sysctl: first set of kernel/sysctl cleanups
Subsystems (7)
PATCH v2 4/9 watchdog: move watchdog sysctl interface to watchdog.c
In reply to Luis ChamberlainFrom: Xiaoming Ni <[email protected]>
The kernel/sysctl.c is a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.
To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic of proc sysctl.
So, move the watchdog syscl interface to watchdog.c. Use register_sysctl() to register the sysctl interface to avoid merge conflicts when different features modify sysctl.c at the same time.
--- kernel/sysctl.c | 96 ------------------------------------------- kernel/watchdog.c | 101 ++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 101 insertions(+), 96 deletions(-)
kernel/sysctl.c +0 -96
@@ -102,16 +102,10 @@#ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE #include <linux/stackleak.h> #endif -#ifdef CONFIG_LOCKUP_DETECTOR -#include <linux/nmi.h> -#endif #if defined(CONFIG_SYSCTL )/* Constants used for minimum and maximum */ -#ifdef CONFIG_LOCKUP_DETECTOR -static int sixty =60 ; -#endif static unsigned long zero_ul;static unsigned long one_ul =1 ;
@@ -2308,96 +2302,6 @@ static struct ctl_table kern_table[] = { .mode = 0444, .proc_handler = proc_dointvec, }, -#if defined(CONFIG_LOCKUP_DETECTOR) - { - .procname = "watchdog", - .data = &watchdog_user_enabled, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_watchdog, - .extra1 = SYSCTL_ZERO, - .extra2 = SYSCTL_ONE, - }, - { - .procname = "watchdog_thresh", - .data = &watchdog_thresh, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_watchdog_thresh, - .extra1 = SYSCTL_ZERO, - .extra2 = &sixty, - }, - { - .procname = "nmi_watchdog", - .data = &nmi_watchdog_user_enabled, - .maxlen = sizeof(int), - .mode = NMI_WATCHDOG_SYSCTL_PERM, - .proc_handler = proc_nmi_watchdog, - .extra1 = SYSCTL_ZERO, - .extra2 = SYSCTL_ONE, - }, - { - .procname = "watchdog_cpumask", - .data = &watchdog_cpumask_bits, - .maxlen = NR_CPUS, - .mode = 0644, - .proc_handler = proc_watchdog_cpumask, - }, -#ifdef CONFIG_SOFTLOCKUP_DETECTOR - { - .procname = "soft_watchdog", - .data = &soft_watchdog_user_enabled, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_soft_watchdog, - .extra1 = SYSCTL_ZERO, - .extra2 = SYSCTL_ONE, - }, - { - .procname = "softlockup_panic", - .data = &softlockup_panic, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_dointvec_minmax, - .extra1 = SYSCTL_ZERO, - .extra2 = SYSCTL_ONE, - }, -#ifdef CONFIG_SMP - { - .procname = "softlockup_all_cpu_backtrace", - .data = &sysctl_softlockup_all_cpu_backtrace, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_dointvec_minmax, - .extra1 = SYSCTL_ZERO, - .extra2 = SYSCTL_ONE, - }, -#endif /* CONFIG_SMP */ -#endif -#ifdef CONFIG_HARDLOCKUP_DETECTOR - { - .procname = "hardlockup_panic", - .data = &hardlockup_panic, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_dointvec_minmax, - .extra1 = SYSCTL_ZERO, - .extra2 = SYSCTL_ONE, - }, -#ifdef CONFIG_SMP - { - .procname = "hardlockup_all_cpu_backtrace", - .data = &sysctl_hardlockup_all_cpu_backtrace, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_dointvec_minmax, - .extra1 = SYSCTL_ZERO, - .extra2 = SYSCTL_ONE, - }, -#endif /* CONFIG_SMP */ -#endif -#endif -#if defined(CONFIG_X86_LOCAL_APIC ) && defined(CONFIG_X86 ) { .procname = "unknown_nmi_panic",
kernel/watchdog.c +101 -0
111 lines not on this page · raw message
Signature · 1 line
2.33.0
PATCH v2 8/9 aio: move aio sysctl to aio.c
In reply to Luis ChamberlainFrom: Xiaoming Ni <[email protected]>
The kernel/sysctl.c is a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.
To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic.
Move aio sysctl to aio.c and use the new register_sysctl_init() to register the sysctl interface for aio.
--- fs/aio.c | 31 +++++++++++++++++++++++++++++-- include/linux/aio.h | 4 ---- kernel/sysctl.c | 17 ----------------- 3 files changed, 29 insertions(+), 23 deletions(-)
fs/aio.c +29 -2
@@ -219,9 +219,35 @@ struct aio_kiocb {/*------ sysctl variables----*/ static DEFINE_SPINLOCK (aio_nr_lock); -unsigned long aio_nr;/* current system wide number of aio requests */ -unsigned long aio_max_nr =0x10000 ;/* system wide maximum number of aio requests */ +static unsigned long aio_nr;/* current system wide number of aio requests */ +static unsigned long aio_max_nr =0x10000 ;/* system wide maximum number of aio requests */ /*----end sysctl variables---*/ +#ifdef CONFIG_SYSCTL +static struct ctl_table aio_sysctls[] = { + { + .procname ="aio-nr" , + .data = &aio_nr, + .maxlen =sizeof (aio_nr), + .mode =0444 , + .proc_handler = proc_doulongvec_minmax, + }, + { + .procname ="aio-max-nr" , + .data = &aio_max_nr, + .maxlen =sizeof (aio_max_nr), + .mode =0644 , + .proc_handler = proc_doulongvec_minmax, + }, + {} +}; + +static void __init aio_sysctl_init (void ) +{ +register_sysctl_init ("fs" , aio_sysctls); +} +#else +#define aio_sysctl_init () do { } while (0) +#endif static struct kmem_cache *kiocb_cachep;static struct kmem_cache *kioctx_cachep;
@@ -274,6 +300,7 @@ static int __init aio_setup(void) kiocb_cachep =KMEM_CACHE (aio_kiocb,SLAB_HWCACHE_ALIGN |SLAB_PANIC ); kioctx_cachep =KMEM_CACHE (kioctx,SLAB_HWCACHE_ALIGN |SLAB_PANIC ); +aio_sysctl_init ();return 0 ; }__initcall (aio_setup);
include/linux/aio.h +0 -4
@@ -20,8 +20,4 @@ static inline void kiocb_set_cancel_fn(struct kiocb *req, kiocb_cancel_fn *cancel) { }#endif /* CONFIG_AIO */ -/* for sysctl: */ -extern unsigned long aio_nr; -extern unsigned long aio_max_nr; -#endif /* __LINUX__AIO_H */
kernel/sysctl.c +0 -17
@@ -20,7 +20,6 @@*/ #include <linux/module.h> -#include <linux/aio.h> #include <linux/mm.h> #include <linux/swap.h> #include <linux/slab.h>
@@ -3110,22 +3109,6 @@ static struct ctl_table fs_table[] = { .proc_handler = proc_dointvec, },#endif -#ifdef CONFIG_AIO - { - .procname = "aio-nr", - .data = &aio_nr, - .maxlen = sizeof(aio_nr), - .mode = 0444, - .proc_handler = proc_doulongvec_minmax, - }, - { - .procname = "aio-max-nr", - .data = &aio_max_nr, - .maxlen = sizeof(aio_max_nr), - .mode = 0644, - .proc_handler = proc_doulongvec_minmax, - }, -#endif /* CONFIG_AIO */#ifdef CONFIG_INOTIFY_USER { .procname = "inotify",
Signature · 1 line
2.33.0
PATCH v2 5/9 sysctl: make ngroups_max const
In reply to Luis ChamberlainFrom: Stephen Kitt <[email protected]>
ngroups_max is a read-only sysctl entry, reflecting NGROUPS_MAX. Make it const, in the same way as cap_last_cap.
