Skip to contentSkip to conversation

lockup during selinux policy load.

PATCH In review Dave Jones 3 participants 10 messages Sep 16, 2013 – Oct 3, 2013 SELinux security module

lore
Share & tools
lore

Dave Jones

author
#0 lore raw

On a slow machine (with debugging enabled), during a yum update I get the soft lockup detector kicking in when it gets to reloading the selinux policy. It looks like this..

BUG: soft lockup - CPU#2 stuck for 23s! [load_policy:19045]
irq event stamp: 2368864
hardirqs last  enabled at (2368863): [<ffffffff8152ab5d>] __slab_alloc.constprop.78+0x4c0/0x4d7
hardirqs last disabled at (2368864): [<ffffffff8153c82a>] apic_timer_interrupt+0x6a/0x80
softirqs last  enabled at (2368554): [<ffffffff810480b9>] __do_softirq+0x169/0x200
softirqs last disabled at (2368539): [<ffffffff8104837d>] irq_exit+0x11d/0x140
CPU: 2 PID: 19045 Comm: load_policy Not tainted 3.11.0+ #16
Hardware name:                  /D510MO, BIOS MOPNV10J.86A.0175.2010.0308.0620 03/08/2010
task: ffff88005ab38000 ti: ffff880019620000 task.ti: ffff880019620000
RIP: 0010:[<ffffffff81274273>]  [<ffffffff81274273>] strcmp+0x23/0x40
RSP: 0018:ffff880019621818  EFLAGS: 00000246
RAX: 0000000000000063 RBX: ffff880018090ca8 RCX: 0000000000006070
RDX: ffff88000781d8f0 RSI: ffff88000781d8f1 RDI: ffff880079caed21
RBP: ffff880019621818 R08: ffff88006345a290 R09: ffff880018091680
R10: 0000000000000001 R11: 0000000000000001 R12: 0000000000000292
R13: 0000000180160016 R14: ffff88007b804488 R15: ffffffff81221430
FS:  00007f73e1212800(0000) GS:ffff88007e600000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f73e0857aa0 CR3: 0000000076fea000 CR4: 00000000000007e0
Stack:
 ffff880019621828 ffffffff81221ddf ffff880019621850 ffffffff81221c27
 ffff8800180ac6d8 ffff880019621b88 ffffffff82759620 ffff8800196218a0
 ffffffff8122e96c 0000000181221e05 ffff880018091680 0000000000000080
Call Trace:
 [<ffffffff81221ddf>] symcmp+0xf/0x20
 [<ffffffff81221c27>] hashtab_search+0x47/0x80
 [<ffffffff8122e96c>] mls_convert_context+0xdc/0x1c0
 [<ffffffff812294e8>] convert_context+0x378/0x460
 [<ffffffff81229170>] ? security_context_to_sid_core+0x240/0x240
 [<ffffffff812221b5>] sidtab_map+0x45/0x80
 [<ffffffff8122bb9f>] security_load_policy+0x3ff/0x580
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff810786dd>] ? sched_clock_local+0x1d/0x80
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103096a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff810786dd>] ? sched_clock_local+0x1d/0x80
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103096a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff81534ddc>] ? retint_restore_args+0xe/0xe
 [<ffffffff8109c82d>] ? trace_hardirqs_on_caller+0xfd/0x1c0
 [<ffffffff81279a2e>] ? trace_hardirqs_on_thunk+0x3a/0x3f
 [<ffffffff810d28a8>] ? rcu_irq_exit+0x68/0xb0
 [<ffffffff81534ddc>] ? retint_restore_args+0xe/0xe
 [<ffffffff8121e947>] sel_write_load+0xa7/0x770
 [<ffffffff81139633>] ? vfs_write+0x1c3/0x200
 [<ffffffff81210e8e>] ? security_file_permission+0x1e/0xa0
 [<ffffffff8113952b>] vfs_write+0xbb/0x200
 [<ffffffff811581c7>] ? fget_light+0x397/0x4b0
 [<ffffffff81139c27>] SyS_write+0x47/0xa0
 [<ffffffff8153bde4>] tracesys+0xdd/0xe2
Code: 0f 1f 84 00 00 00 00 00 55 48 89 e5 eb 0e 66 2e 0f 1f 84 00 00 00 00 00 84 c0 74 1c 48 83 c7 01 0f b6 47 ff 48 83 c6 01 3a 46 ff <74> eb 19 c0 83 c8 01 5d c3 0f 1f 40 00 31 c0 5d c3 66 66 66 2e 

