x86/cpu: Do some janitorial work
Subsystems (4)
Borislav Petkov
maintainerv2 cover x86/cpu: Do some janitorial work
From: Borislav Petkov <[email protected]>
Hi all,
here's v2 with some adjustments due to dependent changes which went upstream, in the meantime. I figure I should send them out one more time before I queue them.
Also, patch 1 handles printing of naked feature numbers for which we don't generate a string for /proc/cpuinfo now too.
Thx.
changelog:
----------
v1
--
Beef up clearcpuid= so that it supports the strings from /proc/cpuinfo too and remove a bunch of useless, and, in some cases even downright dangerous, chicken bits we have added over the years.
Borislav Petkov (6):
x86/cpu: Allow feature bit names from /proc/cpuinfo in clearcpuid= x86/cpu: Remove "nosep"
x86/cpu: Remove CONFIG_X86_SMAP and "nosmap"
x86/cpu: Remove "nosmep"
x86/cpu: Remove "noexec"
x86/cpu: Remove "noclflush"
.../admin-guide/kernel-parameters.txt | 24 ++--- Documentation/x86/cpuinfo.rst | 5 +- Documentation/x86/x86_64/boot-options.rst | 9 -- arch/x86/Kconfig | 11 --- arch/x86/include/asm/cpufeature.h | 7 +- arch/x86/include/asm/disabled-features.h | 8 +- arch/x86/include/asm/proto.h | 1 - arch/x86/include/asm/smap.h | 24 ----- arch/x86/kernel/cpu/common.c | 94 +++++++++---------- arch/x86/kernel/setup.c | 28 +++++- arch/x86/mm/Makefile | 3 +- arch/x86/mm/init_64.c | 1 - arch/x86/mm/setup_nx.c | 62 ------------ scripts/Makefile.build | 1 - scripts/link-vmlinux.sh | 3 - .../arch/x86/include/asm/disabled-features.h | 8 +- 16 files changed, 91 insertions(+), 198 deletions(-) delete mode 100644 arch/x86/mm/setup_nx.c
Signature · 1 line
2.29.2
Borislav Petkov
maintainerPATCH v2 1/6 x86/cpu: Allow feature bit names from /proc/cpuinfo in clearcpuid=
From: Borislav Petkov <[email protected]>
Having to give the X86_FEATURE array indices in order to disable a feature bit for testing is not really user-friendly. So accept the feature bit names too.
Some feature bits don't have names so there the array indices are still accepted, of course.
Clearing CPUID flags is not something which should be done in production so taint the kernel too.
An exemplary cmdline would then be something like:
clearcpuid=de,440,smca,succory,bmi1,3dnow
("succory" is wrong on purpose). And it says:
[ 0.000000] Clearing CPUID bits: de 13:24 smca bmi1 3dnow
--- .../admin-guide/kernel-parameters.txt | 11 +++- arch/x86/include/asm/cpufeature.h | 7 ++- arch/x86/kernel/cpu/common.c | 57 +++++++++++++++---- 3 files changed, 58 insertions(+), 17 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +8 -3
@@ -631,12 +631,17 @@ Defaults to zero when built as a module and to 10 seconds when built into the kernel. - clearcpuid=BITNUM[,BITNUM...] [X86] + clearcpuid=X[,X...] [X86] Disable CPUID feature X for the kernel. See arch/x86/include/asm/cpufeatures.h for the valid bit - numbers. Note the Linux specific bits are not necessarily - stable over kernel options, but the vendor specific + numbers X. Note the Linux-specific bits are not necessarily + stable over kernel options, but the vendor-specific ones should be. + X can also be a string as appearing in the flags: line + in /proc/cpuinfo which does not have the above + instability issue. However, not all features have names + in /proc/cpuinfo. + Note that using this option will taint your kernel. Also note that user programs calling CPUID directly or using the feature without checking anything will still see it. This just prevents it from
arch/x86/include/asm/cpufeature.h +5 -2
@@ -34,14 +34,17 @@ enum cpuid_leafs CPUID_8000_001F_EAX, }; +#define X86_CAP_FMT_NUM "%d:%d" +#define x86_cap_flag_num (flag) ((flag) >> 5), ((flag) & 31) +#ifdef CONFIG_X86_FEATURE_NAMES extern const char *const x86_cap_flags[NCAPINTS *32 ];extern const char *const x86_power_flags[32 ];#define X86_CAP_FMT "%s"#define x86_cap_flag (flag) x86_cap_flags[flag]#else -#define X86_CAP_FMT "%d:%d" -#define x86_cap_flag (flag) ((flag) >> 5), ((flag) & 31) +#define X86_CAP_FMT X86_CAP_FMT_NUM +#define x86_cap_flag x86_cap_flag_num#endif /*
arch/x86/kernel/cpu/common.c +45 -12
74 lines not on this page · raw message
Signature · 1 line
2.29.2
Re: [PATCH v2 1/6] x86/cpu: Allow feature bit names from /proc/cpuinfo in clearcpuid=
Quoted 14 lines · Borislav Petkov
From: Borislav Petkov <[email protected]>
Having to give the
X86_FEATUREarray indices in order to disable a feature bit for testing is not really user-friendly. So accept the feature bit names too.Some feature bits don't have names so there the array indices are still accepted, of course.
Clearing CPUID flags is not something which should be done in production so taint the kernel too.
An exemplary cmdline would then be something like:
Quoted 7 lines
clearcpuid=de,440,smca,succory,bmi1,3dnow
("succory" is wrong on purpose). And it says:
[ 0.000000] Clearing CPUID bits: de 13:24 smca bmi1 3dnowSigned-off-by Borislav Petkov <[email protected]>
I like the taint flag addition!
Even though it reports what it does actually clear, do you think it might be more "friendly" to yell about unknown stuff too? i.e.:
[ 0.000000] Clearing CPUID bits: unknown bit 'succory' [ 0.000000] Clearing CPUID bits: de 13:24 smca bmi1 3dnow
Either way:
Signature · 1 line
Kees Cook
Re: [PATCH v2 1/6] x86/cpu: Allow feature bit names from /proc/cpuinfo in clearcpuid=
Quoted 1 line · Kees Cook
I like the taint flag addition!
IKR. See below.
Quoted 2 lines
Even though it reports what it does actually clear, do you think it might be more "friendly" to yell about unknown stuff too? i.e.:
Well, this thing is not supposed to be used by normal users anyway. There was even talk of completely removing it but CPU folks wanna do some experiments with it, that's why we kept it.
And for exactly the same reason it taints the kernel - so that we know that it is an "out-of-spec" situation.
I even caused an explosion while testing it due to creating an impossible configuration of features.
So, folks who use it better know what they're doing.
Quoted 2 lines
[ 0.000000] Clearing CPUID bits: unknown bit 'succory' [ 0.000000] Clearing CPUID bits: de 13:24 smca bmi1 3dnow
How's the below (I don't want to disrupt the pr_cont flow too much):
On the cmdline I have:
... clearcpuid=de,440,smca,succory,bmi1,3dnow,bla,foo ...
which says:
[ 0.000000] Clearing CPUID bits: de 13:24 smca (unknown: succory) bmi1 3dnow (unknown: bla) (unknown: foo)
Quoted 1 line
Reviewed-by Kees Cook <[email protected]>
Thx.
---
From: Borislav Petkov <[email protected]>
Date: Thu, 27 Jan 2022 12:56:21 +0100
Subject: [PATCH] x86/cpu: Allow feature bit names from /proc/cpuinfo in clearcpuid=
Having to give the X86_FEATURE array indices in order to disable a feature bit for testing is not really user-friendly. So accept the feature bit names too.
Some feature bits don't have names so there the array indices are still accepted, of course.
Clearing CPUID flags is not something which should be done in production so taint the kernel too.