--- kernel/sysctl.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
kernel/sysctl.c +2 -2
@@ -124,7 +124,7 @@ static unsigned long dirty_bytes_min = 2 * PAGE_SIZE;static int maxolduid =65535 ;static int minolduid; -static int ngroups_max =NGROUPS_MAX ; +static const int ngroups_max =NGROUPS_MAX ;static const int cap_last_cap =CAP_LAST_CAP ;
@@ -2290,7 +2290,7 @@ static struct ctl_table kern_table[] = {#endif { .procname = "ngroups_max", - .data = &ngroups_max, + .data = (void *)&ngroups_max, .maxlen = sizeof (int), .mode = 0444, .proc_handler = proc_dointvec,
Signature · 1 line
2.33.0
PATCH v2 9/9 dnotify: move dnotify sysctl to dnotify.c
In reply to Luis ChamberlainFrom: Xiaoming Ni <[email protected]>
The kernel/sysctl.c is a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.
To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic.
So move dnotify sysctls to dnotify.c and use the new register_sysctl_init() to register the sysctl interface.
--- fs/notify/dnotify/dnotify.c | 21 ++++++++++++++++++++- include/linux/dnotify.h | 1 - kernel/sysctl.c | 10 ---------- 3 files changed, 20 insertions(+), 12 deletions(-)
fs/notify/dnotify/dnotify.c +20 -1
@@ -19,7 +19,25 @@#include <linux/fdtable.h> #include <linux/fsnotify_backend.h> -int dir_notify_enable__read_mostly =1 ; +static int dir_notify_enable__read_mostly =1 ; +#ifdef CONFIG_SYSCTL +static struct ctl_table dnotify_sysctls[] = { + { + .procname ="dir-notify-enable" , + .data = &dir_notify_enable, + .maxlen =sizeof (int ), + .mode =0644 , + .proc_handler = proc_dointvec, + }, + {} +}; +static void __init dnotify_sysctl_init (void ) +{ +register_sysctl_init ("fs" , dnotify_sysctls); +} +#else +#define dnotify_sysctl_init () do { } while (0) +#endif static struct kmem_cache *dnotify_struct_cache__read_mostly ;static struct kmem_cache *dnotify_mark_cache__read_mostly ;
@@ -386,6 +404,7 @@ static int __init dnotify_init(void) dnotify_group =fsnotify_alloc_group (&dnotify_fsnotify_ops);if (IS_ERR (dnotify_group))panic ("unable to allocate fsnotify group for dnotify\n" ); +dnotify_sysctl_init ();return 0 ; }
include/linux/dnotify.h +0 -1
@@ -29,7 +29,6 @@ struct dnotify_struct { FS_CREATE | FS_DN_RENAME |\ FS_MOVED_FROM | FS_MOVED_TO) -extern int dir_notify_enable;extern void dnotify_flush (struct file *,fl_owner_t );extern int fcntl_dirnotify (int ,struct file *,unsigned long );
kernel/sysctl.c +0 -10
@@ -48,7 +48,6 @@#include <linux/times.h> #include <linux/limits.h> #include <linux/dcache.h> -#include <linux/dnotify.h> #include <linux/syscalls.h> #include <linux/vmstat.h> #include <linux/nfs_fs.h>
@@ -3090,15 +3089,6 @@ static struct ctl_table fs_table[] = { .proc_handler = proc_dointvec, },#endif -#ifdef CONFIG_DNOTIFY - { - .procname = "dir-notify-enable", - .data = &dir_notify_enable, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_dointvec, - }, -#endif #ifdef CONFIG_MMU #ifdef CONFIG_FILE_LOCKING {
Signature · 1 line
2.33.0
PATCH v2 2/9 sysctl: Move some boundary constants from sysctl.c to sysctl_vals
In reply to Luis ChamberlainFrom: Xiaoming Ni <[email protected]>
sysctl has helpers which let us specify boundary values for a min or max int value. Since these are used for a boundary check only they don't change, so move these variables to sysctl_vals to avoid adding duplicate variables. This will help with our cleanup of kernel/sysctl.c.
--- fs/proc/proc_sysctl.c | 2 +- include/linux/sysctl.h | 12 +++++++++--- kernel/sysctl.c | 44 ++++++++++++++++++------------------------ 3 files changed, 29 insertions(+), 29 deletions(-)
fs/proc/proc_sysctl.c +1 -1
@@ -26,7 +26,7 @@ static const struct file_operations proc_sys_dir_file_operations;static const struct inode_operations proc_sys_dir_operations;/* shared constants to be used in various sysctls */ -const int sysctl_vals[] = {0 ,1 ,INT_MAX }; +const int sysctl_vals[] = {-1 ,0 ,1 ,2 ,4 ,100 ,200 ,1000 ,INT_MAX };EXPORT_SYMBOL (sysctl_vals);/* Support for permanently empty directories */
include/linux/sysctl.h +9 -3
@@ -38,9 +38,15 @@ struct ctl_table_header;struct ctl_dir ;/* Keep the same order as in fs/proc/proc_sysctl.c */ -#define SYSCTL_ZERO ((void *)&sysctl_vals[0]) -#define SYSCTL_ONE ((void *)&sysctl_vals[1]) -#define SYSCTL_INT_MAX ((void *)&sysctl_vals[2]) +#define SYSCTL_NEG_ONE ((void *)&sysctl_vals[0]) +#define SYSCTL_ZERO ((void *)&sysctl_vals[1]) +#define SYSCTL_ONE ((void *)&sysctl_vals[2]) +#define SYSCTL_TWO ((void *)&sysctl_vals[3]) +#define SYSCTL_FOUR ((void *)&sysctl_vals[4]) +#define SYSCTL_ONE_HUNDRED ((void *)&sysctl_vals[5]) +#define SYSCTL_TWO_HUNDRED ((void *)&sysctl_vals[6]) +#define SYSCTL_ONE_THOUSAND ((void *)&sysctl_vals[7]) +#define SYSCTL_INT_MAX ((void *)&sysctl_vals[8])extern const int sysctl_vals[];
kernel/sysctl.c +19 -25
167 lines not on this page · raw message
Signature · 1 line
2.33.0
PATCH v2 3/9 hung_task: Move hung_task sysctl interface to hung_task.c
In reply to Luis ChamberlainFrom: Xiaoming Ni <[email protected]>
The kernel/sysctl.c is a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.
To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic.
So move hung_task sysctl interface to hung_task.c and use register_sysctl() to register the sysctl interface.