23s in the kernel is an eternity. Short of rearchitecting how policy loads are done, perhaps we could do something like this ? (untested, and 10000 is arbitarily chosen, may need to be adjusted)

thoughts ?

Dave

security/selinux/ss/hashtab.c +8 -1
index 933e735..69a0587 100644
@@ -75,14 +75,21 @@ void *hashtab_search(struct hashtab *h, const void *key)
 {
 	u32 hvalue;
 	struct hashtab_node *cur;
+	int count;
 
 	if (!h)
 		return NULL;
 
 	hvalue = h->hash_value(h, key);
 	cur = h->htable[hvalue];
-	while (cur && h->keycmp(h, key, cur->key) > 0)
+	while (cur && h->keycmp(h, key, cur->key) > 0) {
 		cur = cur->next;
+		count++;
+		if (count == 10000) {
+			touch_softlockup_watchdog();
+			count = 0;
+		}
+	}
 
 	if (cur == NULL || (h->keycmp(h, key, cur->key) != 0))
 		return NULL;
#1 lore raw
Quoted 59 lines · Dave Jones

On a slow machine (with debugging enabled), during a yum update I get the soft lockup detector kicking in when it gets to reloading the selinux policy. It looks like this..

BUG: soft lockup - CPU#2 stuck for 23s! [load_policy:19045]
irq event stamp: 2368864
hardirqs last  enabled at (2368863): [<ffffffff8152ab5d>] __slab_alloc.constprop.78+0x4c0/0x4d7
hardirqs last disabled at (2368864): [<ffffffff8153c82a>] apic_timer_interrupt+0x6a/0x80
softirqs last  enabled at (2368554): [<ffffffff810480b9>] __do_softirq+0x169/0x200
softirqs last disabled at (2368539): [<ffffffff8104837d>] irq_exit+0x11d/0x140
CPU: 2 PID: 19045 Comm: load_policy Not tainted 3.11.0+ #16
Hardware name:                  /D510MO, BIOS MOPNV10J.86A.0175.2010.0308.0620 03/08/2010
task: ffff88005ab38000 ti: ffff880019620000 task.ti: ffff880019620000
RIP: 0010:[<ffffffff81274273>]  [<ffffffff81274273>] strcmp+0x23/0x40
RSP: 0018:ffff880019621818  EFLAGS: 00000246
RAX: 0000000000000063 RBX: ffff880018090ca8 RCX: 0000000000006070
RDX: ffff88000781d8f0 RSI: ffff88000781d8f1 RDI: ffff880079caed21
RBP: ffff880019621818 R08: ffff88006345a290 R09: ffff880018091680
R10: 0000000000000001 R11: 0000000000000001 R12: 0000000000000292
R13: 0000000180160016 R14: ffff88007b804488 R15: ffffffff81221430
FS:  00007f73e1212800(0000) GS:ffff88007e600000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f73e0857aa0 CR3: 0000000076fea000 CR4: 00000000000007e0
Stack:
 ffff880019621828 ffffffff81221ddf ffff880019621850 ffffffff81221c27
 ffff8800180ac6d8 ffff880019621b88 ffffffff82759620 ffff8800196218a0
 ffffffff8122e96c 0000000181221e05 ffff880018091680 0000000000000080
Call Trace:
 [<ffffffff81221ddf>] symcmp+0xf/0x20
 [<ffffffff81221c27>] hashtab_search+0x47/0x80
 [<ffffffff8122e96c>] mls_convert_context+0xdc/0x1c0
 [<ffffffff812294e8>] convert_context+0x378/0x460
 [<ffffffff81229170>] ? security_context_to_sid_core+0x240/0x240
 [<ffffffff812221b5>] sidtab_map+0x45/0x80
 [<ffffffff8122bb9f>] security_load_policy+0x3ff/0x580
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff810786dd>] ? sched_clock_local+0x1d/0x80
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103096a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff810786dd>] ? sched_clock_local+0x1d/0x80
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103096a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff81534ddc>] ? retint_restore_args+0xe/0xe
 [<ffffffff8109c82d>] ? trace_hardirqs_on_caller+0xfd/0x1c0
 [<ffffffff81279a2e>] ? trace_hardirqs_on_thunk+0x3a/0x3f
 [<ffffffff810d28a8>] ? rcu_irq_exit+0x68/0xb0
 [<ffffffff81534ddc>] ? retint_restore_args+0xe/0xe
 [<ffffffff8121e947>] sel_write_load+0xa7/0x770
 [<ffffffff81139633>] ? vfs_write+0x1c3/0x200
 [<ffffffff81210e8e>] ? security_file_permission+0x1e/0xa0
 [<ffffffff8113952b>] vfs_write+0xbb/0x200
 [<ffffffff811581c7>] ? fget_light+0x397/0x4b0
 [<ffffffff81139c27>] SyS_write+0x47/0xa0
 [<ffffffff8153bde4>] tracesys+0xdd/0xe2
Code: 0f 1f 84 00 00 00 00 00 55 48 89 e5 eb 0e 66 2e 0f 1f 84 00 00 00 00 00 84 c0 74 1c 48 83 c7 01 0f b6 47 ff 48 83 c6 01 3a 46 ff <74> eb 19 c0 83 c8 01 5d c3 0f 1f 40 00 31 c0 5d c3 66 66 66 2e 
Quoted 4 lines

