| Message ID | 20241228145746.2783627-14-yukaixiong@huawei.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | sysctl: move sysctls from vm_table into its own files | expand |
On Sat, Dec 28, 2024 at 10:17 AM Kaixiong Yu <yukaixiong@huawei.com> wrote: > > When CONFIG_X86_32 is defined and CONFIG_UML is not defined, > vdso_enabled belongs to arch/x86/entry/vdso/vdso32-setup.c. > So, move it into its own file. > > Before this patch, vdso_enabled was allowed to be set to > a value exceeding 1 on x86_32 architecture. After this patch is > applied, vdso_enabled is not permitted to set the value more than 1. > It does not matter, because according to the function load_vdso32(), > only vdso_enabled is set to 1, VDSO would be enabled. Other values > all mean "disabled". The same limitation could be seen in the > function vdso32_setup(). > > Signed-off-by: Kaixiong Yu <yukaixiong@huawei.com> > Reviewed-by: Kees Cook <kees@kernel.org> > --- > v4: > - const qualify struct ctl_table vdso_table > --- > --- > arch/x86/entry/vdso/vdso32-setup.c | 16 +++++++++++----- > kernel/sysctl.c | 8 +------- > 2 files changed, 12 insertions(+), 12 deletions(-) > > diff --git a/arch/x86/entry/vdso/vdso32-setup.c b/arch/x86/entry/vdso/vdso32-setup.c > index 76e4e74f35b5..f71625f99bf9 100644 > --- a/arch/x86/entry/vdso/vdso32-setup.c > +++ b/arch/x86/entry/vdso/vdso32-setup.c > @@ -51,15 +51,17 @@ __setup("vdso32=", vdso32_setup); > __setup_param("vdso=", vdso_setup, vdso32_setup, 0); > #endif > > -#ifdef CONFIG_X86_64 > > #ifdef CONFIG_SYSCTL > -/* Register vsyscall32 into the ABI table */ > #include <linux/sysctl.h> > > -static struct ctl_table abi_table2[] = { > +static const struct ctl_table vdso_table[] = { > { > +#ifdef CONFIG_X86_64 > .procname = "vsyscall32", > +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML)) vdso32-setup,.c is not used when building UML, so this can be reduced to "#else". > + .procname = "vdso_enabled", > +#endif > .data = &vdso32_enabled, > .maxlen = sizeof(int), > .mode = 0644, > @@ -71,10 +73,14 @@ static struct ctl_table abi_table2[] = { > > static __init int ia32_binfmt_init(void) > { > - register_sysctl("abi", abi_table2); > +#ifdef CONFIG_X86_64 > + /* Register vsyscall32 into the ABI table */ > + register_sysctl("abi", vdso_table); > +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML)) Same as above. > + register_sysctl_init("vm", vdso_table); > +#endif > return 0; > } > __initcall(ia32_binfmt_init); > #endif /* CONFIG_SYSCTL */ > > -#endif /* CONFIG_X86_64 */ Brian Gerst
On 2024/12/30 7:05, Brian Gerst wrote: > On Sat, Dec 28, 2024 at 10:17 AM Kaixiong Yu <yukaixiong@huawei.com> wrote: >> When CONFIG_X86_32 is defined and CONFIG_UML is not defined, >> vdso_enabled belongs to arch/x86/entry/vdso/vdso32-setup.c. >> So, move it into its own file. >> >> Before this patch, vdso_enabled was allowed to be set to >> a value exceeding 1 on x86_32 architecture. After this patch is >> applied, vdso_enabled is not permitted to set the value more than 1. >> It does not matter, because according to the function load_vdso32(), >> only vdso_enabled is set to 1, VDSO would be enabled. Other values >> all mean "disabled". The same limitation could