--- include/linux/sched/sysctl.h | 14 +------ kernel/hung_task.c | 81 ++++++++++++++++++++++++++++++++++-- kernel/sysctl.c | 61 --------------------------- 3 files changed, 79 insertions(+), 77 deletions(-)
include/linux/sched/sysctl.h +1 -13
@@ -7,20 +7,8 @@struct ctl_table ;#ifdef CONFIG_DETECT_HUNG_TASK - -#ifdef CONFIG_SMP -extern unsigned int sysctl_hung_task_all_cpu_backtrace; -#else -#define sysctl_hung_task_all_cpu_backtrace 0 -#endif /* CONFIG_SMP */ - -extern int sysctl_hung_task_check_count; -extern unsigned int sysctl_hung_task_panic; +/* used for hung_task and block/ */ extern unsigned long sysctl_hung_task_timeout_secs; -extern unsigned long sysctl_hung_task_check_interval_secs; -extern int sysctl_hung_task_warnings; -int proc_dohung_task_timeout_secs (struct ctl_table *table,int write, -void *buffer,size_t *lenp,loff_t *ppos);#else /* Avoid need for ifdefs elsewhere in the code */ enum { sysctl_hung_task_timeout_secs =0 };
kernel/hung_task.c +78 -3
108 lines not on this page · raw message
kernel/sysctl.c +0 -61
73 lines not on this page · raw message
Signature · 1 line
2.33.0
PATCH v2 1/9 sysctl: add a new register_sysctl_init() interface
In reply to Luis ChamberlainFrom: Xiaoming Ni <[email protected]>
The kernel/sysctl.c is a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.
To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic.
Today though folks heavily rely on tables on kernel/sysctl.c so they can easily just extend this table with their needed sysctls. In order to help users move their sysctls out we need to provide a helper which can be used during code initialization.
We special-case the initialization use of register_sysctl() since it *is* safe to fail, given all that sysctls do is provide a dynamic interface to query or modify at runtime an existing variable. So the use case of register_sysctl() on init should *not* stop if the sysctls don't end up getting registered. It would be counter productive to stop boot if a simple sysctl registration failed.
Provide a helper for init then, and document the recommended init levels to use for callers of this routine. We will later use this in subsequent patches to start slimming down kernel/sysctl.c tables and moving sysctl registration to the code which actually needs these sysctls.
[mcgrof: major commit log and documentation rephrasing
also moved to fs/proc/proc_sysctl.c ]
--- fs/proc/proc_sysctl.c | 33 +++++++++++++++++++++++++++++++++ include/linux/sysctl.h | 3 +++ 2 files changed, 36 insertions(+)
fs/proc/proc_sysctl.c +33 -0
@@ -16,6 +16,7 @@#include <linux/module.h> #include <linux/bpf-cgroup.h> #include <linux/mount.h> +#include <linux/kmemleak.h> #include "internal.h" static const struct dentry_operations proc_sys_dentry_operations;
@@ -1384,6 +1385,38 @@ struct ctl_table_header *register_sysctl(const char *path, struct ctl_table *tab }EXPORT_SYMBOL (register_sysctl); +/** +* __register_sysctl_init() - register sysctl table to path +* @path: path name for sysctl base +* @table: This is the sysctl table that needs to be registered to the path +* @table_name: The name of sysctl table, only used for log printing when +* registration fails +* +* The sysctl interface is used by userspace to query or modify at runtime +* a predefined value set on a variable. These variables however have default +* values pre-set. Code which depends on these variables will always work even +* if register_sysctl() fails. If register_sysctl() fails you'd just loose the +* ability to query or modify the sysctls dynamically at run time. Chances of +* register_sysctl() failing on init are extremely low, and so for both reasons +* this function does not return any error as it is used by initialization code. +* +* Context: Can only be called after your respective sysctl base path has been +* registered. So for instance, most base directories are registered early on +* init before init levels are processed through proc_sys_init() and +* sysctl_init(). +*/ +void __init __register_sysctl_init (const char *path,struct ctl_table *table, +const char *table_name) +{ +struct ctl_table_header *hdr =register_sysctl (path, table); + +if (unlikely (!hdr)) { +pr_err ("failed when register_sysctl %s to %s\n" , table_name, path); +return ; + } +kmemleak_not_leak (hdr); +} +static char *append_path (const char *path,char *pos,const char *name) {int namelen;
include/linux/sysctl.h +3 -0
@@ -199,6 +199,9 @@ struct ctl_table_header *register_sysctl_paths(const struct ctl_path *path,void unregister_sysctl_table (struct ctl_table_header * table);extern int sysctl_init (void ); +extern void __register_sysctl_init (const char *path,struct ctl_table *table, +const char *table_name); +#define register_sysctl_init (path, table) __register_sysctl_init(path, table, #table)void do_sysctl_args (void );extern int pwrsw_enabled;
Signature · 1 line
2.33.0
v2 cover sysctl: first set of kernel/sysctl cleanups
Finally had time to respin the series of the work we had started last year on cleaning up the kernel/sysct.c kitchen sink. People keeps stuffing their sysctls in that file and this creates a maintenance burden. So this effort is aimed at placing sysctls where they actually belong.
I'm going to split patches up into series as there is quite a bit of work.
This first set adds register_sysctl_init() for uses of registerting a sysctl on the init path, adds const where missing to a few places, generalizes common values so to be more easy to share, and starts the move of a few kernel/sysctl.c out where they belong.
The majority of rework on v2 in this first patch set is 0-day fixes. Eric W. Biederman's feedback is later addressed in subsequent patch sets.
I'll only post the first two patch sets for now. We can address the rest once the first two patch sets get completely reviewed / Acked. Since the sysctls are all over the place I can either put up a tree to keep track of these changes and later send a pull request to Linus or we can have them trickle into Andrew's tree. Let me know what folks prefer.
Changes in v2:
* 0-day compile issues
* added reviewed-by tags
* enhanced commit logs
* Added patch by Stephen Kitt
Stephen Kitt (1):
sysctl: make ngroups_max const
Xiaoming Ni (8):
sysctl: add a new register_sysctl_init() interface
sysctl: Move some boundary constants from sysctl.c to sysctl_vals
hung_task: Move hung_task sysctl interface to hung_task.c
watchdog: move watchdog sysctl interface to watchdog.c
sysctl: use const for typically used max/min proc sysctls
sysctl: use SYSCTL_ZERO to replace some static int zero uses
aio: move aio sysctl to aio.c
dnotify: move dnotify sysctl to dnotify.c
fs/aio.c | 31 +++- fs/notify/dnotify/dnotify.c | 21 ++- fs/proc/proc_sysctl.c | 35 ++++- include/linux/aio.h | 4 - include/linux/dnotify.h | 1 - include/linux/sched/sysctl.h | 14 +- include/linux/sysctl.h | 15 +- kernel/hung_task.c | 81 +++++++++- kernel/sysctl.c | 285 ++++++----------------------------- kernel/watchdog.c | 101 +++++++++++++ 10 files changed, 322 insertions(+), 266 deletions(-)
Signature · 1 line
2.33.0
PATCH v2 7/9 sysctl: use SYSCTL_ZERO to replace some static int zero uses
From: Xiaoming Ni <[email protected]>
Use the variable SYSCTL_ZERO to replace some static int boundary variables with a value of 0 (minolduid, min_extfrag_threshold, min_wakeup_granularity_ns).