23s in the kernel is an eternity. Short of rearchitecting how policy loads are done, perhaps we could do something like this ? (untested, and 10000 is arbitarily chosen, may need to be adjusted)

thoughts ?

Maybe put a cond_resched() within the ebitmap_for_each_positive_bit() loop in mls_convert_context()?

Quoted 32 lines

Dave

security/selinux/ss/hashtab.c +8 -1
index 933e735..69a0587 100644
@@ -75,14 +75,21 @@ void *hashtab_search(struct hashtab *h, const void *key)
 {
 	u32 hvalue;
 	struct hashtab_node *cur;
+	int count;
 
 	if (!h)
 		return NULL;
 
 	hvalue = h->hash_value(h, key);
 	cur = h->htable[hvalue];
-	while (cur && h->keycmp(h, key, cur->key) > 0)
+	while (cur && h->keycmp(h, key, cur->key) > 0) {
 		cur = cur->next;
+		count++;
+		if (count == 10000) {
+			touch_softlockup_watchdog();
+			count = 0;
+		}
+	}
 
 	if (cur == NULL || (h->keycmp(h, key, cur->key) != 0))
 		return NULL;
author
#2 lore raw

PATCH 1/1 conditionally reschedule while loading selinux policy.

On a slow machine (with debugging enabled), upgrading selinux policy may take a considerable amount of time. Long enough that the softlockup detector gets triggered.

The backtrace looks like this..