An exemplary cmdline would then be something like:
clearcpuid=de,440,smca,succory,bmi1,3dnow
("succory" is wrong on purpose). And it says:
[ ... ] Clearing CPUID bits: de 13:24 smca (unknown: succory) bmi1 3dnow
--- .../admin-guide/kernel-parameters.txt | 11 +++- arch/x86/include/asm/cpufeature.h | 7 ++- arch/x86/kernel/cpu/common.c | 62 +++++++++++++++---- 3 files changed, 63 insertions(+), 17 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +8 -3
@@ -631,12 +631,17 @@ Defaults to zero when built as a module and to 10 seconds when built into the kernel. - clearcpuid=BITNUM[,BITNUM...] [X86] + clearcpuid=X[,X...] [X86] Disable CPUID feature X for the kernel. See arch/x86/include/asm/cpufeatures.h for the valid bit - numbers. Note the Linux specific bits are not necessarily - stable over kernel options, but the vendor specific + numbers X. Note the Linux-specific bits are not necessarily + stable over kernel options, but the vendor-specific ones should be. + X can also be a string as appearing in the flags: line + in /proc/cpuinfo which does not have the above + instability issue. However, not all features have names + in /proc/cpuinfo. + Note that using this option will taint your kernel. Also note that user programs calling CPUID directly or using the feature without checking anything will still see it. This just prevents it from
arch/x86/include/asm/cpufeature.h +5 -2
@@ -34,14 +34,17 @@ enum cpuid_leafs CPUID_8000_001F_EAX, }; +#define X86_CAP_FMT_NUM "%d:%d" +#define x86_cap_flag_num (flag) ((flag) >> 5), ((flag) & 31) +#ifdef CONFIG_X86_FEATURE_NAMES extern const char *const x86_cap_flags[NCAPINTS *32 ];extern const char *const x86_power_flags[32 ];#define X86_CAP_FMT "%s"#define x86_cap_flag (flag) x86_cap_flags[flag]#else -#define X86_CAP_FMT "%d:%d" -#define x86_cap_flag (flag) ((flag) >> 5), ((flag) & 31) +#define X86_CAP_FMT X86_CAP_FMT_NUM +#define x86_cap_flag x86_cap_flag_num#endif /*
arch/x86/kernel/cpu/common.c +50 -12
79 lines not on this page · raw message
Signature · 7 lines
2.29.2
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette[tip: x86/cpu] x86/cpu: Allow feature bit names from /proc/cpuinfo in clearcpuid=
In reply to Borislav PetkovThe following commit has been merged into the x86/cpu branch of tip:
The following commit has been merged into the x86/cpu branch of tip:
Commit-ID: 1625c833db93516faaac5feedadf8d19c14238b6 Gitweb: https://git.kernel.org/tip/1625c833db93516faaac5feedadf8d19c14238b6 Author: Borislav Petkov <[email protected]> AuthorDate: Thu, 27 Jan 2022 12:56:21 +01:00 Committer: Borislav Petkov <[email protected]> CommitterDate: Mon, 04 Apr 2022 10:16:52 +02:00
x86/cpu: Allow feature bit names from /proc/cpuinfo in clearcpuid=
Having to give the X86_FEATURE array indices in order to disable a feature bit for testing is not really user-friendly. So accept the feature bit names too.
Some feature bits don't have names so there the array indices are still accepted, of course.
Clearing CPUID flags is not something which should be done in production so taint the kernel too.
An exemplary cmdline would then be something like:
clearcpuid=de,440,smca,succory,bmi1,3dnow
("succory" is wrong on purpose). And it says:
[ ... ] Clearing CPUID bits: de 13:24 smca (unknown: succory) bmi1 3dnow
[ Fix CONFIG_X86_FEATURE_NAMES=n build error as reported by the 0day
robot: https://lore.kernel.org/r/[email protected] ]
--- Documentation/admin-guide/kernel-parameters.txt | 11 ++- arch/x86/include/asm/cpufeature.h | 7 +- arch/x86/kernel/cpu/common.c | 64 +++++++++++++--- 3 files changed, 65 insertions(+), 17 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +8 -3
@@ -631,12 +631,17 @@ Defaults to zero when built as a module and to 10 seconds when built into the kernel. - clearcpuid=BITNUM[,BITNUM...] [X86] + clearcpuid=X[,X...] [X86] Disable CPUID feature X for the kernel. See arch/x86/include/asm/cpufeatures.h for the valid bit - numbers. Note the Linux specific bits are not necessarily - stable over kernel options, but the vendor specific + numbers X. Note the Linux-specific bits are not necessarily + stable over kernel options, but the vendor-specific ones should be. + X can also be a string as appearing in the flags: line + in /proc/cpuinfo which does not have the above + instability issue. However, not all features have names + in /proc/cpuinfo. + Note that using this option will taint your kernel. Also note that user programs calling CPUID directly or using the feature without checking anything will still see it. This just prevents it from
arch/x86/include/asm/cpufeature.h +5 -2
@@ -34,14 +34,17 @@ enum cpuid_leafs CPUID_8000_001F_EAX, }; +#define X86_CAP_FMT_NUM "%d:%d" +#define x86_cap_flag_num (flag) ((flag) >> 5), ((flag) & 31) +#ifdef CONFIG_X86_FEATURE_NAMES extern const char *const x86_cap_flags[NCAPINTS *32 ];extern const char *const x86_power_flags[32 ];#define X86_CAP_FMT "%s"#define x86_cap_flag (flag) x86_cap_flags[flag]#else -#define X86_CAP_FMT "%d:%d" -#define x86_cap_flag (flag) ((flag) >> 5), ((flag) & 31) +#define X86_CAP_FMT X86_CAP_FMT_NUM +#define x86_cap_flag x86_cap_flag_num#endif /*
arch/x86/kernel/cpu/common.c +52 -12
81 lines not on this page · raw message
Borislav Petkov
maintainerPATCH v2 2/6 x86/cpu: Remove "nosep"
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
That chicken bit was added by
4f88651125e2 ("[PATCH] i386: allow disabling X86_FEATURE_SEP at boot")
but measuring int80 vsyscall performance on 32-bit doesn't matter anymore.
If still needed, one can boot with
clearcpuid=sep
to disable that feature for testing.
--- Documentation/admin-guide/kernel-parameters.txt | 2 -- arch/x86/kernel/cpu/common.c | 7 ------- 2 files changed, 9 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +0 -2
@@ -3599,8 +3599,6 @@ nosbagart [IA-64] - nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support. - nosgx [X86-64,SGX] Disables Intel SGX kernel support. nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
arch/x86/kernel/cpu/common.c +0 -7
@@ -220,13 +220,6 @@ static int __init cachesize_setup(char *str) }__setup ("cachesize=" , cachesize_setup); -static int __init x86_sep_setup (char *s) -{ -setup_clear_cpu_cap (X86_FEATURE_SEP ); -return 1 ; -} -__setup ("nosep" , x86_sep_setup); -/* Standard macro to see if a specific flag is changeable */ static inline int flag_is_changeable_p (u32 flag) {
Signature · 1 line
2.29.2
Re: [PATCH v2 2/6] x86/cpu: Remove "nosep"
Quoted 9 lines · Borislav Petkov
From: Borislav Petkov <[email protected]>
That chicken bit was added by
4f88651125e2("[PATCH] i386: allow disablingX86_FEATURE_SEPat boot")but measuring int80 vsyscall performance on 32-bit doesn't matter anymore.
Quoted 7 lines
If still needed, one can boot with
clearcpuid=sep
to disable that feature for testing.
Signed-off-by Borislav Petkov <[email protected]>
Signature · 1 line
Kees Cook
[tip: x86/cpu] x86/cpu: Remove "nosep"
In reply to Borislav PetkovThe following commit has been merged into the x86/cpu branch of tip:
The following commit has been merged into the x86/cpu branch of tip:
Commit-ID: c949110ef4e31cb5d3387bd8273fd5de66b5227b Gitweb: https://git.kernel.org/tip/c949110ef4e31cb5d3387bd8273fd5de66b5227b Author: Borislav Petkov <[email protected]> AuthorDate: Thu, 27 Jan 2022 12:56:22 +01:00 Committer: Borislav Petkov <[email protected]> CommitterDate: Mon, 04 Apr 2022 10:16:55 +02:00
x86/cpu: Remove "nosep"
That chicken bit was added by
4f88651125e2 ("[PATCH] i386: allow disabling X86_FEATURE_SEP at boot")
but measuring int80 vsyscall performance on 32-bit doesn't matter anymore.