--- kernel/sysctl.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
kernel/sysctl.c +7 -7
@@ -122,7 +122,7 @@ static const unsigned long dirty_bytes_min = 2 * PAGE_SIZE;/* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */ static const int maxolduid =65535 ; -static const int minolduid; +/* minolduid is SYSCTL_ZERO */ static const int ngroups_max =NGROUPS_MAX ;static const int cap_last_cap =CAP_LAST_CAP ;
@@ -170,7 +170,7 @@ int sysctl_legacy_va_layout;#endif #ifdef CONFIG_COMPACTION -static const int min_extfrag_threshold; +/* min_extfrag_threshold is SYSCTL_ZERO */ ;static const int max_extfrag_threshold =1000 ;#endif
@@ -2175,7 +2175,7 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = (void *)&minolduid, + .extra1 = SYSCTL_ZERO, .extra2 = (void *)&maxolduid, }, {
@@ -2184,7 +2184,7 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = (void *)&minolduid, + .extra1 = SYSCTL_ZERO, .extra2 = (void *)&maxolduid, },#ifdef CONFIG_S390
@@ -2758,7 +2758,7 @@ static struct ctl_table vm_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = (void *)&min_extfrag_threshold, + .extra1 = SYSCTL_ZERO, .extra2 = (void *)&max_extfrag_threshold, }, {
@@ -3070,7 +3070,7 @@ static struct ctl_table fs_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = (void *)&minolduid, + .extra1 = SYSCTL_ZERO, .extra2 = (void *)&maxolduid, }, {
@@ -3079,7 +3079,7 @@ static struct ctl_table fs_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = (void *)&minolduid, + .extra1 = SYSCTL_ZERO, .extra2 = (void *)&maxolduid, },#ifdef CONFIG_FILE_LOCKING
Signature · 1 line
2.33.0
PATCH v2 6/9 sysctl: use const for typically used max/min proc sysctls
In reply to Luis ChamberlainFrom: Xiaoming Ni <[email protected]>
When proc_dointvec_minmax() or proc_doulongvec_minmax() are used we are using the extra1 and extra2 parameters on the sysctl table only for a min and max boundary, these extra1 and extra2 arguments are then used for read-only operations. So make them const to reflect this.
--- kernel/sysctl.c | 53 ++++++++++++++++++++++++------------------------- 1 file changed, 26 insertions(+), 27 deletions(-)
kernel/sysctl.c +26 -27
@@ -107,27 +107,26 @@/* Constants used for minimum and maximum */ -static unsigned long zero_ul; -static unsigned long one_ul =1 ; -static unsigned long long_max =LONG_MAX ; +static const unsigned long zero_ul; +static const unsigned long one_ul =1 ; +static const unsigned long long_max =LONG_MAX ;#ifdef CONFIG_PRINTK -static int ten_thousand =10000 ; +static const int ten_thousand =10000 ;#endif #ifdef CONFIG_PERF_EVENTS -static int six_hundred_forty_kb =640 *1024 ; +static const int six_hundred_forty_kb =640 *1024 ;#endif /* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */ -static unsigned long dirty_bytes_min =2 *PAGE_SIZE ; +static const unsigned long dirty_bytes_min =2 *PAGE_SIZE ;/* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */ -static int maxolduid =65535 ; -static int minolduid; +static const int maxolduid =65535 ; +static const int minolduid;static const int ngroups_max =NGROUPS_MAX ;static const int cap_last_cap =CAP_LAST_CAP ; -#ifdef CONFIG_INOTIFY_USER #include <linux/inotify.h> #endif
@@ -171,8 +170,8 @@ int sysctl_legacy_va_layout;#endif #ifdef CONFIG_COMPACTION -static int min_extfrag_threshold; -static int max_extfrag_threshold =1000 ; +static const int min_extfrag_threshold; +static const int max_extfrag_threshold =1000 ;#endif #endif /* CONFIG_SYSCTL */
@@ -2176,8 +2175,8 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = &minolduid, - .extra2 = &maxolduid, + .extra1 = (void *)&minolduid, + .extra2 = (void *)&maxolduid, }, { .procname = "overflowgid",
@@ -2185,8 +2184,8 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = &minolduid, - .extra2 = &maxolduid, + .extra1 = (void *)&minolduid, + .extra2 = (void *)&maxolduid, },#ifdef CONFIG_S390 {
@@ -2260,7 +2259,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, - .extra2 = &ten_thousand, + .extra2 = (void *)&ten_thousand, }, { .procname = "printk_devkmsg",
@@ -2472,7 +2471,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = perf_event_max_stack_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &six_hundred_forty_kb, + .extra2 = (void *)&six_hundred_forty_kb, }, { .procname = "perf_event_max_contexts_per_stack",
@@ -2628,7 +2627,7 @@ static struct ctl_table vm_table[] = { .maxlen = sizeof(dirty_background_bytes), .mode = 0644, .proc_handler = dirty_background_bytes_handler, - .extra1 = &one_ul, + .extra1 = (void *)&one_ul, }, { .procname = "dirty_ratio",
@@ -2645,7 +2644,7 @@ static struct ctl_table vm_table[] = { .maxlen = sizeof(vm_dirty_bytes), .mode = 0644, .proc_handler = dirty_bytes_handler, - .extra1 = &dirty_bytes_min, + .extra1 = (void *)&dirty_bytes_min, }, { .procname = "dirty_writeback_centisecs",
@@ -2759,8 +2758,8 @@ static struct ctl_table vm_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = &min_extfrag_threshold, - .extra2 = &max_extfrag_threshold, + .extra1 = (void *)&min_extfrag_threshold, + .extra2 = (void *)&max_extfrag_threshold, }, { .procname = "compact_unevictable_allowed",
@@ -3046,8 +3045,8 @@ static struct ctl_table fs_table[] = { .maxlen = sizeof(files_stat.max_files), .mode = 0644, .proc_handler = proc_doulongvec_minmax, - .extra1 = &zero_ul, - .extra2 = &long_max, + .extra1 = (void *)&zero_ul, + .extra2 = (void *)&long_max, }, { .procname = "nr_open",
@@ -3071,8 +3070,8 @@ static struct ctl_table fs_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = &minolduid, - .extra2 = &maxolduid, + .extra1 = (void *)&minolduid, + .extra2 = (void *)&maxolduid, }, { .procname = "overflowgid",
@@ -3080,8 +3079,8 @@ static struct ctl_table fs_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = &minolduid, - .extra2 = &maxolduid, + .extra1 = (void *)&minolduid, + .extra2 = (void *)&maxolduid, },#ifdef CONFIG_FILE_LOCKING {
Signature · 1 line
2.33.0
Andrew Morton
maintainerQuoted 4 lines · On Tue, 23 Nov 2021 12:23:38 -0800 Luis Chamberlain <[email protected]> wrote:
Since the sysctls are all over the place I can either put up a tree to keep track of these changes and later send a pull request to Linus or we can have them trickle into Andrew's tree. Let me know what folks prefer.
I grabbed. I staged them behind all-of-linux-next, to get visibility into changes in the various trees which might conflict.
Quoted 9 lines · Andrew Morton
Quoted 4 lines · On Tue, 23 Nov 2021 12:23:38 -0800 Luis Chamberlain <[email protected]> wrote:
Since the sysctls are all over the place I can either put up a tree to keep track of these changes and later send a pull request to Linus or we can have them trickle into Andrew's tree. Let me know what folks prefer.
I grabbed. I staged them behind all-of-linux-next, to get visibility into changes in the various trees which might conflict.
Groovy thanks, if nothing barks back next week I can send out a third set of changes. I'll note that I've been testing these with 0-day prior to posting sets.
Luis
Re: [PATCH v2 2/9] sysctl: Move some boundary constants from sysctl.c to sysctl_vals
In reply to Luis ChamberlainQuoted 6 lines · Luis Chamberlain
From: Xiaoming Ni <[email protected]>
sysctl has helpers which let us specify boundary values for a min or max int value. Since these are used for a boundary check only they don't change, so move these variables to
sysctl_valsto avoid adding duplicate variables. This will help with our cleanup ofkernel/sysctl.c.
Ouch.