Quoted 21 lines
BUG: soft lockup - CPU#2 stuck for 23s! [load_policy:19045]
Call Trace:
 [<ffffffff81221ddf>] symcmp+0xf/0x20
 [<ffffffff81221c27>] hashtab_search+0x47/0x80
 [<ffffffff8122e96c>] mls_convert_context+0xdc/0x1c0
 [<ffffffff812294e8>] convert_context+0x378/0x460
 [<ffffffff81229170>] ? security_context_to_sid_core+0x240/0x240
 [<ffffffff812221b5>] sidtab_map+0x45/0x80
 [<ffffffff8122bb9f>] security_load_policy+0x3ff/0x580
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff810786dd>] ? sched_clock_local+0x1d/0x80
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103096a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff810786dd>] ? sched_clock_local+0x1d/0x80
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103096a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff81534ddc>] ? retint_restore_args+0xe/0xe
 [<ffffffff8109c82d>] ? trace_hardirqs_on_caller+0xfd/0x1c0
 [<ffffffff81279a2e>] ? trace_hardirqs_on_thunk+0x3a/0x3f
 [<ffffffff810d28a8>] ? rcu_irq_exit+0x68/0xb0
Quoted 8 lines
 [<ffffffff81534ddc>] ? retint_restore_args+0xe/0xe
 [<ffffffff8121e947>] sel_write_load+0xa7/0x770
 [<ffffffff81139633>] ? vfs_write+0x1c3/0x200
 [<ffffffff81210e8e>] ? security_file_permission+0x1e/0xa0
 [<ffffffff8113952b>] vfs_write+0xbb/0x200
 [<ffffffff811581c7>] ? fget_light+0x397/0x4b0
 [<ffffffff81139c27>] SyS_write+0x47/0xa0
 [<ffffffff8153bde4>] tracesys+0xdd/0xe2

Stephen Smalley suggested:

Quoted 2 lines

Maybe put a cond_resched() within the ebitmap_for_each_positive_bit() loop in mls_convert_context()?

That seems to do the trick. Tested by downgrading and re-upgrading selinux-policy-targeted.

security/selinux/ss/mls.c +2 -0
index 40de8d3..9ef8e51 100644
@@ -500,6 +500,8 @@ int mls_convert_context(struct policydb *oldp,
 			rc = ebitmap_set_bit(&bitmap, catdatum->value - 1, 1);
 			if (rc)
 				return rc;
+
+			cond_resched();
 		}
 		ebitmap_destroy(&c->range.level[l].cat);
 		c->range.level[l].cat = bitmap;
Acked-by on v1 1/1
#3 lore raw

Re: [PATCH] conditionally reschedule while loading selinux policy.

Quoted 36 lines · Dave Jones

On a slow machine (with debugging enabled), upgrading selinux policy may take a considerable amount of time. Long enough that the softlockup detector gets triggered.

The backtrace looks like this..

Quoted 21 lines
BUG: soft lockup - CPU#2 stuck for 23s! [load_policy:19045]
Call Trace:
 [<ffffffff81221ddf>] symcmp+0xf/0x20
 [<ffffffff81221c27>] hashtab_search+0x47/0x80
 [<ffffffff8122e96c>] mls_convert_context+0xdc/0x1c0
 [<ffffffff812294e8>] convert_context+0x378/0x460
 [<ffffffff81229170>] ? security_context_to_sid_core+0x240/0x240
 [<ffffffff812221b5>] sidtab_map+0x45/0x80
 [<ffffffff8122bb9f>] security_load_policy+0x3ff/0x580
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff810786dd>] ? sched_clock_local+0x1d/0x80
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103096a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff810786dd>] ? sched_clock_local+0x1d/0x80
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103096a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff810788a8>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff81534ddc>] ? retint_restore_args+0xe/0xe
 [<ffffffff8109c82d>] ? trace_hardirqs_on_caller+0xfd/0x1c0
 [<ffffffff81279a2e>] ? trace_hardirqs_on_thunk+0x3a/0x3f
 [<ffffffff810d28a8>] ? rcu_irq_exit+0x68/0xb0
Quoted 8 lines
 [<ffffffff81534ddc>] ? retint_restore_args+0xe/0xe
 [<ffffffff8121e947>] sel_write_load+0xa7/0x770
 [<ffffffff81139633>] ? vfs_write+0x1c3/0x200
 [<ffffffff81210e8e>] ? security_file_permission+0x1e/0xa0
 [<ffffffff8113952b>] vfs_write+0xbb/0x200
 [<ffffffff811581c7>] ? fget_light+0x397/0x4b0
 [<ffffffff81139c27>] SyS_write+0x47/0xa0
 [<ffffffff8153bde4>] tracesys+0xdd/0xe2
Quoted 8 lines

Stephen Smalley suggested:

Quoted 2 lines

Maybe put a cond_resched() within the ebitmap_for_each_positive_bit() loop in mls_convert_context()?