If still needed, one can boot with
clearcpuid=sep
to disable that feature for testing.
--- Documentation/admin-guide/kernel-parameters.txt | 2 -- arch/x86/kernel/cpu/common.c | 7 ------- 2 files changed, 9 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +0 -2
@@ -3665,8 +3665,6 @@ nosbagart [IA-64] - nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support. - nosgx [X86-64,SGX] Disables Intel SGX kernel support. nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
arch/x86/kernel/cpu/common.c +0 -7
@@ -298,13 +298,6 @@ static int __init cachesize_setup(char *str) }__setup ("cachesize=" , cachesize_setup); -static int __init x86_sep_setup (char *s) -{ -setup_clear_cpu_cap (X86_FEATURE_SEP ); -return 1 ; -} -__setup ("nosep" , x86_sep_setup); -/* Standard macro to see if a specific flag is changeable */ static inline int flag_is_changeable_p (u32 flag) {
Borislav Petkov
maintainerPATCH v2 3/6 x86/cpu: Remove CONFIG_X86_SMAP and "nosmap"
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
Those were added as part of the SMAP enablement but SMAP is currently an integral part of kernel proper and there's no need to disable it anymore.
Rip out that functionality.
If still needed - clearcpuid=smap.
--- .../admin-guide/kernel-parameters.txt | 2 +- Documentation/x86/cpuinfo.rst | 5 ++-- arch/x86/Kconfig | 11 --------- arch/x86/include/asm/disabled-features.h | 8 +------ arch/x86/include/asm/smap.h | 24 ------------------- arch/x86/kernel/cpu/common.c | 15 +----------- scripts/Makefile.build | 1 - scripts/link-vmlinux.sh | 3 --- .../arch/x86/include/asm/disabled-features.h | 8 +------ 9 files changed, 6 insertions(+), 71 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +1 -1
@@ -3394,7 +3394,7 @@ noexec=on: enable non-executable mappings (default) noexec=off: disable non-executable mappings - nosmap [X86,PPC] + nosmap [PPC] Disable SMAP (Supervisor Mode Access Prevention) even if it is supported by processor.
Documentation/x86/cpuinfo.rst +2 -3
@@ -140,9 +140,8 @@ from #define X86_FEATURE_UMIP (16*32 + 2). In addition, there exists a variety of custom command-line parameters that disable specific features. The list of parameters includes, but is not limited -to, nofsgsbase, nosmap, and nosmep. 5-level paging can also be disabled using -"no5lvl". SMAP and SMEP are disabled with the aforementioned parameters, -respectively. +to, nofsgsbase, and nosmep. 5-level paging can also be disabled using +"no5lvl". SMEP is disabled with the aforementioned parameter. e: The feature was known to be non-functional. ----------------------------------------------
arch/x86/Kconfig +0 -11
@@ -1834,17 +1834,6 @@ config ARCH_RANDOM If supported, this is a high bandwidth, cryptographically secure hardware random number generator. -config X86_SMAP - def_bool y - prompt "Supervisor Mode Access Prevention" if EXPERT - help - Supervisor Mode Access Prevention (SMAP) is a security - feature in newer Intel processors. There is a small - performance cost if this enabled and turned on; there is - also a small increase in the kernel size if this is enabled. - - If unsure, say Y. - config X86_UMIP def_bool y prompt "User Mode Instruction Prevention" if EXPERT
arch/x86/include/asm/disabled-features.h +1 -7
@@ -10,12 +10,6 @@* cpu_feature_enabled(). */ -#ifdef CONFIG_X86_SMAP -# define DISABLE_SMAP 0 -#else -# define DISABLE_SMAP (1<<(X86_FEATURE_SMAP & 31)) -#endif -#ifdef CONFIG_X86_UMIP # define DISABLE_UMIP 0#else
@@ -77,7 +71,7 @@#define DISABLED_MASK6 0#define DISABLED_MASK7 (DISABLE_PTI)#define DISABLED_MASK8 0 -#define DISABLED_MASK9 (DISABLE_SMAP|DISABLE_SGX) +#define DISABLED_MASK9 (DISABLE_SGX)#define DISABLED_MASK10 0#define DISABLED_MASK11 0#define DISABLED_MASK12 0
arch/x86/include/asm/smap.h +0 -24
@@ -19,25 +19,14 @@#ifdef __ASSEMBLY__ -#ifdef CONFIG_X86_SMAP -#define ASM_CLAC \ ALTERNATIVE "", __ASM_CLAC, X86_FEATURE_SMAP#define ASM_STAC \ ALTERNATIVE "", __ASM_STAC, X86_FEATURE_SMAP -#else /* CONFIG_X86_SMAP */ - -#define ASM_CLAC -#define ASM_STAC - -#endif /* CONFIG_X86_SMAP */ -#else /* __ASSEMBLY__ */ -#ifdef CONFIG_X86_SMAP - static __always_inline void clac(void) {/* Note: a barrier is implicit in alternative() */
@@ -76,19 +65,6 @@ static __always_inline void smap_restore(unsigned long flags)#define ASM_STAC \ ALTERNATIVE("", __ASM_STAC, X86_FEATURE_SMAP) -#else /* CONFIG_X86_SMAP */ - -static inline void clac (void ) { } -static inline void stac (void ) { } - -static inline unsigned long smap_save (void ) {return 0 ; } -static inline void smap_restore (unsigned long flags) { } - -#define ASM_CLAC -#define ASM_STAC - -#endif /* CONFIG_X86_SMAP */ -#endif /* __ASSEMBLY__ */ #endif /* _ASM_X86_SMAP_H */
arch/x86/kernel/cpu/common.c +1 -14
@@ -304,13 +304,6 @@ static __always_inline void setup_smep(struct cpuinfo_x86 *c)cr4_set_bits (X86_CR4_SMEP ); } -static __init int setup_disable_smap (char *arg) -{ -setup_clear_cpu_cap (X86_FEATURE_SMAP ); -return 1 ; -} -__setup ("nosmap" , setup_disable_smap); - static __always_inline void setup_smap(struct cpuinfo_x86 *c) {unsigned long eflags =native_save_fl ();
@@ -318,14 +311,8 @@ static __always_inline void setup_smap(struct cpuinfo_x86 *c)/* This should have been cleared long ago */ BUG_ON (eflags &X86_EFLAGS_AC ); -if (cpu_has (c,X86_FEATURE_SMAP )) { -#ifdef CONFIG_X86_SMAP +if (cpu_has (c,X86_FEATURE_SMAP ))cr4_set_bits (X86_CR4_SMAP ); -#else -clear_cpu_cap (c,X86_FEATURE_SMAP ); -cr4_clear_bits (X86_CR4_SMAP ); -#endif - } } static __always_inline void setup_umip(struct cpuinfo_x86 *c)
scripts/Makefile.build +0 -1
@@ -233,7 +233,6 @@ objtool_args = \ $(if $(CONFIG_FRAME_POINTER),, --no-fp) \ $(if $(CONFIG_GCOV_KERNEL)$(CONFIG_LTO_CLANG), --no-unreachable)\ $(if $(CONFIG_RETPOLINE), --retpoline) \ - $(if $(CONFIG_X86_SMAP), --uaccess) \ $(if $(CONFIG_FTRACE_MCOUNT_USE_OBJTOOL), --mcount) \ $(if $(CONFIG_SLS), --sls)
scripts/link-vmlinux.sh +0 -3
@@ -140,9 +140,6 @@ objtool_link() if is_enabled CONFIG_RETPOLINE; then objtoolopt="${objtoolopt} --retpoline" fi - if is_enabled CONFIG_X86_SMAP; then - objtoolopt="${objtoolopt} --uaccess" - fi if is_enabled CONFIG_SLS; then objtoolopt="${objtoolopt} --sls" fi
tools/arch/x86/include/asm/disabled-features.h +1 -7
@@ -10,12 +10,6 @@* cpu_feature_enabled(). */ -#ifdef CONFIG_X86_SMAP -# define DISABLE_SMAP 0 -#else -# define DISABLE_SMAP (1<<(X86_FEATURE_SMAP & 31)) -#endif -#ifdef CONFIG_X86_UMIP # define DISABLE_UMIP 0#else
@@ -77,7 +71,7 @@#define DISABLED_MASK6 0#define DISABLED_MASK7 (DISABLE_PTI)#define DISABLED_MASK8 0 -#define DISABLED_MASK9 (DISABLE_SMAP|DISABLE_SGX) +#define DISABLED_MASK9 (DISABLE_SGX)#define DISABLED_MASK10 0#define DISABLED_MASK11 0#define DISABLED_MASK12 0
Signature · 1 line
2.29.2
Re: [PATCH v2 3/6] x86/cpu: Remove CONFIG_X86_SMAP and "nosmap"
Quoted 41 lines · Borislav Petkov
From: Borislav Petkov <[email protected]>
Those were added as part of the SMAP enablement but SMAP is currently an integral part of kernel proper and there's no need to disable it anymore.