be seen in the >> function vdso32_setup(). >> >> Signed-off-by: Kaixiong Yu <yukaixiong@huawei.com> >> Reviewed-by: Kees Cook <kees@kernel.org> >> --- >> v4: >> - const qualify struct ctl_table vdso_table >> --- >> --- >> arch/x86/entry/vdso/vdso32-setup.c | 16 +++++++++++----- >> kernel/sysctl.c | 8 +------- >> 2 files changed, 12 insertions(+), 12 deletions(-) >> >> diff --git a/arch/x86/entry/vdso/vdso32-setup.c b/arch/x86/entry/vdso/vdso32-setup.c >> index 76e4e74f35b5..f71625f99bf9 100644 >> --- a/arch/x86/entry/vdso/vdso32-setup.c >> +++ b/arch/x86/entry/vdso/vdso32-setup.c >> @@ -51,15 +51,17 @@ __setup("vdso32=", vdso32_setup); >> __setup_param("vdso=", vdso_setup, vdso32_setup, 0); >> #endif >> >> -#ifdef CONFIG_X86_64 >> >> #ifdef CONFIG_SYSCTL >> -/* Register vsyscall32 into the ABI table */ >> #include <linux/sysctl.h> >> >> -static struct ctl_table abi_table2[] = { >> +static const struct ctl_table vdso_table[] = { >> { >> +#ifdef CONFIG_X86_64 >> .procname = "vsyscall32", >> +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML)) > vdso32-setup,.c is not used when building UML, so this can be reduced > to "#else". I will take your advice. Thanks. >> + .procname = "vdso_enabled", >> +#endif >> .data = &vdso32_enabled, >> .maxlen = sizeof(int), >> .mode = 0644, >> @@ -71,10 +73,14 @@ static struct ctl_table abi_table2[] = { >> >> static __init int ia32_binfmt_init(void) >> { >> - register_sysctl("abi", abi_table2); >> +#ifdef CONFIG_X86_64 >> + /* Register vsyscall32 into the ABI table */ >> + register_sysctl("abi", vdso_table); >> +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML)) > Same as above. > > I will take your advice. Thanks. >> + register_sysctl_init("vm", vdso_table); >> +#endif >> return 0; >> } >> __initcall(ia32_binfmt_init); >> #endif /* CONFIG_SYSCTL */ >> >> -#endif /* CONFIG_X86_64 */ > > Brian Gerst > . >
On 2024/12/30 7:05, Brian Gerst wrote: > On Sat, Dec 28, 2024 at 10:17 AM Kaixiong Yu <yukaixiong@huawei.com> wrote: >> When CONFIG_X86_32 is defined and CONFIG_UML is not defined, >> vdso_enabled belongs to arch/x86/entry/vdso/vdso32-setup.c. >> So, move it into its own file. >> >> Before this patch, vdso_enabled was allowed to be set to >> a value exceeding 1 on x86_32 architecture. After this patch is >> applied, vdso_enabled is not permitted to set the value more than 1. >> It does not matter, because according to the function load_vdso32(), >> only vdso_enabled is set to 1, VDSO would be enabled. Other values >> all mean "disabled". The same limitation could be seen in the >> function vdso32_setup(). >> >> Signed-off-by: Kaixiong Yu <yukaixiong@huawei.com> >> Reviewed-by: Kees Cook <kees@kernel.org> >> --- >> v4: >> - const qualify struct ctl_table vdso_table >> --- >> --- >> arch/x86/entry/vdso/vdso32-setup.c | 16 +++++++++++----- >> kernel/sysctl.c | 8 +------- >> 2 files changed, 12 insertions(+), 12 deletions(-) >> >> diff --git a/arch/x86/entry/vdso/vdso32-setup.c b/arch/x86/entry/vdso/vdso32-setup.c >> index 76e4e74f35b5..f71625f99bf9 100644 >> --- a/arch/x86/entry/vdso/vdso32-setup.c >> +++ b/arch/x86/entry/vdso/vdso32-setup.c >> @@ -51,15 +51,17 @@ __setup("vdso32=", vdso32_setup); >> __setup_param("vdso=", vdso_setup, vdso32_setup, 