Skip to contentSkip to conversation

PATCH - NGROUPS 2.6.2rc2 + fixups

PATCH Tim Hockin 6 participants 6 messages Jan 29, 2004 – Feb 19, 2004

Subsystems (11)
lore
Share & tools
lore

Tim Hockin

author
#0 lore raw

Linus, Andrew,

Attached is a patch to remove the NGROUPS limit (again). I've fixed up the issues that people have pointed out this week. It incorporates a fixed limit which can be easily changed in a user wanted even more groups.

This patch is against linux-2.6.2rc2. If it is OK, I can make it bitkeeper accessible.

What think?

Signature · 4 lines
Tim Hockin
Sun Microsystems, Linux Software Engineering
[email protected]
All opinions are my own, not Sun's

===== include/linux/sched.h 1.178 vs edited =====

include/linux/sched.h +28 -2
@@ -329,6 +329,33 @@
 struct io_context;			/* See blkdev.h */
 void exit_io_context(void);
 
+#define NGROUPS_SMALL		32
+#define NGROUPS_PER_BLOCK	((int)(EXEC_PAGESIZE / sizeof(gid_t)))
+struct group_info {
+	int ngroups;
+	atomic_t usage;
+	gid_t small_block[NGROUPS_SMALL];
+	int nblocks;
+	gid_t *blocks[0];
+};
+
+#define get_group_info(group_info) do { \
+	atomic_inc(&(group_info)->usage); \
+} while (0)
+
+#define put_group_info(group_info) do { \
+	if (atomic_dec_and_test(&(group_info)->usage)) \
+		groups_free(group_info); \
+} while (0)
+
+struct group_info *groups_alloc(int gidsetsize);
+void groups_free(struct group_info *group_info);
+int set_current_groups(struct group_info *group_info);
+/* access the groups "array" with this macro */
+#define GROUP_AT(gi, i) \
+    ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
+
+
 struct task_struct {
 	volatile long state;	/* -1 unrunnable, 0 runnable, >0 stopped */
 	struct thread_info *thread_info;
@@ -403,8 +430,7 @@
 /* process credentials */
 	uid_t uid,euid,suid,fsuid;
 	gid_t gid,egid,sgid,fsgid;
-	int ngroups;
-	gid_t	groups[NGROUPS];
+	struct group_info *group_info;
 	kernel_cap_t   cap_effective, cap_inheritable, cap_permitted;
 	int keep_capabilities:1;
 	struct user_struct *user;

===== include/linux/init_task.h 1.27 vs edited =====

include/linux/init_task.h +3 -0
@@ -56,6 +56,8 @@
 	.siglock	= SPIN_LOCK_UNLOCKED, 		\
 }
 
+extern struct group_info init_groups;
+
 /*
  *  INIT_TASK is used to set up the first task table, touch at
  * your own risk!. Base=0, limit=0x1fffff (=2MB)
@@ -87,6 +89,7 @@
 	.real_timer	= {						\
 		.function	= it_real_fn				\
 	},								\
+	.group_info	= &init_groups,					\
 	.cap_effective	= CAP_INIT_EFF_SET,				\
 	.cap_inheritable = CAP_INIT_INH_SET,				\
 	.cap_permitted	= CAP_FULL_SET,					\

===== include/linux/limits.h 1.3 vs edited =====

include/linux/limits.h +1 -1
@@ -3,7 +3,7 @@
 
 #define NR_OPEN	        1024
 
-#define NGROUPS_MAX       32	/* supplemental group IDs are available */
+#define NGROUPS_MAX    65536	/* supplemental group IDs are available */
 #define ARG_MAX       131072	/* # bytes of args + environ for exec() */
 #define CHILD_MAX        999    /* no limit :-) */
 #define OPEN_MAX         256	/* # open files a process may have */