Rip out that functionality.
If still needed - clearcpuid=smap.
Signed-off-by Borislav Petkov <[email protected]>Reviewed-by Lai Jiangshan <[email protected]>--- .../admin-guide/kernel-parameters.txt | 2 +- Documentation/x86/cpuinfo.rst | 5 ++-- arch/x86/Kconfig | 11 --------- arch/x86/include/asm/disabled-features.h | 8 +------ arch/x86/include/asm/smap.h | 24 ------------------- arch/x86/kernel/cpu/common.c | 15 +----------- scripts/Makefile.build | 1 - scripts/link-vmlinux.sh | 3 --- .../arch/x86/include/asm/disabled-features.h | 8 +------ 9 files changed, 6 insertions(+), 71 deletions(-)[...]
scripts/Makefile.build +0
-1@@ -233,7 +233,6 @@ objtool_args = \ $(if $(CONFIG_FRAME_POINTER),, --no-fp) \ $(if $(CONFIG_GCOV_KERNEL)$(CONFIG_LTO_CLANG), --no-unreachable)\ $(if $(CONFIG_RETPOLINE), --retpoline) \ - $(if $(CONFIG_X86_SMAP), --uaccess) \ $(if $(CONFIG_FTRACE_MCOUNT_USE_OBJTOOL), --mcount) \ $(if $(CONFIG_SLS), --sls)scripts/link-vmlinux.sh +0
-0
scripts/link-vmlinux.sh objtool_link() · line 145
Quoted 9 lines
scripts/link-vmlinux.sh +0
-3@@ -140,9 +140,6 @@ objtool_link() if is_enabled CONFIG_RETPOLINE; then objtoolopt="${objtoolopt} --retpoline" fi - if is_enabled CONFIG_X86_SMAP; then - objtoolopt="${objtoolopt} --uaccess" - fi if is_enabled CONFIG_SLS; then objtoolopt="${objtoolopt} --sls" fi
I was expecting to see --uaccess unconditionally added to objtool arguments. I think that needs to happen?
Signature · 1 line
Kees Cook
Borislav Petkov
maintainerRe: [PATCH v2 3/6] x86/cpu: Remove CONFIG_X86_SMAP and "nosmap"
Quoted 46 lines · Kees Cook
Quoted 45 lines · Borislav Petkov
From: Borislav Petkov <[email protected]>
Those were added as part of the SMAP enablement but SMAP is currently an integral part of kernel proper and there's no need to disable it anymore.
Rip out that functionality.
If still needed - clearcpuid=smap.
Signed-off-by Borislav Petkov <[email protected]>Reviewed-by Lai Jiangshan <[email protected]>--- .../admin-guide/kernel-parameters.txt | 2 +- Documentation/x86/cpuinfo.rst | 5 ++-- arch/x86/Kconfig | 11 --------- arch/x86/include/asm/disabled-features.h | 8 +------ arch/x86/include/asm/smap.h | 24 ------------------- arch/x86/kernel/cpu/common.c | 15 +----------- scripts/Makefile.build | 1 - scripts/link-vmlinux.sh | 3 --- .../arch/x86/include/asm/disabled-features.h | 8 +------ 9 files changed, 6 insertions(+), 71 deletions(-)[...]
scripts/Makefile.build +0
-1@@ -233,7 +233,6 @@ objtool_args = \ $(if $(CONFIG_FRAME_POINTER),, --no-fp) \ $(if $(CONFIG_GCOV_KERNEL)$(CONFIG_LTO_CLANG), --no-unreachable)\ $(if $(CONFIG_RETPOLINE), --retpoline) \ - $(if $(CONFIG_X86_SMAP), --uaccess) \ $(if $(CONFIG_FTRACE_MCOUNT_USE_OBJTOOL), --mcount) \ $(if $(CONFIG_SLS), --sls)scripts/link-vmlinux.sh +0
-1@@ -140,9 +140,6 @@ objtool_link() if is_enabled CONFIG_RETPOLINE; then objtoolopt="${objtoolopt} --retpoline" fi - if is_enabled CONFIG_X86_SMAP; then
Quoted 8 lines
scripts/link-vmlinux.sh objtool_link() · line 145
Quoted 5 lines
scripts/link-vmlinux.sh +0
-2- objtoolopt="${objtoolopt} --uaccess" - fi if is_enabled CONFIG_SLS; then objtoolopt="${objtoolopt} --sls" fiI was expecting to see --uaccess unconditionally added to objtool arguments. I think that needs to happen?
Hmm, I think you're right.
Mr. Z?
Signature · 4 lines
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquetteBorislav Petkov
maintainerRe: [PATCH v2 3/6] x86/cpu: Remove CONFIG_X86_SMAP and "nosmap"
Quoted 3 lines · Borislav Petkov
Hmm, I think you're right.
Mr. Z?
He says yes.
Ok, that was a good catch, thanks!
This chunk looks now like this - I'll send a new version later.
---
scripts/Makefile.build +1 -1
@@ -233,7 +233,7 @@ objtool_args = \ $(if $(CONFIG_FRAME_POINTER),, --no-fp) \ $(if $(CONFIG_GCOV_KERNEL)$(CONFIG_LTO_CLANG), --no-unreachable)\ $(if $(CONFIG_RETPOLINE), --retpoline) \ - $(if $(CONFIG_X86_SMAP), --uaccess) \ + --uaccess \ $(if $(CONFIG_FTRACE_MCOUNT_USE_OBJTOOL), --mcount) \ $(if $(CONFIG_SLS), --sls)
scripts/link-vmlinux.sh +1 -4
@@ -106,7 +106,7 @@ modpost_link() objtool_link() { local objtoolcmd; - local objtoolopt; + local objtoolopt="--uaccess" if is_enabled CONFIG_LTO_CLANG && is_enabled CONFIG_STACK_VALIDATION; then # Don't perform vmlinux validation unless explicitly requested,
@@ -140,9 +140,6 @@ objtool_link() if is_enabled CONFIG_RETPOLINE; then objtoolopt="${objtoolopt} --retpoline" fi - if is_enabled CONFIG_X86_SMAP; then - objtoolopt="${objtoolopt} --uaccess" - fi if is_enabled CONFIG_SLS; then objtoolopt="${objtoolopt} --sls" fi
Signature · 4 lines
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquetteKees Cook
Reviewed-byRe: [PATCH v2 3/6] x86/cpu: Remove CONFIG_X86_SMAP and "nosmap"
Quoted 39 lines · Borislav Petkov
Quoted 3 lines · Borislav Petkov
Hmm, I think you're right.