That seems to do the trick. Tested by downgrading and re-upgrading selinux-policy-targeted.

Quoted 16 lines
security/selinux/ss/mls.c +2 -0
index 40de8d3..9ef8e51 100644
@@ -500,6 +500,8 @@ int mls_convert_context(struct policydb *oldp,
 			rc = ebitmap_set_bit(&bitmap, catdatum->value - 1, 1);
 			if (rc)
 				return rc;
+
+			cond_resched();
 		}
 		ebitmap_destroy(&c->range.level[l].cat);
 		c->range.level[l].cat = bitmap;
author
#4 lore raw

Re: [PATCH] conditionally reschedule while loading selinux policy.

In reply to Dave Jones
Quoted 4 lines · Dave Jones

On a slow machine (with debugging enabled), upgrading selinux policy may take a considerable amount of time. Long enough that the softlockup detector gets triggered.

Quoted 11 lines

The backtrace looks like this..

Quoted 9 lines
BUG: soft lockup - CPU#2 stuck for 23s! [load_policy:19045]
Call Trace:
 [<ffffffff81221ddf>] symcmp+0xf/0x20
 [<ffffffff81221c27>] hashtab_search+0x47/0x80
 [<ffffffff8122e96c>] mls_convert_context+0xdc/0x1c0
 [<ffffffff812294e8>] convert_context+0x378/0x460
 [<ffffffff81229170>] ? security_context_to_sid_core+0x240/0x240
 [<ffffffff812221b5>] sidtab_map+0x45/0x80
 [<ffffffff8122bb9f>] security_load_policy+0x3ff/0x580

With that patch applied, the problem seems to have moved elsewhere..

BUG: soft lockup - CPU#3 stuck for 22s! [load_policy:8119]
 irq event stamp: 1590886
 hardirqs last  enabled at (1590885): [<ffffffff8152c3dd>] __slab_alloc.constprop.78+0x4c0/0x4d7
 hardirqs last disabled at (1590886): [<ffffffff8153e06a>] apic_timer_interrupt+0x6a/0x80
 softirqs last  enabled at (1590336): [<ffffffff810480d9>] __do_softirq+0x169/0x200
 softirqs last disabled at (1590331): [<ffffffff8104839d>] irq_exit+0x11d/0x140
 RIP: 0010:[<ffffffff81223182>]  [<ffffffff81223182>] hashtab_insert+0x62/0x110

Call Trace:
 [<ffffffff812283b5>] policydb_read+0xc25/0x1200
 [<ffffffff810b1639>] ? is_module_text_address+0x19/0x40
 [<ffffffff8122cefe>] security_load_policy+0x10e/0x580
 [<ffffffff81078878>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff810786ad>] ? sched_clock_local+0x1d/0x80
 [<ffffffff81078878>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff810786ad>] ? sched_clock_local+0x1d/0x80
 [<ffffffff81078878>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103098a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff81078878>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8153669c>] ? retint_restore_args+0xe/0xe
 [<ffffffff810a196d>] ? trace_hardirqs_on_caller+0xfd/0x1c0
 [<ffffffff8127b07e>] ? trace_hardirqs_on_thunk+0x3a/0x3f
 [<ffffffff810d2958>] ? rcu_irq_exit+0x68/0xb0
 [<ffffffff8153669c>] ? retint_restore_args+0xe/0xe
 [<ffffffff8121ff97>] sel_write_load+0xa7/0x770
 [<ffffffff811394c3>] ? vfs_write+0x1c3/0x200
 [<ffffffff812124de>] ? security_file_permission+0x1e/0xa0
 [<ffffffff811393bb>] vfs_write+0xbb/0x200
 [<ffffffff81158337>] ? fget_light+0x397/0x4b0
 [<ffffffff81139ab7>] SyS_write+0x47/0xa0
 [<ffffffff8153d634>] tracesys+0xdd/0xe2

We're holding a bunch of locks here, so we can't just cond_resched. Thoughts ?