===== kernel/sys.c 1.69 vs edited =====

kernel/sys.c +201 -35

281 lines not on this page · raw message

===== kernel/fork.c 1.154 vs edited =====

kernel/fork.c +3 -0
@@ -86,6 +86,7 @@
 
 	security_task_free(tsk);
 	free_uid(tsk->user);
+	put_group_info(tsk->group_info);
 	free_task(tsk);
 }
 
@@ -878,6 +879,7 @@
 
 	atomic_inc(&p->user->__count);
 	atomic_inc(&p->user->processes);
+	get_group_info(p->group_info);
 
 	/*
 	 * If multiple threads are within copy_process(), then this check
@@ -1084,6 +1086,7 @@
 bad_fork_cleanup_put_domain:
 	module_put(p->thread_info->exec_domain->module);
 bad_fork_cleanup_count:
+	put_group_info(p->group_info);
 	atomic_dec(&p->user->processes);
 	free_uid(p->user);
 bad_fork_free:

===== kernel/uid16.c 1.5 vs edited =====

kernel/uid16.c +74 -23

119 lines not on this page · raw message

===== fs/proc/array.c 1.55 vs edited =====

fs/proc/array.c +4 -2
@@ -176,8 +176,10 @@
 		p->files ? p->files->max_fds : 0);
 	task_unlock(p);
 
-	for (g = 0; g < p->ngroups; g++)
-		buffer += sprintf(buffer, "%d ", p->groups[g]);
+	get_group_info(p->group_info);
+	for (g = 0; g < min(p->group_info->ngroups,NGROUPS_SMALL); g++)
+		buffer += sprintf(buffer, "%d ", GROUP_AT(p->group_info,g));
+	put_group_info(p->group_info);
 
 	buffer += sprintf(buffer, "\n");
 	return buffer;

===== include/linux/sunrpc/auth.h 1.9 vs edited =====

include/linux/sunrpc/auth.h +1 -2
@@ -28,8 +28,7 @@
 struct auth_cred {
 	uid_t	uid;
 	gid_t	gid;
-	int	ngroups;
-	gid_t	*groups;
+	struct group_info *group_info;
 };
 
 /*

===== include/linux/sunrpc/svcauth.h 1.9 vs edited =====

include/linux/sunrpc/svcauth.h +2 -1
@@ -16,10 +16,11 @@
 #include <linux/sunrpc/cache.h>
 #include <linux/hash.h>
 
+#define SVC_CRED_NGROUPS	32
 struct svc_cred {
 	uid_t			cr_uid;
 	gid_t			cr_gid;
-	gid_t			cr_groups[NGROUPS];
+	gid_t			cr_groups[SVC_CRED_NGROUPS];
 };
 
 struct svc_rqst;		/* forward decl */

===== net/sunrpc/auth.c 1.12 vs edited =====

net/sunrpc/auth.c +21 -14
@@ -246,34 +246,41 @@
 struct rpc_cred *
 rpcauth_lookupcred(struct rpc_auth *auth, int taskflags)
 {
-	struct auth_cred acred = {
-		.uid = current->fsuid,
-		.gid = current->fsgid,
-		.ngroups = current->ngroups,
-		.groups = current->groups,
-	};
+	struct auth_cred acred;
+	struct rpc_cred *ret;
+
+	get_group_info(current->group_info);
+	acred.uid = current->fsuid;
+	acred.gid = current->fsgid;
+	acred.group_info = current->group_info;
+
 	dprintk("RPC:     looking up %s cred\n",
 		auth->au_ops->au_name);
-	return rpcauth_lookup_credcache(auth, &acred, taskflags);
+	ret = rpcauth_lookup_credcache(auth, &acred, taskflags);
+	put_group_info(current->group_info);
+	return ret;
 }
 
 struct rpc_cred *
 rpcauth_bindcred(struct rpc_task *task)
 {
 	struct rpc_auth *auth = task->tk_auth;
-	struct auth_cred acred = {
-		.uid = current->fsuid,
-		.gid = current->fsgid,
-		.ngroups = current->ngroups,
-		.groups = current->groups,
-	};
+	struct auth_cred acred;
+	struct rpc_cred *ret;
+
+	get_group_info(current->group_info);
+	acred.uid = current->fsuid;
+	acred.gid = current->fsgid;
+	acred.group_info = current->group_info;
 