Mr. Z?
He says yes.
Ok, that was a good catch, thanks!
This chunk looks now like this - I'll send a new version later.
---
scripts/Makefile.build +1
-1@@ -233,7 +233,7 @@ objtool_args = \ $(if $(CONFIG_FRAME_POINTER),, --no-fp) \ $(if $(CONFIG_GCOV_KERNEL)$(CONFIG_LTO_CLANG), --no-unreachable)\ $(if $(CONFIG_RETPOLINE), --retpoline) \ - $(if $(CONFIG_X86_SMAP), --uaccess) \ + --uaccess \ $(if $(CONFIG_FTRACE_MCOUNT_USE_OBJTOOL), --mcount) \ $(if $(CONFIG_SLS), --sls)scripts/link-vmlinux.sh +1
-1@@ -106,7 +106,7 @@ modpost_link() objtool_link() { local objtoolcmd; - local objtoolopt; + local objtoolopt="--uaccess" if is_enabled CONFIG_LTO_CLANG && is_enabled CONFIG_STACK_VALIDATION; then # Don't perform vmlinux validation unless explicitly requested,
scripts/link-vmlinux.sh objtool_link() · line 145
Quoted 9 lines
scripts/link-vmlinux.sh +0
-3@@ -140,9 +140,6 @@ objtool_link() if is_enabled CONFIG_RETPOLINE; then objtoolopt="${objtoolopt} --retpoline" fi - if is_enabled CONFIG_X86_SMAP; then - objtoolopt="${objtoolopt} --uaccess" - fi if is_enabled CONFIG_SLS; then objtoolopt="${objtoolopt} --sls" fi
Cool; yeah, that's kind of what I was expecting. I have a knee-jerk "I don't want to touch this again later" reaction to seeing it always applied instead of gated by CONFIG_X86 or something, but then, I doubt that'll be hard to change when/if objtool becomes multi-architecture.
Consider the update as:
Signature · 1 line
Kees Cook
Borislav Petkov
maintainerRe: [PATCH v2 3/6] x86/cpu: Remove CONFIG_X86_SMAP and "nosmap"
Quoted 4 lines · Kees Cook
Cool; yeah, that's kind of what I was expecting. I have a knee-jerk "I don't want to touch this again later" reaction to seeing it always applied instead of gated by
CONFIG_X86or something, but then, I doubt that'll be hard to change when/if objtool becomes multi-architecture.
Yah, what I had doesn't build with 32-bit allnoconfig etc because of CONFIG_STACK_VALIDATION=n or so. So I ended up doing, see end.
In this case, it gets set only when the earlier LTO_CLANG/STACK/VMLINUX validation options are enabled.
We could do an is_enabled check around it but what would be the correct criterium to enable uaccess validation?
Quoted 3 lines
Consider the update as:
Reviewed-by Kees Cook <[email protected]>
Thx.
---
scripts/Makefile.build +1 -1
@@ -233,7 +233,7 @@ objtool_args = \ $(if $(CONFIG_FRAME_POINTER),, --no-fp) \ $(if $(CONFIG_GCOV_KERNEL)$(CONFIG_LTO_CLANG), --no-unreachable)\ $(if $(CONFIG_RETPOLINE), --retpoline) \ - $(if $(CONFIG_X86_SMAP), --uaccess) \ + --uaccess \ $(if $(CONFIG_FTRACE_MCOUNT_USE_OBJTOOL), --mcount) \ $(if $(CONFIG_SLS), --sls)
scripts/link-vmlinux.sh +3 -3
@@ -140,9 +140,9 @@ objtool_link() if is_enabled CONFIG_RETPOLINE; then objtoolopt="${objtoolopt} --retpoline" fi - if is_enabled CONFIG_X86_SMAP; then - objtoolopt="${objtoolopt} --uaccess" - fi + + objtoolopt="${objtoolopt} --uaccess" + if is_enabled CONFIG_SLS; then objtoolopt="${objtoolopt} --sls" fi
Signature · 4 lines
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette[tip: x86/cpu] x86/cpu: Remove CONFIG_X86_SMAP and "nosmap"
In reply to Borislav PetkovThe following commit has been merged into the x86/cpu branch of tip:
The following commit has been merged into the x86/cpu branch of tip:
Commit-ID: dbae0a934f09208075ec3e73491bd0844e1397b3 Gitweb: https://git.kernel.org/tip/dbae0a934f09208075ec3e73491bd0844e1397b3 Author: Borislav Petkov <[email protected]> AuthorDate: Thu, 27 Jan 2022 12:56:23 +01:00 Committer: Borislav Petkov <[email protected]> CommitterDate: Mon, 04 Apr 2022 10:16:57 +02:00
x86/cpu: Remove CONFIG_X86_SMAP and "nosmap"
Those were added as part of the SMAP enablement but SMAP is currently an integral part of kernel proper and there's no need to disable it anymore.
Rip out that functionality. Leave --uaccess default on for objtool as this is what objtool should do by default anyway.
If still needed - clearcpuid=smap.
--- Documentation/admin-guide/kernel-parameters.txt | 2 +- Documentation/x86/cpuinfo.rst | 5 +-- arch/x86/Kconfig | 11 +------- arch/x86/include/asm/disabled-features.h | 8 +----- arch/x86/include/asm/smap.h | 24 +---------------- arch/x86/kernel/cpu/common.c | 15 +---------- scripts/Makefile.build | 2 +- scripts/link-vmlinux.sh | 6 ++-- tools/arch/x86/include/asm/disabled-features.h | 8 +----- 9 files changed, 10 insertions(+), 71 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +1 -1
@@ -3461,7 +3461,7 @@ noexec=on: enable non-executable mappings (default) noexec=off: disable non-executable mappings - nosmap [X86,PPC] + nosmap [PPC] Disable SMAP (Supervisor Mode Access Prevention) even if it is supported by processor.