I kind of, sort of, have to protest.
Where the macros that use sysctl_vals don't have a type they have caused mysterious code breakage because people did not realize they can not be used with sysctls that take a long instead of an int.
This came up with when the internal storage for ucounts see kernel/ucount.c changed from an int to a long. We were quite a while tracking what was going on until we realized that the code could not use SYSCTL_ZERO and SYSCTL_INT_MAX and that we had to defined our own that were long.
So before we extend something like this can we please change the macro naming convention so that it includes the size of the type we want.
I am also not a fan of sysctl_vals living in proc_sysctl. They have nothing to do with the code in that file. They would do much better in kernel/sysctl.c close to the helpers that use them.
Eric
Quoted 237 lines
Signed-off-by Xiaoming Ni <[email protected]>Reviewed-by Kees Cook <[email protected]>[mcgrof: major rebase]Signed-off-by Luis Chamberlain <[email protected]>--- fs/proc/proc_sysctl.c | 2 +- include/linux/sysctl.h | 12 +++++++++--- kernel/sysctl.c | 44 ++++++++++++++++++------------------------ 3 files changed, 29 insertions(+), 29 deletions(-)fs/proc/proc_sysctl.c +1
-1@@ -26,7 +26,7 @@ static const struct file_operations proc_sys_dir_file_operations;static const struct inode_operations proc_sys_dir_operations;/* shared constants to be used in various sysctls */ -const int sysctl_vals[] = {0 ,1 ,INT_MAX }; +const int sysctl_vals[] = {-1 ,0 ,1 ,2 ,4 ,100 ,200 ,1000 ,INT_MAX };EXPORT_SYMBOL (sysctl_vals);/* Support for permanently empty directories */ include/linux/sysctl.h +9
-3@@ -38,9 +38,15 @@ struct ctl_table_header;struct ctl_dir ;/* Keep the same order as in fs/proc/proc_sysctl.c */ -#define SYSCTL_ZERO ((void *)&sysctl_vals[0]) -#define SYSCTL_ONE ((void *)&sysctl_vals[1]) -#define SYSCTL_INT_MAX ((void *)&sysctl_vals[2]) +#define SYSCTL_NEG_ONE ((void *)&sysctl_vals[0]) +#define SYSCTL_ZERO ((void *)&sysctl_vals[1]) +#define SYSCTL_ONE ((void *)&sysctl_vals[2]) +#define SYSCTL_TWO ((void *)&sysctl_vals[3]) +#define SYSCTL_FOUR ((void *)&sysctl_vals[4]) +#define SYSCTL_ONE_HUNDRED ((void *)&sysctl_vals[5]) +#define SYSCTL_TWO_HUNDRED ((void *)&sysctl_vals[6]) +#define SYSCTL_ONE_THOUSAND ((void *)&sysctl_vals[7]) +#define SYSCTL_INT_MAX ((void *)&sysctl_vals[8])extern const int sysctl_vals[];kernel/sysctl.c +19
-25@@ -113,15 +113,9 @@static int sixty =60 ;#endif -static int __maybe_unused neg_one =-1 ; -static int __maybe_unused two =2 ; -static int __maybe_unused four =4 ;static unsigned long zero_ul;static unsigned long one_ul =1 ;static unsigned long long_max =LONG_MAX ; -static int one_hundred =100 ; -static int two_hundred =200 ; -static int one_thousand =1000 ;#ifdef CONFIG_PRINTK static int ten_thousand =10000 ;#endif @@ -1962,7 +1956,7 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = &neg_one, + .extra1 = SYSCTL_NEG_ONE, .extra2 = SYSCTL_ONE, },#endif @@ -2304,7 +2298,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax_sysadmin, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, },#endif {@@ -2564,7 +2558,7 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = &neg_one, + .extra1 = SYSCTL_NEG_ONE, },#endif #ifdef CONFIG_RT_MUTEXES @@ -2626,7 +2620,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = perf_cpu_time_max_percent_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "perf_event_max_stack",@@ -2644,7 +2638,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = perf_event_max_stack_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_thousand, + .extra2 = SYSCTL_ONE_THOUSAND, },#endif {@@ -2675,7 +2669,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = bpf_unpriv_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "bpf_stats_enabled",@@ -2729,7 +2723,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = overcommit_policy_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "panic_on_oom",@@ -2738,7 +2732,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "oom_kill_allocating_task",@@ -2783,7 +2777,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = dirty_background_ratio_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "dirty_background_bytes",@@ -2800,7 +2794,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = dirty_ratio_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "dirty_bytes",@@ -2840,7 +2834,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, - .extra2 = &two_hundred, + .extra2 = SYSCTL_TWO_HUNDRED, },#ifdef CONFIG_HUGETLB_PAGE {@@ -2897,7 +2891,7 @@ static struct ctl_table vm_table[] = { .mode = 0200, .proc_handler = drop_caches_sysctl_handler, .extra1 = SYSCTL_ONE, - .extra2 = &four, + .extra2 = SYSCTL_FOUR, },#ifdef CONFIG_COMPACTION {@@ -2914,7 +2908,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = compaction_proactiveness_sysctl_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "extfrag_threshold",@@ -2959,7 +2953,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = watermark_scale_factor_sysctl_handler, .extra1 = SYSCTL_ONE, - .extra2 = &one_thousand, + .extra2 = SYSCTL_ONE_THOUSAND, }, { .procname = "percpu_pagelist_high_fraction",@@ -3038,7 +3032,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = sysctl_min_unmapped_ratio_sysctl_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "min_slab_ratio",@@ -3047,7 +3041,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = sysctl_min_slab_ratio_sysctl_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, },#endif #ifdef CONFIG_SMP @@ -3337,7 +3331,7 @@ static struct ctl_table fs_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "protected_regular",@@ -3346,7 +3340,7 @@ static struct ctl_table fs_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "suid_dumpable",@@ -3355,7 +3349,7 @@ static struct ctl_table fs_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax_coredump, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, },#if defined(CONFIG_BINFMT_MISC ) || defined(CONFIG_BINFMT_MISC_MODULE ) {
Re: [PATCH v2 2/9] sysctl: Move some boundary constants from sysctl.c to sysctl_vals
Quoted 17 lines · Eric W. Biederman
Quoted 6 lines · Luis Chamberlain
From: Xiaoming Ni <[email protected]>
sysctl has helpers which let us specify boundary values for a min or max int value. Since these are used for a boundary check only they don't change, so move these variables to
sysctl_valsto avoid adding duplicate variables. This will help with our cleanup ofkernel/sysctl.c.Ouch.
I kind of, sort of, have to protest.
Where the macros that use
sysctl_valsdon't have a type they have caused mysterious code breakage because people did not realize they can not be used with sysctls that take a long instead of an int.
Quoted 6 lines
This came up with when the internal storage for ucounts see
kernel/ucount.cchanged from an int to a long. We were quite a while
tracking what was going on until we realized that the code could not useSYSCTL_ZEROandSYSCTL_INT_MAXand that we had to defined our ownthatSYSCTL_ZEROandSYSCTL_ZEROinvolve dozens of files and are used in hundreds of places. were long.
static unsigned long zero_ul;
static unsigned long one_ul = 1;
static unsigned long long_max = LONG_MAX;EXPORT_SYMBOL(proc_doulongvec_minmax);
Yes, min/max of type unsigned long is used in multiple sysctl interfaces. It is necessary to add an unsigned long sysctl_val array to avoid repeated definitions in different .c files.