 	dprintk("RPC: %4d looking up %s cred\n",
 		task->tk_pid, task->tk_auth->au_ops->au_name);
 	task->tk_msg.rpc_cred = rpcauth_lookup_credcache(auth, &acred, task->tk_flags);
 	if (task->tk_msg.rpc_cred == 0)
 		task->tk_status = -ENOMEM;
-	return task->tk_msg.rpc_cred;
+	ret = task->tk_msg.rpc_cred;
+	put_group_info(current->group_info);
+	return ret;
 }
 
 void

===== net/sunrpc/auth_unix.c 1.11 vs edited =====

net/sunrpc/auth_unix.c +4 -4
@@ -82,7 +82,7 @@
 		cred->uc_gid = cred->uc_pgid = 0;
 		cred->uc_gids[0] = NOGROUP;
 	} else {
-		int groups = acred->ngroups;
+		int groups = acred->group_info->ngroups;
 		if (groups > NFS_NGROUPS)
 			groups = NFS_NGROUPS;
 
@@ -91,7 +91,7 @@
 		cred->uc_puid = current->uid;
 		cred->uc_pgid = current->gid;
 		for (i = 0; i < groups; i++)
-			cred->uc_gids[i] = (gid_t) acred->groups[i];
+			cred->uc_gids[i] = GROUP_AT(acred->group_info, i);
 		if (i < NFS_NGROUPS)
 		  cred->uc_gids[i] = NOGROUP;
 	}
@@ -126,11 +126,11 @@
 		 || cred->uc_pgid != current->gid)
 			return 0;
 
-		groups = acred->ngroups;
+		groups = acred->group_info->ngroups;
 		if (groups > NFS_NGROUPS)
 			groups = NFS_NGROUPS;
 		for (i = 0; i < groups ; i++)
-			if (cred->uc_gids[i] != (gid_t) acred->groups[i])
+			if (cred->uc_gids[i] != GROUP_AT(acred->group_info, i))
 				return 0;
 		return 1;
 	}

===== net/sunrpc/svcauth_unix.c 1.20 vs edited =====

net/sunrpc/svcauth_unix.c +2 -2
@@ -434,11 +434,11 @@
 	if (slen > 16 || (len -= (slen + 2)*4) < 0)
 		goto badcred;
 	for (i = 0; i < slen; i++)
-		if (i < NGROUPS)
+		if (i < SVC_CRED_NGROUPS)
 			cred->cr_groups[i] = ntohl(svc_getu32(argv));
 		else
 			svc_getu32(argv);
-	if (i < NGROUPS)
+	if (i < SVC_CRED_NGROUPS)
 		cred->cr_groups[i] = NOGROUP;
 
 	if (svc_getu32(argv) != RPC_AUTH_NULL || svc_getu32(argv) != 0) {

===== fs/nfsd/auth.c 1.2 vs edited =====

fs/nfsd/auth.c +7 -4
@@ -10,12 +10,15 @@
 #include <linux/sunrpc/svcauth.h>
 #include <linux/nfsd/nfsd.h>
 
+extern asmlinkage long sys_setgroups(int gidsetsize, gid_t *grouplist);
+
 #define	CAP_NFSD_MASK (CAP_FS_MASK|CAP_TO_MASK(CAP_SYS_RESOURCE))
 void
 nfsd_setuser(struct svc_rqst *rqstp, struct svc_export *exp)
 {
 	struct svc_cred	*cred = &rqstp->rq_cred;
 	int		i;
+	gid_t		groups[SVC_CRED_NGROUPS];
 