Dave

#5 lore raw

Re: [PATCH] conditionally reschedule while loading selinux policy.

Quoted 51 lines · Dave Jones
Quoted 4 lines · Dave Jones

On a slow machine (with debugging enabled), upgrading selinux policy may take a considerable amount of time. Long enough that the softlockup detector gets triggered.

Quoted 11 lines

The backtrace looks like this..

Quoted 9 lines
BUG: soft lockup - CPU#2 stuck for 23s! [load_policy:19045]
Call Trace:
 [<ffffffff81221ddf>] symcmp+0xf/0x20
 [<ffffffff81221c27>] hashtab_search+0x47/0x80
 [<ffffffff8122e96c>] mls_convert_context+0xdc/0x1c0
 [<ffffffff812294e8>] convert_context+0x378/0x460
 [<ffffffff81229170>] ? security_context_to_sid_core+0x240/0x240
 [<ffffffff812221b5>] sidtab_map+0x45/0x80
 [<ffffffff8122bb9f>] security_load_policy+0x3ff/0x580

With that patch applied, the problem seems to have moved elsewhere..

BUG: soft lockup - CPU#3 stuck for 22s! [load_policy:8119]
 irq event stamp: 1590886
 hardirqs last  enabled at (1590885): [<ffffffff8152c3dd>] __slab_alloc.constprop.78+0x4c0/0x4d7
 hardirqs last disabled at (1590886): [<ffffffff8153e06a>] apic_timer_interrupt+0x6a/0x80
 softirqs last  enabled at (1590336): [<ffffffff810480d9>] __do_softirq+0x169/0x200
 softirqs last disabled at (1590331): [<ffffffff8104839d>] irq_exit+0x11d/0x140
 RIP: 0010:[<ffffffff81223182>]  [<ffffffff81223182>] hashtab_insert+0x62/0x110

Call Trace:
 [<ffffffff812283b5>] policydb_read+0xc25/0x1200
 [<ffffffff810b1639>] ? is_module_text_address+0x19/0x40
 [<ffffffff8122cefe>] security_load_policy+0x10e/0x580
 [<ffffffff81078878>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff810786ad>] ? sched_clock_local+0x1d/0x80
 [<ffffffff81078878>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff810786ad>] ? sched_clock_local+0x1d/0x80
 [<ffffffff81078878>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8103098a>] ? __change_page_attr_set_clr+0x82a/0xa50
 [<ffffffff81078878>] ? sched_clock_cpu+0xa8/0x100
 [<ffffffff8153669c>] ? retint_restore_args+0xe/0xe
 [<ffffffff810a196d>] ? trace_hardirqs_on_caller+0xfd/0x1c0
 [<ffffffff8127b07e>] ? trace_hardirqs_on_thunk+0x3a/0x3f
 [<ffffffff810d2958>] ? rcu_irq_exit+0x68/0xb0
 [<ffffffff8153669c>] ? retint_restore_args+0xe/0xe
 [<ffffffff8121ff97>] sel_write_load+0xa7/0x770
 [<ffffffff811394c3>] ? vfs_write+0x1c3/0x200
 [<ffffffff812124de>] ? security_file_permission+0x1e/0xa0
 [<ffffffff811393bb>] vfs_write+0xbb/0x200
 [<ffffffff81158337>] ? fget_light+0x397/0x4b0
 [<ffffffff81139ab7>] SyS_write+0x47/0xa0
 [<ffffffff8153d634>] tracesys+0xdd/0xe2
Quoted 1 line

We're holding a bunch of locks here, so we can't just cond_resched. Thoughts ?