Quoted 4 lines
So before we extend something like this can we please change the macro naming convention so that it includes the size of the type we want.
The int type is the most widely used type. By default, numeric constants are also of the int type. SYSCTL_ZERO and SYSCTL_ZERO involve dozens of files and are used in hundreds of places. Whether only non-int macros need to be named with their type size?
Quoted 5 lines
I am also not a fan of
sysctl_valsliving inproc_sysctl. They have nothing to do with the code in that file. They would do much better inkernel/sysctl.cclose to the helpers that use them.
yes
Thanks
Xiaoming Ni
Quoted 242 lines
Eric
Quoted 229 lines
Signed-off-by Xiaoming Ni <[email protected]>Reviewed-by Kees Cook <[email protected]>[mcgrof: major rebase]Signed-off-by Luis Chamberlain <[email protected]>--- fs/proc/proc_sysctl.c | 2 +- include/linux/sysctl.h | 12 +++++++++--- kernel/sysctl.c | 44 ++++++++++++++++++------------------------ 3 files changed, 29 insertions(+), 29 deletions(-)fs/proc/proc_sysctl.c +1
-1@@ -26,7 +26,7 @@ static const struct file_operations proc_sys_dir_file_operations;static const struct inode_operations proc_sys_dir_operations;/* shared constants to be used in various sysctls */ -const int sysctl_vals[] = {0 ,1 ,INT_MAX }; +const int sysctl_vals[] = {-1 ,0 ,1 ,2 ,4 ,100 ,200 ,1000 ,INT_MAX };EXPORT_SYMBOL (sysctl_vals);/* Support for permanently empty directories */ include/linux/sysctl.h +9
-3@@ -38,9 +38,15 @@ struct ctl_table_header;struct ctl_dir ;/* Keep the same order as in fs/proc/proc_sysctl.c */ -#define SYSCTL_ZERO ((void *)&sysctl_vals[0]) -#define SYSCTL_ONE ((void *)&sysctl_vals[1]) -#define SYSCTL_INT_MAX ((void *)&sysctl_vals[2]) +#define SYSCTL_NEG_ONE ((void *)&sysctl_vals[0]) +#define SYSCTL_ZERO ((void *)&sysctl_vals[1]) +#define SYSCTL_ONE ((void *)&sysctl_vals[2]) +#define SYSCTL_TWO ((void *)&sysctl_vals[3]) +#define SYSCTL_FOUR ((void *)&sysctl_vals[4]) +#define SYSCTL_ONE_HUNDRED ((void *)&sysctl_vals[5]) +#define SYSCTL_TWO_HUNDRED ((void *)&sysctl_vals[6]) +#define SYSCTL_ONE_THOUSAND ((void *)&sysctl_vals[7]) +#define SYSCTL_INT_MAX ((void *)&sysctl_vals[8])extern const int sysctl_vals[];kernel/sysctl.c +18
-24@@ -113,15 +113,9 @@static int sixty =60 ;#endif -static int __maybe_unused neg_one =-1 ; -static int __maybe_unused two =2 ; -static int __maybe_unused four =4 ;static unsigned long zero_ul;static unsigned long one_ul =1 ;static unsigned long long_max =LONG_MAX ; -static int one_hundred =100 ; -static int two_hundred =200 ; -static int one_thousand =1000 ;#ifdef CONFIG_PRINTK static int ten_thousand =10000 ;#endif @@ -1962,7 +1956,7 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = &neg_one, + .extra1 = SYSCTL_NEG_ONE, .extra2 = SYSCTL_ONE, },#endif @@ -2304,7 +2298,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax_sysadmin, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, },#endif {@@ -2564,7 +2558,7 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec_minmax, - .extra1 = &neg_one, + .extra1 = SYSCTL_NEG_ONE, },#endif #ifdef CONFIG_RT_MUTEXES @@ -2626,7 +2620,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = perf_cpu_time_max_percent_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "perf_event_max_stack",@@ -2644,7 +2638,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = perf_event_max_stack_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_thousand, + .extra2 = SYSCTL_ONE_THOUSAND, },#endif {@@ -2675,7 +2669,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = bpf_unpriv_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "bpf_stats_enabled",@@ -2729,7 +2723,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = overcommit_policy_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "panic_on_oom",@@ -2738,7 +2732,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "oom_kill_allocating_task",@@ -2783,7 +2777,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = dirty_background_ratio_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "dirty_background_bytes",@@ -2800,7 +2794,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = dirty_ratio_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "dirty_bytes",@@ -2840,7 +2834,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, - .extra2 = &two_hundred, + .extra2 = SYSCTL_TWO_HUNDRED, },#ifdef CONFIG_HUGETLB_PAGE {@@ -2897,7 +2891,7 @@ static struct ctl_table vm_table[] = { .mode = 0200, .proc_handler = drop_caches_sysctl_handler, .extra1 = SYSCTL_ONE, - .extra2 = &four, + .extra2 = SYSCTL_FOUR, },#ifdef CONFIG_COMPACTION {@@ -2914,7 +2908,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = compaction_proactiveness_sysctl_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "extfrag_threshold",@@ -2959,7 +2953,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = watermark_scale_factor_sysctl_handler, .extra1 = SYSCTL_ONE, - .extra2 = &one_thousand, + .extra2 = SYSCTL_ONE_THOUSAND, }, { .procname = "percpu_pagelist_high_fraction",@@ -3038,7 +3032,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = sysctl_min_unmapped_ratio_sysctl_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, }, { .procname = "min_slab_ratio",@@ -3047,7 +3041,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = sysctl_min_slab_ratio_sysctl_handler, .extra1 = SYSCTL_ZERO, - .extra2 = &one_hundred, + .extra2 = SYSCTL_ONE_HUNDRED, },#endif #ifdef CONFIG_SMP @@ -3337,7 +3331,7 @@ static struct ctl_table fs_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "protected_regular",@@ -3346,7 +3340,7 @@ static struct ctl_table fs_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, }, { .procname = "suid_dumpable",kernel/sysctl.c static struct ctl_table fs_table[] = { · line 3355
Quoted 8 lines
kernel/sysctl.c +1
-1@@ -3355,7 +3349,7 @@ static struct ctl_table fs_table[] = { .mode = 0644, .proc_handler = proc_dointvec_minmax_coredump, .extra1 = SYSCTL_ZERO, - .extra2 = &two, + .extra2 = SYSCTL_TWO, },#if defined(CONFIG_BINFMT_MISC ) || defined(CONFIG_BINFMT_MISC_MODULE ) {.
Re: [PATCH v2 9/9] dnotify: move dnotify sysctl to dnotify.c
In reply to Luis ChamberlainQuoted 10 lines · Luis Chamberlain
From: Xiaoming Ni <[email protected]>
The
kernel/sysctl.cis a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic.