 	if (exp->ex_flags & NFSEXP_ALLSQUASH) {
 		cred->cr_uid = exp->ex_anon_uid;
@@ -26,7 +29,7 @@
 			cred->cr_uid = exp->ex_anon_uid;
 		if (!cred->cr_gid)
 			cred->cr_gid = exp->ex_anon_gid;
-		for (i = 0; i < NGROUPS; i++)
+		for (i = 0; i < SVC_CRED_NGROUPS; i++)
 			if (!cred->cr_groups[i])
 				cred->cr_groups[i] = exp->ex_anon_gid;
 	}
@@ -39,13 +42,13 @@
 		current->fsgid = cred->cr_gid;
 	else
 		current->fsgid = exp->ex_anon_gid;
-	for (i = 0; i < NGROUPS; i++) {
+	for (i = 0; i < SVC_CRED_NGROUPS; i++) {
 		gid_t group = cred->cr_groups[i];
 		if (group == (gid_t) NOGROUP)
 			break;
-		current->groups[i] = group;
+		groups[i] = group;
 	}
-	current->ngroups = i;
+	sys_setgroups(i, groups);
 
 	if ((cred->cr_uid)) {
 		cap_t(current->cap_effective) &= ~CAP_NFSD_MASK;

===== fs/nfsd/nfs4state.c 1.21 vs edited =====

fs/nfsd/nfs4state.c +1 -1
@@ -244,7 +244,7 @@
 
 	target->cr_uid = source->cr_uid;
 	target->cr_gid = source->cr_gid;
-	for(i = 0; i < NGROUPS; i++)
+	for(i = 0; i < SVC_CRED_NGROUPS; i++)
 		target->cr_groups[i] = source->cr_groups[i];
 }
 

===== include/linux/security.h 1.28 vs edited =====

include/linux/security.h +6 -7
@@ -554,9 +554,8 @@
  *	Return 0 if permission is granted.
  * @task_setgroups:
  *	Check permission before setting the supplementary group set of the
- *	current process to @grouplist.
- *	@gidsetsize contains the number of elements in @grouplist.
- *	@grouplist contains the array of gids.
+ *	current process.
+ *	@group_info contains the new group information.
  *	Return 0 if permission is granted.
  * @task_setnice:
  *	Check permission before setting the nice value of @p to @nice.
@@ -1116,7 +1115,7 @@
 	int (*task_setpgid) (struct task_struct * p, pid_t pgid);
 	int (*task_getpgid) (struct task_struct * p);
 	int (*task_getsid) (struct task_struct * p);
-	int (*task_setgroups) (int gidsetsize, gid_t * grouplist);
+	int (*task_setgroups) (struct group_info *group_info);
 	int (*task_setnice) (struct task_struct * p, int nice);
 	int (*task_setrlimit) (unsigned int resource, struct rlimit * new_rlim);
 	int (*task_setscheduler) (struct task_struct * p, int policy,
@@ -1670,9 +1669,9 @@
 	return security_ops->task_getsid (p);
 }
 
-static inline int security_task_setgroups (int gidsetsize, gid_t *grouplist)
+static inline int security_task_setgroups (struct group_info *group_info)
 {
-	return security_ops->task_setgroups (gidsetsize, grouplist);
+	return security_ops->task_setgroups (group_info);
 }
 
 static inline int security_task_setnice (struct task_struct *p, int nice)
@@ -2299,7 +2298,7 @@
 	return 0;
 }
 
-static inline int security_task_setgroups (int gidsetsize, gid_t *grouplist)
+static inline int security_task_setgroups (struct group_info *group_info)
 {
 	return 0;
 }

===== security/dummy.c 1.30 vs edited =====

security/dummy.c +1 -1
@@ -539,7 +539,7 @@
 	return 0;
 }
 
-static int dummy_task_setgroups (int gidsetsize, gid_t * grouplist)
+static int dummy_task_setgroups (struct group_info *group_info)
 {
 	return 0;
 }

===== security/selinux/hooks.c 1.20 vs edited =====

security/selinux/hooks.c +1 -1
@@ -2265,7 +2265,7 @@
 	return task_has_perm(current, p, PROCESS__GETSESSION);
 }
 