Sorry, what locks are we holding there? You ought to be able to do a cond_resched() anywhere during policydb_read() AFAIK; it is loading the policy into a new structure that isn't being accessed by anything else yet and the policy_rwlock is only held by security_load_policy after calling policydb_read and only to switch it into place as the active policydb.

author
#6 lore raw

Re: [PATCH] conditionally reschedule while loading selinux policy.

Quoted 13 lines · Stephen Smalley
Quoted 12 lines

With that patch applied, the problem seems to have moved elsewhere..

BUG: soft lockup - CPU#3 stuck for 22s! [load_policy:8119]
 irq event stamp: 1590886
 hardirqs last  enabled at (1590885): [<ffffffff8152c3dd>] __slab_alloc.constprop.78+0x4c0/0x4d7
 hardirqs last disabled at (1590886): [<ffffffff8153e06a>] apic_timer_interrupt+0x6a/0x80
 softirqs last  enabled at (1590336): [<ffffffff810480d9>] __do_softirq+0x169/0x200
 softirqs last disabled at (1590331): [<ffffffff8104839d>] irq_exit+0x11d/0x140
 RIP: 0010:[<ffffffff81223182>]  [<ffffffff81223182>] hashtab_insert+0x62/0x110

Call Trace:
 [<ffffffff812283b5>] policydb_read+0xc25/0x1200

...

Quoted 8 lines
Quoted 1 line

We're holding a bunch of locks here, so we can't just cond_resched. Thoughts ?

Sorry, what locks are we holding there? You ought to be able to do a cond_resched() anywhere during policydb_read() AFAIK; it is loading the policy into a new structure that isn't being accessed by anything else yet and the policy_rwlock is only held by security_load_policy after calling policydb_read and only to switch it into place as the active policydb.

Hmm, I thought I had tried this already, and got a lot of spew, but it turns out for some reason I had previously patched hashtab_search instead.

I'll try running with this for a while..

Dave

security/selinux/ss/hashtab.c +3 -0
index 933e735..2cc4961 100644
@@ -6,6 +6,7 @@
 #include <linux/kernel.h>
 #include <linux/slab.h>
 #include <linux/errno.h>
+#include <linux/sched.h>
 #include "hashtab.h"
 
 struct hashtab *hashtab_create(u32 (*hash_value)(struct hashtab *h, const void *key),
@@ -40,6 +41,8 @@ int hashtab_insert(struct hashtab *h, void *key, void *datum)
 	u32 hvalue;
 	struct hashtab_node *prev, *cur, *newnode;
 
+	cond_resched();
+
 	if (!h || h->nel == HASHTAB_MAX_NODES)
 		return -EINVAL;
 
maintainer
#7 lore raw

Re: [PATCH] conditionally reschedule while loading selinux policy.

Quoted 29 lines · Dave Jones
Quoted 21 lines · Stephen Smalley
Quoted 15 lines

With that patch applied, the problem seems to have moved elsewhere..

BUG: soft lockup - CPU#3 stuck for 22s! [load_policy:8119]

 irq event stamp: 1590886
 hardirqs last  enabled at (1590885): [<ffffffff8152c3dd>]
 __slab_alloc.constprop.78+0x4c0/0x4d7 hardirqs last disabled at
 (1590886): [<ffffffff8153e06a>] apic_timer_interrupt+0x6a/0x80
 softirqs last  enabled at (1590336): [<ffffffff810480d9>]
 __do_softirq+0x169/0x200 softirqs last disabled at (1590331):
 [<ffffffff8104839d>] irq_exit+0x11d/0x140 RIP:
 0010:[<ffffffff81223182>]  [<ffffffff81223182>]
 hashtab_insert+0x62/0x110 > > 
Call Trace:
 [<ffffffff812283b5>] policydb_read+0xc25/0x1200

...

Quoted 2 lines

We're holding a bunch of locks here, so we can't just cond_resched. Thoughts ?

Quoted 6 lines

Sorry, what locks are we holding there? You ought to be able to do a cond_resched() anywhere during policydb_read() AFAIK; it is loading the policy into a new structure that isn't being accessed by anything else yet and the policy_rwlock is only held by security_load_policy after calling policydb_read and only to switch it into place as the active policydb.