Documentation/x86/cpuinfo.rst +2 -3
@@ -140,9 +140,8 @@ from #define X86_FEATURE_UMIP (16*32 + 2). In addition, there exists a variety of custom command-line parameters that disable specific features. The list of parameters includes, but is not limited -to, nofsgsbase, nosmap, and nosmep. 5-level paging can also be disabled using -"no5lvl". SMAP and SMEP are disabled with the aforementioned parameters, -respectively. +to, nofsgsbase, and nosmep. 5-level paging can also be disabled using +"no5lvl". SMEP is disabled with the aforementioned parameter. e: The feature was known to be non-functional. ----------------------------------------------
arch/x86/Kconfig +0 -11
@@ -1816,17 +1816,6 @@ config ARCH_RANDOM If supported, this is a high bandwidth, cryptographically secure hardware random number generator. -config X86_SMAP - def_bool y - prompt "Supervisor Mode Access Prevention" if EXPERT - help - Supervisor Mode Access Prevention (SMAP) is a security - feature in newer Intel processors. There is a small - performance cost if this enabled and turned on; there is - also a small increase in the kernel size if this is enabled. - - If unsure, say Y. - config X86_UMIP def_bool y prompt "User Mode Instruction Prevention" if EXPERT
arch/x86/include/asm/disabled-features.h +1 -7
@@ -10,12 +10,6 @@* cpu_feature_enabled(). */ -#ifdef CONFIG_X86_SMAP -# define DISABLE_SMAP 0 -#else -# define DISABLE_SMAP (1<<(X86_FEATURE_SMAP & 31)) -#endif -#ifdef CONFIG_X86_UMIP # define DISABLE_UMIP 0#else
@@ -80,7 +74,7 @@#define DISABLED_MASK6 0#define DISABLED_MASK7 (DISABLE_PTI)#define DISABLED_MASK8 0 -#define DISABLED_MASK9 (DISABLE_SMAP|DISABLE_SGX) +#define DISABLED_MASK9 (DISABLE_SGX)#define DISABLED_MASK10 0#define DISABLED_MASK11 0#define DISABLED_MASK12 0
arch/x86/include/asm/smap.h +0 -24
@@ -19,25 +19,14 @@#ifdef __ASSEMBLY__ -#ifdef CONFIG_X86_SMAP -#define ASM_CLAC \ ALTERNATIVE "", __ASM_CLAC, X86_FEATURE_SMAP#define ASM_STAC \ ALTERNATIVE "", __ASM_STAC, X86_FEATURE_SMAP -#else /* CONFIG_X86_SMAP */ - -#define ASM_CLAC -#define ASM_STAC - -#endif /* CONFIG_X86_SMAP */ -#else /* __ASSEMBLY__ */ -#ifdef CONFIG_X86_SMAP - static __always_inline void clac(void) {/* Note: a barrier is implicit in alternative() */
@@ -76,19 +65,6 @@ static __always_inline void smap_restore(unsigned long flags)#define ASM_STAC \ ALTERNATIVE("", __ASM_STAC, X86_FEATURE_SMAP) -#else /* CONFIG_X86_SMAP */ - -static inline void clac (void ) { } -static inline void stac (void ) { } - -static inline unsigned long smap_save (void ) {return 0 ; } -static inline void smap_restore (unsigned long flags) { } - -#define ASM_CLAC -#define ASM_STAC - -#endif /* CONFIG_X86_SMAP */ -#endif /* __ASSEMBLY__ */ #endif /* _ASM_X86_SMAP_H */
arch/x86/kernel/cpu/common.c +1 -14
@@ -382,13 +382,6 @@ static __always_inline void setup_smep(struct cpuinfo_x86 *c)cr4_set_bits (X86_CR4_SMEP ); } -static __init int setup_disable_smap (char *arg) -{ -setup_clear_cpu_cap (X86_FEATURE_SMAP ); -return 1 ; -} -__setup ("nosmap" , setup_disable_smap); - static __always_inline void setup_smap(struct cpuinfo_x86 *c) {unsigned long eflags =native_save_fl ();
@@ -396,14 +389,8 @@ static __always_inline void setup_smap(struct cpuinfo_x86 *c)/* This should have been cleared long ago */ BUG_ON (eflags &X86_EFLAGS_AC ); -if (cpu_has (c,X86_FEATURE_SMAP )) { -#ifdef CONFIG_X86_SMAP +if (cpu_has (c,X86_FEATURE_SMAP ))cr4_set_bits (X86_CR4_SMAP ); -#else -clear_cpu_cap (c,X86_FEATURE_SMAP ); -cr4_clear_bits (X86_CR4_SMAP ); -#endif - } } static __always_inline void setup_umip(struct cpuinfo_x86 *c)
scripts/Makefile.build +1 -1
@@ -233,7 +233,7 @@ objtool_args = \ $(if $(CONFIG_FRAME_POINTER),, --no-fp) \ $(if $(CONFIG_GCOV_KERNEL)$(CONFIG_LTO_CLANG), --no-unreachable)\ $(if $(CONFIG_RETPOLINE), --retpoline) \ - $(if $(CONFIG_X86_SMAP), --uaccess) \ + --uaccess \ $(if $(CONFIG_FTRACE_MCOUNT_USE_OBJTOOL), --mcount) \ $(if $(CONFIG_SLS), --sls)
scripts/link-vmlinux.sh +3 -3
@@ -146,9 +146,9 @@ objtool_link() if is_enabled CONFIG_RETPOLINE; then objtoolopt="${objtoolopt} --retpoline" fi - if is_enabled CONFIG_X86_SMAP; then - objtoolopt="${objtoolopt} --uaccess" - fi + + objtoolopt="${objtoolopt} --uaccess" + if is_enabled CONFIG_SLS; then objtoolopt="${objtoolopt} --sls" fi
tools/arch/x86/include/asm/disabled-features.h +1 -7
@@ -10,12 +10,6 @@* cpu_feature_enabled(). */ -#ifdef CONFIG_X86_SMAP -# define DISABLE_SMAP 0 -#else -# define DISABLE_SMAP (1<<(X86_FEATURE_SMAP & 31)) -#endif -#ifdef CONFIG_X86_UMIP # define DISABLE_UMIP 0#else
@@ -80,7 +74,7 @@#define DISABLED_MASK6 0#define DISABLED_MASK7 (DISABLE_PTI)#define DISABLED_MASK8 0 -#define DISABLED_MASK9 (DISABLE_SMAP|DISABLE_SGX) +#define DISABLED_MASK9 (DISABLE_SGX)#define DISABLED_MASK10 0#define DISABLED_MASK11 0#define DISABLED_MASK12 0
Borislav Petkov
maintainerPATCH v2 4/6 x86/cpu: Remove "nosmep"
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
There should be no need to disable SMEP anymore.
--- Documentation/admin-guide/kernel-parameters.txt | 2 +- Documentation/x86/cpuinfo.rst | 4 ++-- arch/x86/kernel/cpu/common.c | 7 ------- 3 files changed, 3 insertions(+), 10 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +1 -1
@@ -3398,7 +3398,7 @@ Disable SMAP (Supervisor Mode Access Prevention) even if it is supported by processor. - nosmep [X86,PPC64s] + nosmep [PPC64s] Disable SMEP (Supervisor Mode Execution Prevention) even if it is supported by processor.
Documentation/x86/cpuinfo.rst +2 -2
@@ -140,8 +140,8 @@ from #define X86_FEATURE_UMIP (16*32 + 2). In addition, there exists a variety of custom command-line parameters that disable specific features. The list of parameters includes, but is not limited -to, nofsgsbase, and nosmep. 5-level paging can also be disabled using -"no5lvl". SMEP is disabled with the aforementioned parameter. +to, nofsgsbase, nosgx, noxsave, etc. 5-level paging can also be disabled using +"no5lvl". e: The feature was known to be non-functional. ----------------------------------------------
arch/x86/kernel/cpu/common.c +0 -7
@@ -291,13 +291,6 @@ static inline void squash_the_stupid_serial_number(struct cpuinfo_x86 *c) }#endif -static __init int setup_disable_smep (char *arg) -{ -setup_clear_cpu_cap (X86_FEATURE_SMEP ); -return 1 ; -} -__setup ("nosmep" , setup_disable_smep); - static __always_inline void setup_smep(struct cpuinfo_x86 *c) {if (cpu_has (c,X86_FEATURE_SMEP ))
Signature · 1 line
2.29.2
Re: [PATCH v2 4/6] x86/cpu: Remove "nosmep"
Quoted 5 lines · Borislav Petkov
From: Borislav Petkov <[email protected]>
There should be no need to disable SMEP anymore.
Signed-off-by Borislav Petkov <[email protected]>
Signature · 1 line
Kees Cook
[tip: x86/cpu] x86/cpu: Remove "nosmep"
In reply to Borislav PetkovThe following commit has been merged into the x86/cpu branch of tip:
The following commit has been merged into the x86/cpu branch of tip:
Commit-ID: 385d2ae0a1b5efacb30e13a0f0e521490441d9bb Gitweb: https://git.kernel.org/tip/385d2ae0a1b5efacb30e13a0f0e521490441d9bb Author: Borislav Petkov <[email protected]> AuthorDate: Thu, 27 Jan 2022 12:56:24 +01:00 Committer: Borislav Petkov <[email protected]> CommitterDate: Mon, 04 Apr 2022 10:17:00 +02:00
x86/cpu: Remove "nosmep"
There should be no need to disable SMEP anymore.
--- Documentation/admin-guide/kernel-parameters.txt | 2 +- Documentation/x86/cpuinfo.rst | 4 ++-- arch/x86/kernel/cpu/common.c | 7 ------- 3 files changed, 3 insertions(+), 10 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +1 -1
@@ -3465,7 +3465,7 @@ Disable SMAP (Supervisor Mode Access Prevention) even if it is supported by processor. - nosmep [X86,PPC64s] + nosmep [PPC64s] Disable SMEP (Supervisor Mode Execution Prevention) even if it is supported by processor.