Quoted 6 lines
So move dnotify sysctls to
dnotify.cand use the newregister_sysctl_init()to register the sysctl interface.Signed-off-by Xiaoming Ni <[email protected]>[mcgrof: adjust the commit log to justify the move]Signed-off-by Luis Chamberlain <[email protected]>
Looks sane. Feel free to add:
Honza
Quoted 88 lines
--- fs/notify/dnotify/dnotify.c | 21 ++++++++++++++++++++- include/linux/dnotify.h | 1 - kernel/sysctl.c | 10 ---------- 3 files changed, 20 insertions(+), 12 deletions(-)fs/notify/dnotify/dnotify.c +20
-1@@ -19,7 +19,25 @@#include <linux/fdtable.h> #include <linux/fsnotify_backend.h> -int dir_notify_enable__read_mostly =1 ; +static int dir_notify_enable__read_mostly =1 ; +#ifdef CONFIG_SYSCTL +static struct ctl_table dnotify_sysctls[] = { + { + .procname ="dir-notify-enable" , + .data = &dir_notify_enable, + .maxlen =sizeof (int ), + .mode =0644 , + .proc_handler = proc_dointvec, + }, + {} +}; +static void __init dnotify_sysctl_init (void ) +{ +register_sysctl_init ("fs" , dnotify_sysctls); +} +#else +#define dnotify_sysctl_init () do { } while (0) +#endif static struct kmem_cache *dnotify_struct_cache__read_mostly ;static struct kmem_cache *dnotify_mark_cache__read_mostly ;@@ -386,6 +404,7 @@ static int __init dnotify_init(void) dnotify_group =fsnotify_alloc_group (&dnotify_fsnotify_ops);if (IS_ERR (dnotify_group))panic ("unable to allocate fsnotify group for dnotify\n" ); +dnotify_sysctl_init ();return 0 ; }include/linux/dnotify.h +0
-1@@ -29,7 +29,6 @@ struct dnotify_struct { FS_CREATE | FS_DN_RENAME |\ FS_MOVED_FROM | FS_MOVED_TO) -extern int dir_notify_enable;extern void dnotify_flush (struct file *,fl_owner_t );extern int fcntl_dirnotify (int ,struct file *,unsigned long );kernel/sysctl.c +0
-10@@ -48,7 +48,6 @@#include <linux/times.h> #include <linux/limits.h> #include <linux/dcache.h> -#include <linux/dnotify.h> #include <linux/syscalls.h> #include <linux/vmstat.h> #include <linux/nfs_fs.h> @@ -3090,15 +3089,6 @@ static struct ctl_table fs_table[] = { .proc_handler = proc_dointvec, },#endif -#ifdef CONFIG_DNOTIFY - { - .procname = "dir-notify-enable", - .data = &dir_notify_enable, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_dointvec, - }, -#endif #ifdef CONFIG_MMU #ifdef CONFIG_FILE_LOCKING {Signature · 1 line
2.33.0
Signature · 2 lines
Jan Kara <[email protected]> SUSE Labs, CR
Re: [PATCH v2 8/9] aio: move aio sysctl to aio.c
In reply to Luis ChamberlainQuoted 10 lines · Luis Chamberlain
From: Xiaoming Ni <[email protected]>
The
kernel/sysctl.cis a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic.
Quoted 6 lines
Move aio sysctl to
aio.cand use the newregister_sysctl_init()to register the sysctl interface for aio.Signed-off-by Xiaoming Ni <[email protected]>[mcgrof: adjust commit log to justify the move]Signed-off-by Luis Chamberlain <[email protected]>
Makes sense. Feel free to add:
Honza
Quoted 107 lines
--- fs/aio.c | 31 +++++++++++++++++++++++++++++-- include/linux/aio.h | 4 ---- kernel/sysctl.c | 17 ----------------- 3 files changed, 29 insertions(+), 23 deletions(-)fs/aio.c +29
-2@@ -219,9 +219,35 @@ struct aio_kiocb {/*------ sysctl variables----*/ static DEFINE_SPINLOCK (aio_nr_lock); -unsigned long aio_nr;/* current system wide number of aio requests */ -unsigned long aio_max_nr =0x10000 ;/* system wide maximum number of aio requests */ +static unsigned long aio_nr;/* current system wide number of aio requests */ +static unsigned long aio_max_nr =0x10000 ;/* system wide maximum number of aio requests */ /*----end sysctl variables---*/ +#ifdef CONFIG_SYSCTL +static struct ctl_table aio_sysctls[] = { + { + .procname ="aio-nr" , + .data = &aio_nr, + .maxlen =sizeof (aio_nr), + .mode =0444 , + .proc_handler = proc_doulongvec_minmax, + }, + { + .procname ="aio-max-nr" , + .data = &aio_max_nr, + .maxlen =sizeof (aio_max_nr), + .mode =0644 , + .proc_handler = proc_doulongvec_minmax, + }, + {} +}; + +static void __init aio_sysctl_init (void ) +{ +register_sysctl_init ("fs" , aio_sysctls); +} +#else +#define aio_sysctl_init () do { } while (0) +#endif static struct kmem_cache *kiocb_cachep;static struct kmem_cache *kioctx_cachep;@@ -274,6 +300,7 @@ static int __init aio_setup(void) kiocb_cachep =KMEM_CACHE (aio_kiocb,SLAB_HWCACHE_ALIGN |SLAB_PANIC ); kioctx_cachep =KMEM_CACHE (kioctx,SLAB_HWCACHE_ALIGN |SLAB_PANIC ); +aio_sysctl_init ();return 0 ; }__initcall (aio_setup);include/linux/aio.h +0
-4@@ -20,8 +20,4 @@ static inline void kiocb_set_cancel_fn(struct kiocb *req, kiocb_cancel_fn *cancel) { }#endif /* CONFIG_AIO */ -/* for sysctl: */ -extern unsigned long aio_nr; -extern unsigned long aio_max_nr; -#endif /* __LINUX__AIO_H */ kernel/sysctl.c +0
-17@@ -20,7 +20,6 @@*/ #include <linux/module.h> -#include <linux/aio.h> #include <linux/mm.h> #include <linux/swap.h> #include <linux/slab.h> @@ -3110,22 +3109,6 @@ static struct ctl_table fs_table[] = { .proc_handler = proc_dointvec, },#endif -#ifdef CONFIG_AIO - { - .procname = "aio-nr", - .data = &aio_nr, - .maxlen = sizeof(aio_nr), - .mode = 0444, - .proc_handler = proc_doulongvec_minmax, - }, - { - .procname = "aio-max-nr", - .data = &aio_max_nr, - .maxlen = sizeof(aio_max_nr), - .mode = 0644, - .proc_handler = proc_doulongvec_minmax, - }, -#endif /* CONFIG_AIO */#ifdef CONFIG_INOTIFY_USER { .procname = "inotify",Signature · 1 line
2.33.0
Signature · 2 lines
Jan Kara <[email protected]> SUSE Labs, CR
Re: [PATCH v2 2/9] sysctl: Move some boundary constants from sysctl.c to sysctl_vals
In reply to Xiaoming NiQuoted 43 lines · Xiaoming Ni
Quoted 17 lines · Eric W. Biederman
Quoted 6 lines · Luis Chamberlain
From: Xiaoming Ni <[email protected]>
sysctl has helpers which let us specify boundary values for a min or max int value. Since these are used for a boundary check only they don't change, so move these variables to
sysctl_valsto avoid adding duplicate variables. This will help with our cleanup ofkernel/sysctl.c.Ouch.
I kind of, sort of, have to protest.
Where the macros that use
sysctl_valsdon't have a type they have caused mysterious code breakage because people did not realize they can not be used with sysctls that take a long instead of an int.Quoted 6 lines
This came up with when the internal storage for ucounts see
kernel/ucount.cchanged from an int to a long. We were quite a while
tracking what was going on until we realized that the code could not useSYSCTL_ZEROandSYSCTL_INT_MAXand that we had to defined our ownthatSYSCTL_ZEROandSYSCTL_ZEROinvolve dozens of files and are used in hundreds of places. were long.static unsigned long
zero_ul;
static unsigned longone_ul= 1;
static unsigned longlong_max=LONG_MAX;EXPORT_SYMBOL(proc_doulongvec_minmax);Yes, min/max of type unsigned long is used in multiple sysctl interfaces. It is necessary to add an unsigned long
sysctl_valarray to avoid repeated definitions in different .c files.Quoted 4 lines
So before we extend something like this can we please change the macro naming convention so that it includes the size of the type we want.