Quoted 4 lines

Hmm, I thought I had tried this already, and got a lot of spew, but it turns out for some reason I had previously patched hashtab_search instead.

I'll try running with this for a while..

Hi Dave,

Just checking to see if this patch solved your problem ... ?

Quoted 20 lines
security/selinux/ss/hashtab.c +1 -0
index 933e735..2cc4961 100644
@@ -6,6 +6,7 @@
 #include <linux/kernel.h>
 #include <linux/slab.h>
 #include <linux/errno.h>
+#include <linux/sched.h>
 #include "hashtab.h"
 
 struct hashtab *hashtab_create(u32 (*hash_value)(struct hashtab *h, const

void *key), @@ -40,6 +41,8 @@ int hashtab_insert(struct hashtab *h, void *key, void *datum) u32 hvalue;

security/selinux/ss/hashtab.c +2 -0
 	struct hashtab_node *prev, *cur, *newnode;
 
+	cond_resched();
+
 	if (!h || h->nel == HASHTAB_MAX_NODES)
 		return -EINVAL;
Signature · 8 lines
paul moore
www.paul-moore.com


--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to [email protected] with
the words "unsubscribe selinux" without quotes as the message.
author
#8 lore raw

Re: [PATCH] conditionally reschedule while loading selinux policy.

Quoted 36 lines · Paul Moore
Quoted 10 lines · Dave Jones
Quoted 8 lines · Stephen Smalley
Quoted 1 line

With that patch applied, the problem seems to have moved elsewhere..

Sorry, what locks are we holding there? You ought to be able to do a cond_resched() anywhere during policydb_read() AFAIK; it is loading the policy into a new structure that isn't being accessed by anything else yet and the policy_rwlock is only held by security_load_policy after calling policydb_read and only to switch it into place as the active policydb.

Quoted 4 lines

Hmm, I thought I had tried this already, and got a lot of spew, but it turns out for some reason I had previously patched hashtab_search instead.

I'll try running with this for a while..

Hi Dave,

Just checking to see if this patch solved your problem ... ?

Quoted 16 lines
security/selinux/ss/hashtab.c +1 -0
index 933e735..2cc4961 100644
@@ -6,6 +6,7 @@
 #include <linux/kernel.h>
 #include <linux/slab.h>
 #include <linux/errno.h>
+#include <linux/sched.h>
 #include "hashtab.h"
 
 struct hashtab *hashtab_create(u32 (*hash_value)(struct hashtab *h, const

void *key), @@ -40,6 +41,8 @@ int hashtab_insert(struct hashtab *h, void *key, void *datum) u32 hvalue;
struct hashtab_node *prev, *cur, *newnode;

Quoted 4 lines

security/selinux/ss/hashtab.c int hashtab_insert(struct hashtab *h, void *key, void *datum) · line 47

Quoted 4 lines
security/selinux/ss/hashtab.c +2 -0
+	cond_resched();
+
 	if (!h || h->nel == HASHTAB_MAX_NODES)
 		return -EINVAL;

I couldn't get a backtrace when I downgraded, and reupgraded my policy, but that said, I didn't when I tested that way on my first patch either. It wasn't until I got a newer policy that I saw the 2nd spew. I was planning on sitting on this until the next policy update just to confirm.

Dave

maintainer
#9 lore raw

Re: [PATCH] conditionally reschedule while loading selinux policy.

Quoted 5 lines · Dave Jones

I couldn't get a backtrace when I downgraded, and reupgraded my policy, but that said, I didn't when I tested that way on my first patch either. It wasn't until I got a newer policy that I saw the 2nd spew. I was planning on sitting on this until the next policy update just to confirm.

Okay, no problem. Let me know how it goes.

Thanks,
-Paul

Signature · 8 lines
paul moore
www.paul-moore.com


--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to [email protected] with
the words "unsubscribe selinux" without quotes as the message.