-static int selinux_task_setgroups(int gidsetsize, gid_t *grouplist)
+static int selinux_task_setgroups(struct group_info *group_info)
 {
 	/* See the comment for setuid above. */
 	return 0;

===== arch/ia64/ia32/sys_ia32.c 1.87 vs edited =====

arch/ia64/ia32/sys_ia32.c +81 -27

125 lines not on this page · raw message

===== arch/mips/kernel/sysirix.c 1.18 vs edited =====

arch/mips/kernel/sysirix.c +1 -1
@@ -368,7 +368,7 @@
 			retval = HZ;
 			goto out;
 		case 4:
-			retval = NGROUPS;
+			retval = NGROUPS_MAX;
 			goto out;
 		case 5:
 			retval = NR_OPEN;

===== arch/s390/kernel/compat_linux.c 1.10 vs edited =====

arch/s390/kernel/compat_linux.c +72 -18

112 lines not on this page · raw message

===== arch/sparc64/kernel/sys_sparc32.c 1.86 vs edited =====

arch/sparc64/kernel/sys_sparc32.c +72 -18

112 lines not on this page · raw message

===== include/asm-alpha/param.h 1.2 vs edited =====

include/asm-alpha/param.h +0 -4
@@ -19,10 +19,6 @@
 
 #define EXEC_PAGESIZE	8192
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-arm/param.h 1.5 vs edited =====

include/asm-arm/param.h +0 -4
@@ -26,10 +26,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS         32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP         (-1)
 #endif

===== include/asm-arm26/param.h 1.1 vs edited =====

include/asm-arm26/param.h +0 -4
@@ -22,10 +22,6 @@
 # define HZ		100
 #endif
 
-#ifndef NGROUPS
-#define NGROUPS         32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP         (-1)
 #endif

===== include/asm-cris/param.h 1.2 vs edited =====

include/asm-cris/param.h +0 -4
@@ -14,10 +14,6 @@
 
 #define EXEC_PAGESIZE	8192
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-h8300/param.h 1.1 vs edited =====

include/asm-h8300/param.h +0 -4
@@ -14,10 +14,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-i386/param.h 1.2 vs edited =====

include/asm-i386/param.h +0 -4
@@ -13,10 +13,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-ia64/param.h 1.5 vs edited =====

include/asm-ia64/param.h +0 -4
@@ -12,10 +12,6 @@
 
 #define EXEC_PAGESIZE	65536
 
-#ifndef NGROUPS
-# define NGROUPS	32
-#endif
-
 #ifndef NOGROUP
 # define NOGROUP	(-1)
 #endif

===== include/asm-m68k/param.h 1.2 vs edited =====

include/asm-m68k/param.h +0 -4
@@ -13,10 +13,6 @@
 
 #define EXEC_PAGESIZE	8192
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-m68knommu/param.h 1.1 vs edited =====

include/asm-m68knommu/param.h +0 -4
@@ -44,10 +44,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-mips/param.h 1.3 vs edited =====

include/asm-mips/param.h +0 -4
@@ -33,10 +33,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-parisc/param.h 1.3 vs edited =====

include/asm-parisc/param.h +0 -4
@@ -18,10 +18,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-ppc/param.h 1.6 vs edited =====

include/asm-ppc/param.h +0 -4
@@ -13,10 +13,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-ppc64/param.h 1.2 vs edited =====

include/asm-ppc64/param.h +0 -4
@@ -20,10 +20,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-s390/param.h 1.3 vs edited =====

include/asm-s390/param.h +0 -4
@@ -21,10 +21,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-sh/param.h 1.2 vs edited =====

include/asm-sh/param.h +0 -4
@@ -17,10 +17,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-sparc/param.h 1.2 vs edited =====