0); >> #endif >> >> -#ifdef CONFIG_X86_64 >> >> #ifdef CONFIG_SYSCTL >> -/* Register vsyscall32 into the ABI table */ >> #include <linux/sysctl.h> >> >> -static struct ctl_table abi_table2[] = { >> +static const struct ctl_table vdso_table[] = { >> { >> +#ifdef CONFIG_X86_64 >> .procname = "vsyscall32", >> +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML)) > vdso32-setup,.c is not used when building UML, so this can be reduced > to "#else". > >> + .procname = "vdso_enabled", >> +#endif >> .data = &vdso32_enabled, >> .maxlen = sizeof(int), >> .mode = 0644, >> @@ -71,10 +73,14 @@ static struct ctl_table abi_table2[] = { >> >> static __init int ia32_binfmt_init(void) >> { >> - register_sysctl("abi", abi_table2); >> +#ifdef CONFIG_X86_64 >> + /* Register vsyscall32 into the ABI table */ >> + register_sysctl("abi", vdso_table); >> +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML)) > Same as above. > > > >> + register_sysctl_init("vm", vdso_table); >> +#endif >> return 0; >> } >> __initcall(ia32_binfmt_init); >> #endif /* CONFIG_SYSCTL */ >> >> -#endif /* CONFIG_X86_64 */ > > Brian Gerst > . Hello all； I want to confirm that I should send a new patch series, such as "PATCH v5 -next"， or just modify this patch by "git send-email -in-reply-to xxxxx"，or the maintainer will fix this issue ?
diff --git a/arch/x86/entry/vdso/vdso32-setup.c b/arch/x86/entry/vdso/vdso32-setup.c index 76e4e74f35b5..f71625f99bf9 100644 --- a/arch/x86/entry/vdso/vdso32-setup.c +++ b/arch/x86/entry/vdso/vdso32-setup.c @@ -51,15 +51,17 @@ __setup("vdso32=", vdso32_setup); __setup_param("vdso=", vdso_setup, vdso32_setup, 0); #endif -#ifdef CONFIG_X86_64 #ifdef CONFIG_SYSCTL -/* Register vsyscall32 into the ABI table */ #include <linux/sysctl.h> -static struct ctl_table abi_table2[] = { +static const struct ctl_table vdso_table[] = { { +#ifdef CONFIG_X86_64 .procname = "vsyscall32", +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML)) + .procname = "vdso_enabled", +#endif .data = &vdso32_enabled, .maxlen = sizeof(int), .mode = 0644, @@ -71,10 +73,14 @@ static struct ctl_table abi_table2[] = { static __init int ia32_binfmt_init(void) { - register_sysctl("abi", abi_table2); +#ifdef CONFIG_X86_64 + /* Register vsyscall32 into the ABI table */ + register_sysctl("abi", vdso_table); +#elif (defined(CONFIG_X86_32) && !defined(CONFIG_UML)) + register_sysctl_init("vm", vdso_table); +#endif return 0; } __initcall(ia32_binfmt_init); #endif /* CONFIG_SYSCTL */ -#endif /* CONFIG_X86_64 */ diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 860dea8f1587..7ff07b7560b4 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -2013,17 +2013,11 @@ static struct ctl_table kern_table[] = { }; static struct ctl_table vm_table[] = { -#if (defined(CONFIG_X86_32) && !defined(CONFIG_UML))|| \ - (defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL)) +#if defined(CONFIG_SUPERH) && defined(CONFIG_VSYSCALL) { .procname = "vdso_enabled", -#ifdef CONFIG_X86_32 - .data = &vdso32_enabled, - .maxlen = sizeof(vdso32_enabled), -#else .data = &vdso_enabled, .maxlen = sizeof(vdso_enabled), -#endif .mode = 0644, .proc_handler = proc_dointvec, .extra1 = SYSCTL_ZERO,