code is only executed when the the flush mode is 'cond' or 6222 + * This code is only executed when the flush mode is 'cond' or 6223 6223 * 'always' 6224 6224 */ 6225 6225 if (static_branch_likely(&vmx_l1d_flush_cond)) {

+1 -1

arch/x86/kvm/x86.c

··· 11937 11937 if (current->mm == kvm->mm) { 11938 11938 /* 11939 11939 * Free memory regions allocated on behalf of userspace, 11940 - * unless the the memory map has changed due to process exit 11940 + * unless the memory map has changed due to process exit 11941 11941 * or fd copying. 11942 11942 */ 11943 11943 mutex_lock(&kvm->slots_lock);

+1 -1

arch/x86/platform/efi/efi_thunk_64.S

··· 8 8 * The below thunking functions are only used after ExitBootServices() 9 9 * has been called. This simplifies things considerably as compared with 10 10 * the early EFI thunking because we can leave all the kernel state 11 - * intact (GDT, IDT, etc) and simply invoke the the 32-bit EFI runtime 11 + * intact (GDT, IDT, etc) and simply invoke the 32-bit EFI runtime 12 12 * services from __KERNEL32_CS. This means we can continue to service 13 13 * interrupts across an EFI mixed mode call. 14 14 *

Configure Feed

Configure Feed