include/asm-sparc/param.h +0 -4
@@ -14,10 +14,6 @@
 
 #define EXEC_PAGESIZE	8192    /* Thanks for sun4's we carry baggage... */
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-sparc64/param.h 1.2 vs edited =====

include/asm-sparc64/param.h +0 -4
@@ -14,10 +14,6 @@
 
 #define EXEC_PAGESIZE	8192    /* Thanks for sun4's we carry baggage... */
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-um/param.h 1.1 vs edited =====

include/asm-um/param.h +0 -4
@@ -3,10 +3,6 @@
 
 #define EXEC_PAGESIZE   4096
 
-#ifndef NGROUPS
-#define NGROUPS         32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP         (-1)
 #endif

===== include/asm-v850/param.h 1.1 vs edited =====

include/asm-v850/param.h +0 -4
@@ -18,10 +18,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif

===== include/asm-x86_64/param.h 1.3 vs edited =====

include/asm-x86_64/param.h +0 -4
@@ -13,10 +13,6 @@
 
 #define EXEC_PAGESIZE	4096
 
-#ifndef NGROUPS
-#define NGROUPS		32
-#endif
-
 #ifndef NOGROUP
 #define NOGROUP		(-1)
 #endif
#1 lore raw
Quoted 2 lines · Tim Hockin

What think?

I still don't understand the complexity.

Why the list of pages? Is there really any valid use for this that could overflow a simple "kmalloc()"? How many groups do people really really need?

I just find it wrong to go from 32 to millions. And a few thousand can trivially be handled by a normal "kmalloc()".

Basically, I find overdesign physically nauseating. This is less so than the original stuff, but I'm still finding myself asking "Why?".

Linus

#2 lore raw

In message <[email protected]> you write:

Quoted 11 lines · Tim Hockin
Quoted 2 lines

What think?

I still don't understand the complexity.

Why the list of pages? Is there really any valid use for this that could overflow a simple "kmalloc()"? How many groups do people really really need?

I was happy with kmalloc, and no sorting. Simple patch. Tim complained that he had some wierd-ass users who hit kmalloc limits w/ fragmentation. I added about 10 lines of code to fall back to vmalloc for that very rare case, and you scotched that.

ie. Don't blame Tim: you led us here 8(

Rusty.
--
Anyone who quotes me in their sig is an idiot. -- Rusty Russell.

#3 lore raw

Hi,

anything new about this patch?

I'd like to see this problem solved too, also because of a samba, mad integration with the need of > 256 groups per user.

regards
Stefan

Signature · 6 lines
--------------------------------------------------------------------
Stefan Völkel                            [email protected]
Millenux GmbH                              mobile: +49.170.79177.17
Lilienthalstraße 2                          phone: +49.711.88770.300
70825 Stuttgart-Korntal                       fax: +49.711.88770.349
     -= linux without limits -=- http://linux.zSeries.org/ =-
#4 lore raw
Quoted 6 lines · Stefan Voelkel

Hi,

anything new about this patch?

I'd like to see this problem solved too, also because of a samba, mad integration with the need of > 256 groups per user.

Has anyone else seen a limit of 16 groups with nss-ldap?

#5 lore raw
In reply to Rusty Russell
Quoted 13 lines · Rusty Russell

In message <[email protected]> you write:

Quoted 11 lines · Tim Hockin
Quoted 2 lines

What think?

I still don't understand the complexity.

Why the list of pages? Is there really any valid use for this that could overflow a simple "kmalloc()"? How many groups do people really really need?

Quoted 8 lines

I was happy with kmalloc, and no sorting. Simple patch. Tim complained that he had some wierd-ass users who hit kmalloc limits w/ fragmentation. I added about 10 lines of code to fall back to vmalloc for that very rare case, and you scotched that.

ie. Don't blame Tim: you led us here 8(

Rusty.

So what was the conclusion? It would be nice to get this resolved and merged in finally..

-- Pasi Kärkkäinen

                                   ^
                                .     .
                                 Linux
                              /    -    \
                             Choice.of.the
                           .Next.Generation.