Documentation/x86/cpuinfo.rst +2 -2
@@ -140,8 +140,8 @@ from #define X86_FEATURE_UMIP (16*32 + 2). In addition, there exists a variety of custom command-line parameters that disable specific features. The list of parameters includes, but is not limited -to, nofsgsbase, and nosmep. 5-level paging can also be disabled using -"no5lvl". SMEP is disabled with the aforementioned parameter. +to, nofsgsbase, nosgx, noxsave, etc. 5-level paging can also be disabled using +"no5lvl". e: The feature was known to be non-functional. ----------------------------------------------
arch/x86/kernel/cpu/common.c +0 -7
@@ -369,13 +369,6 @@ static inline void squash_the_stupid_serial_number(struct cpuinfo_x86 *c) }#endif -static __init int setup_disable_smep (char *arg) -{ -setup_clear_cpu_cap (X86_FEATURE_SMEP ); -return 1 ; -} -__setup ("nosmep" , setup_disable_smep); - static __always_inline void setup_smep(struct cpuinfo_x86 *c) {if (cpu_has (c,X86_FEATURE_SMEP ))
Borislav Petkov
maintainerPATCH v2 5/6 x86/cpu: Remove "noexec"
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
It doesn't make any sense to disable non-executable mappings - security-wise or else.
So rip out that switch and move the remaining code into setup.c and delete setup_nx.c
--- .../admin-guide/kernel-parameters.txt | 5 -- Documentation/x86/x86_64/boot-options.rst | 9 --- arch/x86/include/asm/proto.h | 1 - arch/x86/kernel/setup.c | 28 ++++++++- arch/x86/mm/Makefile | 3 +- arch/x86/mm/init_64.c | 1 - arch/x86/mm/setup_nx.c | 62 ------------------- 7 files changed, 26 insertions(+), 83 deletions(-) delete mode 100644 arch/x86/mm/setup_nx.c
Documentation/admin-guide/kernel-parameters.txt +0 -5
@@ -3389,11 +3389,6 @@ noexec [IA-64] - noexec [X86] - On X86-32 available only on PAE configured kernels. - noexec=on: enable non-executable mappings (default) - noexec=off: disable non-executable mappings - nosmap [PPC] Disable SMAP (Supervisor Mode Access Prevention) even if it is supported by processor.
Documentation/x86/x86_64/boot-options.rst +0 -9
@@ -164,15 +164,6 @@ Rebooting newer BIOS, or newer board) using this option will ignore the built-in quirk table, and use the generic default reboot actions. -Non Executable Mappings -======================= - - noexec=on|off - on - Enable(default) - off - Disable - NUMA ====
arch/x86/include/asm/proto.h +0 -1
@@ -35,7 +35,6 @@ void xen_entry_INT80_compat(void);#endif void x86_configure_nx (void ); -void x86_report_nx (void );extern int reboot_force;
arch/x86/kernel/setup.c +25 -3
@@ -740,6 +740,30 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p)return 0 ; } +void x86_configure_nx (void ) +{ +if (boot_cpu_has (X86_FEATURE_NX )) + __supported_pte_mask |= _PAGE_NX; +else + __supported_pte_mask &= ~_PAGE_NX; +} + +static void __init x86_report_nx (void ) +{ +if (!boot_cpu_has (X86_FEATURE_NX )) { +printk (KERN_NOTICE "Notice: NX (Execute Disable) protection " +"missing in CPU!\n" ); + }else { +#if defined(CONFIG_X86_64 ) || defined(CONFIG_X86_PAE ) +printk (KERN_INFO "NX (Execute Disable) protection: active\n" ); +#else +/* 32bit non-PAE kernel, NX cannot be used */ +printk (KERN_NOTICE "Notice: NX (Execute Disable) protection " +"cannot be enabled: non-PAE kernel!\n" ); +#endif + } +} +/* * Determine if we were loaded by an EFI loader. If so, then we have also been * passed the efi memmap, systab, etc., so we should use these data structures
@@ -880,9 +904,7 @@ void __init setup_arch(char **cmdline_p)/* * x86_configure_nx() is called before parse_early_param() to detect * whether hardware doesn't support NX (so that the early EHCI debug -* console setup can safely call set_fixmap()). It may then be called -* again from within noexec_setup() during parsing early parameters -* to honor the respective command line option. +* console setup can safely call set_fixmap()). */ x86_configure_nx ();
arch/x86/mm/Makefile +1 -2
@@ -20,13 +20,12 @@ CFLAGS_REMOVE_mem_encrypt_identity.o = -pg endif obj-y := init.o init_$(BITS).o fault.o ioremap.o extable.o mmap.o \ - pgtable.o physaddr.o setup_nx.o tlb.o cpu_entry_area.o maccess.o + pgtable.o physaddr.o tlb.o cpu_entry_area.o maccess.o obj-y += pat/ # Make sure __phys_addr has no stackprotector CFLAGS_physaddr.o := -fno-stack-protector -CFLAGS_setup_nx.o := -fno-stack-protector CFLAGS_mem_encrypt_identity.o := -fno-stack-protector CFLAGS_fault.o := -I $(srctree)/$(src)/../include/asm/trace
arch/x86/mm/init_64.c +0 -1
@@ -110,7 +110,6 @@ int force_personality32;/* * noexec32=on|off * Control non executable heap for 32bit processes. -* To control the stack too use noexec=off * * on PROT_READ does not imply PROT_EXEC for 32-bit processes (default) * off PROT_READ implies PROT_EXEC
arch/x86/mm/setup_nx.c deleted +0 -62
62 lines not on this page · raw message
Signature · 1 line
2.29.2
Re: [PATCH v2 5/6] x86/cpu: Remove "noexec"
Quoted 10 lines · Borislav Petkov
From: Borislav Petkov <[email protected]>
It doesn't make any sense to disable non-executable mappings - security-wise or else.
So rip out that switch and move the remaining code into
setup.cand deletesetup_nx.cSigned-off-by Borislav Petkov <[email protected]>Reviewed-by Lai Jiangshan <[email protected]>
I've dug through the nx fun again. The verify_cpu() calls are untouched (which were the most fiddly part of all that a decade ago), so that's good. The early handling of __supported_pte_mask in x86_configure_nx() is untouched and the call is retained before the param processing for the set_fixmap() dance, so everything I can remember as being "difficult" with NX ordering here looks good.
Thanks for cleaning all this up!
Signature · 1 line
Kees Cook
Re: [PATCH v2 5/6] x86/cpu: Remove "noexec"
Quoted 5 lines · Kees Cook
Quoted 4 lines · Borislav Petkov
From: Borislav Petkov <[email protected]>
It doesn't make any sense to disable non-executable mappings - security-wise or else.
Heh, I've actually used noexec relatively recently to triage KVM goofs. That said, the same net result can still be achieved via clearcpuid=52, so I've no objection to removing the dedicated parameter.
Borislav Petkov
maintainerRe: [PATCH v2 5/6] x86/cpu: Remove "noexec"
Quoted 3 lines · Sean Christopherson
Heh, I've actually used noexec relatively recently to triage KVM goofs. That said, the same net result can still be achieved via clearcpuid=52, so I've no objection to removing the dedicated parameter.
Even better: clearcpuid=nx works now too.
Signature · 4 lines
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette[tip: x86/cpu] x86/cpu: Remove "noexec"
In reply to Borislav PetkovThe following commit has been merged into the x86/cpu branch of tip:
The following commit has been merged into the x86/cpu branch of tip:
Commit-ID: 76ea0025a214cdf0d2c204f4c21cbffa9fb57c32 Gitweb: https://git.kernel.org/tip/76ea0025a214cdf0d2c204f4c21cbffa9fb57c32 Author: Borislav Petkov <[email protected]> AuthorDate: Thu, 27 Jan 2022 12:56:25 +01:00 Committer: Borislav Petkov <[email protected]> CommitterDate: Mon, 04 Apr 2022 10:17:03 +02:00
x86/cpu: Remove "noexec"
It doesn't make any sense to disable non-executable mappings - security-wise or else.