The int type is the most widely used type. By default, numeric constants are also of the int type.
SYSCTL_ZEROandSYSCTL_ZEROinvolve dozens of files and are used in hundreds of places. Whether only non-int macros need to be named with their type size?Quoted 1 line
Quoted 6 lines
Quoted 4 lines
I am also not a fan of
sysctl_valsliving inproc_sysctl. They have nothing to do with the code in that file. They would do much better inkernel/sysctl.cclose to the helpers that use them.yes
Looking a little more. I think it makes sense to do something like:
include/linux/sysctl.h +2 -2
@@ -121,8 +121,8 @@ struct ctl_table {struct ctl_table *child;/* Deprecated */ proc_handler *proc_handler;/* Callback for text formatting */ struct ctl_table_poll *poll; -void *extra1; -void *extra2; +long min; +long max; } __randomize_layout;
Nearly every use of .extra1 and .extra2 are for min and max values. A long takes the same storage as a void * parameter.
So it would be a net saving in storage as you don't have separate storage for the values anywhere.
There are a few cases where .extra1 is used for something else so keeping a "void *extra" field will probably be needed.
By finishing the removal of the child field adding a "void *extra" field can be done at no storage cost.
Having the min and max parameters in the structure has the major advantage that there is no redirection, and no fancy games. People can just read the value from the structure initializer. Plus a conversion from int to long won't requiring changing the min and max constants.
So really I think instead of doubling down on the error prone case that we have and extending sysctl_vals we should just get rid of it entirely.
It is a bit more work to make that change but the long term result is much better.
Any chance we can do that for a cleanup instead of extending sysctl_vals?
Eric
Re: [PATCH v2 2/9] sysctl: Move some boundary constants from sysctl.c to sysctl_vals
Quoted 16 lines · Eric W. Biederman
Looking a little more. I think it makes sense to do something like:
include/linux/sysctl.h +2
-2@@ -121,8 +121,8 @@ struct ctl_table {struct ctl_table *child;/* Deprecated */ proc_handler *proc_handler;/* Callback for text formatting */ struct ctl_table_poll *poll; -void *extra1; -void *extra2; +long min; +long max; } __randomize_layout;
Quoted 1 line
Any chance we can do that for a cleanup instead of extending
sysctl_vals?
Essentially the change is *big*. We either do that *not yet implemented* change *now, and then have to rebase all the non-upstream work to adapt to that, or we do that after the cleanup.
Both I think are going to cause some suffering. I'd prefer to do it after though. Would you be OK with that?
Luis
Petr Mladek
reviewerRe: [PATCH v2 1/9] sysctl: add a new register_sysctl_init() interface
In reply to Luis ChamberlainQuoted 22 lines · Luis Chamberlain
From: Xiaoming Ni <[email protected]>
The
kernel/sysctl.cis a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic.
Today though folks heavily rely on tables on
kernel/sysctl.cso they can easily just extend this table with their needed sysctls. In order to help users move their sysctls out we need to provide a helper which can be used during code initialization.We special-case the initialization use of
register_sysctl()since it *is* safe to fail, given all that sysctls do is provide a dynamic interface to query or modify at runtime an existing variable. So the use case ofregister_sysctl()on init should *not* stop if the sysctls don't end up getting registered. It would be counter productive to stop boot if a simple sysctl registration failed.
Quoted 5 lines
Provide a helper for init then, and document the recommended init levels to use for callers of this routine. We will later use this in subsequent patches to start slimming down
kernel/sysctl.ctables and moving sysctl registration to the code which actually needs these sysctls.
Do we really need a new helper for this?
Is the failure acceptable only during system initialization?
The warning would be useful even for the original register_sysctl().
It should be up-to-the caller to decide if the failure is fatal or not. It might be enough to document the reasoning why a warning is enough in most cases.
Best Regards,
Petr
Re: [PATCH v2 4/9] watchdog: move watchdog sysctl interface to watchdog.c
In reply to Luis ChamberlainQuoted 15 lines · Luis Chamberlain
From: Xiaoming Ni <[email protected]>
The
kernel/sysctl.cis a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic of proc sysctl.
So, move the watchdog syscl interface to
watchdog.c. Useregister_sysctl()to register the sysctl interface to avoid merge conflicts when different features modifysysctl.cat the same time.
Quoted 4 lines
Signed-off-by Xiaoming Ni <[email protected]>Reviewed-by Kees Cook <[email protected]>[mcgrof: justify the move on the commit log]Signed-off-by Luis Chamberlain <[email protected]>
Best Regards,
Petr
Re: [PATCH v2 3/9] hung_task: Move hung_task sysctl interface to hung_task.c
In reply to Luis ChamberlainQuoted 10 lines · Luis Chamberlain
From: Xiaoming Ni <[email protected]>
The
kernel/sysctl.cis a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic.
Quoted 7 lines
So move
hung_tasksysctl interface tohung_task.cand useregister_sysctl()to register the sysctl interface.Signed-off-by Xiaoming Ni <[email protected]>Reviewed-by Kees Cook <[email protected]>[mcgrof: commit log refresh and fixed 2-3 0day reported compile issues]Signed-off-by Luis Chamberlain <[email protected]>
Best Regards,
Petr
Re: [PATCH v2 1/9] sysctl: add a new register_sysctl_init() interface
In reply to Petr MladekQuoted 23 lines · Petr Mladek
Quoted 22 lines · Luis Chamberlain
From: Xiaoming Ni <[email protected]>
The
kernel/sysctl.cis a kitchen sink where everyone leaves their dirty dishes, this makes it very difficult to maintain.To help with this maintenance let's start by moving sysctls to places where they actually belong. The proc sysctl maintainers do not want to know what sysctl knobs you wish to add for your own piece of code, we just care about the core logic.
Today though folks heavily rely on tables on
kernel/sysctl.cso they can easily just extend this table with their needed sysctls. In order to help users move their sysctls out we need to provide a helper which can be used during code initialization.We special-case the initialization use of
register_sysctl()since it *is* safe to fail, given all that sysctls do is provide a dynamic interface to query or modify at runtime an existing variable. So the use case ofregister_sysctl()on init should *not* stop if the sysctls don't end up getting registered. It would be counter productive to stop boot if a simple sysctl registration failed.
Quoted 8 lines
Quoted 5 lines
Provide a helper for init then, and document the recommended init levels to use for callers of this routine. We will later use this in subsequent patches to start slimming down
kernel/sysctl.ctables and moving sysctl registration to the code which actually needs these sysctls.Do we really need a new helper for this?
Is the failure acceptable only during system initialization?
Yes because it is __init and we allow / guide folks to *think* clearly about not stopping the init process when it comes to sysctls on failure.
Quoted 1 line
The warning would be useful even for the original
register_sysctl().
We can open code those.
Quoted 3 lines
It should be up-to-the caller to decide if the failure is fatal or not. It might be enough to document the reasoning why a warning is enough in most cases.
For most case I have seen so far special casing init seems like a worthy objective. When we're done with the full conversion we can re-visit things but at this point I can't say sharing this outside of init uses makes too much sense.
Luis