So rip out that switch and move the remaining code into setup.c and delete setup_nx.c
--- Documentation/admin-guide/kernel-parameters.txt | 5 +- Documentation/x86/x86_64/boot-options.rst | 9 +-- arch/x86/include/asm/proto.h | 1 +- arch/x86/kernel/setup.c | 28 ++++++- arch/x86/mm/Makefile | 3 +- arch/x86/mm/init_64.c | 1 +- arch/x86/mm/setup_nx.c | 62 +---------------- 7 files changed, 26 insertions(+), 83 deletions(-) delete mode 100644 arch/x86/mm/setup_nx.c
Documentation/admin-guide/kernel-parameters.txt +0 -5
@@ -3456,11 +3456,6 @@ noexec [IA-64] - noexec [X86] - On X86-32 available only on PAE configured kernels. - noexec=on: enable non-executable mappings (default) - noexec=off: disable non-executable mappings - nosmap [PPC] Disable SMAP (Supervisor Mode Access Prevention) even if it is supported by processor.
Documentation/x86/x86_64/boot-options.rst +0 -9
@@ -157,15 +157,6 @@ Rebooting newer BIOS, or newer board) using this option will ignore the built-in quirk table, and use the generic default reboot actions. -Non Executable Mappings -======================= - - noexec=on|off - on - Enable(default) - off - Disable - NUMA ====
arch/x86/include/asm/proto.h +0 -1
@@ -35,7 +35,6 @@ void xen_entry_INT80_compat(void);#endif void x86_configure_nx (void ); -void x86_report_nx (void );extern int reboot_force;
arch/x86/kernel/setup.c +25 -3
@@ -756,6 +756,30 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p)return 0 ; } +void x86_configure_nx (void ) +{ +if (boot_cpu_has (X86_FEATURE_NX )) + __supported_pte_mask |= _PAGE_NX; +else + __supported_pte_mask &= ~_PAGE_NX; +} + +static void __init x86_report_nx (void ) +{ +if (!boot_cpu_has (X86_FEATURE_NX )) { +printk (KERN_NOTICE "Notice: NX (Execute Disable) protection " +"missing in CPU!\n" ); + }else { +#if defined(CONFIG_X86_64 ) || defined(CONFIG_X86_PAE ) +printk (KERN_INFO "NX (Execute Disable) protection: active\n" ); +#else +/* 32bit non-PAE kernel, NX cannot be used */ +printk (KERN_NOTICE "Notice: NX (Execute Disable) protection " +"cannot be enabled: non-PAE kernel!\n" ); +#endif + } +} +/* * Determine if we were loaded by an EFI loader. If so, then we have also been * passed the efi memmap, systab, etc., so we should use these data structures
@@ -896,9 +920,7 @@ void __init setup_arch(char **cmdline_p)/* * x86_configure_nx() is called before parse_early_param() to detect * whether hardware doesn't support NX (so that the early EHCI debug -* console setup can safely call set_fixmap()). It may then be called -* again from within noexec_setup() during parsing early parameters -* to honor the respective command line option. +* console setup can safely call set_fixmap()). */ x86_configure_nx ();
arch/x86/mm/Makefile +1 -2
@@ -20,13 +20,12 @@ CFLAGS_REMOVE_mem_encrypt_identity.o = -pg endif obj-y := init.o init_$(BITS).o fault.o ioremap.o extable.o mmap.o \ - pgtable.o physaddr.o setup_nx.o tlb.o cpu_entry_area.o maccess.o + pgtable.o physaddr.o tlb.o cpu_entry_area.o maccess.o obj-y += pat/ # Make sure __phys_addr has no stackprotector CFLAGS_physaddr.o := -fno-stack-protector -CFLAGS_setup_nx.o := -fno-stack-protector CFLAGS_mem_encrypt_identity.o := -fno-stack-protector CFLAGS_fault.o := -I $(srctree)/$(src)/../include/asm/trace
arch/x86/mm/init_64.c +0 -1
@@ -110,7 +110,6 @@ int force_personality32;/* * noexec32=on|off * Control non executable heap for 32bit processes. -* To control the stack too use noexec=off * * on PROT_READ does not imply PROT_EXEC for 32-bit processes (default) * off PROT_READ implies PROT_EXEC
arch/x86/mm/setup_nx.c deleted +0 -62
62 lines not on this page · raw message
Borislav Petkov
maintainerPATCH v2 6/6 x86/cpu: Remove "noclflush"
In reply to Borislav PetkovFrom: Borislav Petkov <[email protected]>
Not really needed anymore and there's clearcpuid=.
--- Documentation/admin-guide/kernel-parameters.txt | 2 -- arch/x86/kernel/cpu/common.c | 8 -------- 2 files changed, 10 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +0 -2
@@ -3377,8 +3377,6 @@ nocache [ARM] - noclflush [BUGS=X86] Don't use the CLFLUSH instruction - delayacct [KNL] Enable per-task delay accounting nodsp [SH] Disable hardware DSP at boot time.
arch/x86/kernel/cpu/common.c +0 -8
@@ -1727,14 +1727,6 @@ void identify_secondary_cpu(struct cpuinfo_x86 *c)update_srbds_msr (); } -static __init int setup_noclflush (char *arg) -{ -setup_clear_cpu_cap (X86_FEATURE_CLFLUSH ); -setup_clear_cpu_cap (X86_FEATURE_CLFLUSHOPT ); -return 1 ; -} -__setup ("noclflush" , setup_noclflush); -void print_cpu_info (struct cpuinfo_x86 *c) {const char *vendor =NULL ;
Signature · 1 line
2.29.2
Re: [PATCH v2 6/6] x86/cpu: Remove "noclflush"
Quoted 5 lines · Borislav Petkov
From: Borislav Petkov <[email protected]>
Not really needed anymore and there's clearcpuid=.
Signed-off-by Borislav Petkov <[email protected]>
Signature · 1 line
Kees Cook
[tip: x86/cpu] x86/cpu: Remove "noclflush"
In reply to Borislav PetkovThe following commit has been merged into the x86/cpu branch of tip:
The following commit has been merged into the x86/cpu branch of tip:
Commit-ID: f8858b5eff30d1b2be15ef1ea6285964013b95e6 Gitweb: https://git.kernel.org/tip/f8858b5eff30d1b2be15ef1ea6285964013b95e6 Author: Borislav Petkov <[email protected]> AuthorDate: Thu, 27 Jan 2022 12:56:26 +01:00 Committer: Borislav Petkov <[email protected]> CommitterDate: Mon, 04 Apr 2022 10:17:05 +02:00
x86/cpu: Remove "noclflush"
Not really needed anymore and there's clearcpuid=.
--- Documentation/admin-guide/kernel-parameters.txt | 2 -- arch/x86/kernel/cpu/common.c | 8 -------- 2 files changed, 10 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +0 -2
@@ -3444,8 +3444,6 @@ nocache [ARM] - noclflush [BUGS=X86] Don't use the CLFLUSH instruction - delayacct [KNL] Enable per-task delay accounting nodsp [SH] Disable hardware DSP at boot time.
arch/x86/kernel/cpu/common.c +0 -8
@@ -1870,14 +1870,6 @@ void identify_secondary_cpu(struct cpuinfo_x86 *c)update_srbds_msr (); } -static __init int setup_noclflush (char *arg) -{ -setup_clear_cpu_cap (X86_FEATURE_CLFLUSH ); -setup_clear_cpu_cap (X86_FEATURE_CLFLUSHOPT ); -return 1 ; -} -__setup ("noclflush" , setup_noclflush); -void print_cpu_info (struct cpuinfo_x86 *c) {const char *vendor =NULL ;