x86/proc: fix /proc/cpuinfo cpu offline bug

In my test, I found that if a cpu has been offline, the next cpus may not be shown in the /proc/cpuinfo.

if one read() cannot consume the whole /proc/cpuinfo file, c_start() will be called again in the next read() calls. And *pos has been increased by 1 by the caller(seq_read()). if this time the cpu#*pos is offline, c_start() will return NULL, and the next cpus can not be shown.

this fix uses next_cpu_nr(*pos - 1, cpu_online_map) to search the next unshown cpu.

the way to reproduce this bug:
(cpu#0 cpu#1 cpu#2 must be online before test)

step 1) offline cpu#1
step 2) dd ibs=2 if=/proc/cpuinfo

now, only cpu#0 's cpuinfo is shown.
cpu#2 and cpu#3 .... cannot be shown in /proc/cpuinfo it's a bug.

---

index a26c480..01b1244 100644
@@ -160,14 +160,16 @@ static void *c_start(struct seq_file *m, loff_t *pos)
 {
 	if (*pos == 0)	/* just in case, cpu 0 is not the first */
 		*pos = first_cpu(cpu_online_map);
-	if ((*pos) < nr_cpu_ids && cpu_online(*pos))
+	else
+		*pos = next_cpu_nr(*pos - 1, cpu_online_map);
+	if ((*pos) < nr_cpu_ids)
 		return &cpu_data(*pos);
 	return NULL;
 }
 
 static void *c_next(struct seq_file *m, void *v, loff_t *pos)
 {
-	*pos = next_cpu(*pos, cpu_online_map);
+	(*pos)++;
 	return c_start(m, pos);
 }
 

fix is already upstream - see below.

Ingo

-------------------->

Gitweb:     http://git.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=bc8bcc79ea4203c7d04309f1307ab88c86f0b0cf
Commit:     bc8bcc79ea4203c7d04309f1307ab88c86f0b0cf
Parent:     35af28219e684a36cc8b1ff456c370ce22be157d
Author:     Lai Jiangshan <[email protected]>
AuthorDate: Wed Oct 22 12:42:30 2008 +0800
Committer:  Ingo Molnar <[email protected]>
CommitDate: Wed Oct 22 14:29:37 2008 +0200

x86/proc: fix /proc/cpuinfo cpu offline bug

Impact: fix missing CPUs in /proc/cpuinfo after CPU hotunplug/hotreplug

In my test, I found that if a cpu has been offline, the next cpus may not be shown in the /proc/cpuinfo.

if one read() cannot consume the whole /proc/cpuinfo, c_start() will be called again in the next read() calls. And *pos has been increased by 1 by the caller(seq_read()). if this time the cpu#*pos is offline, c_start() will return NULL, and the next cpus can not be shown.

this fix use next_cpu_nr(*pos - 1, cpu_online_map) to search the next unshown cpu.

the most easy way to reproduce this bug:
1) offline cpu#1 (cpu#0 is online)
2) dd ibs=2 if=/proc/cpuinfo
the result is that only cpu#0 is shown. cpu#2 and cpu#3 .... cannot be shown in /proc/cpuinfo it's bug.

---
 arch/x86/kernel/cpu/proc.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

index a26c480..01b1244 100644
@@ -160,14 +160,16 @@ static void *c_start(struct seq_file *m, loff_t *pos)
 {
 	if (*pos == 0)	/* just in case, cpu 0 is not the first */
 		*pos = first_cpu(cpu_online_map);
-	if ((*pos) < nr_cpu_ids && cpu_online(*pos))
+	else
+		*pos = next_cpu_nr(*pos - 1, cpu_online_map);
+	if ((*pos) < nr_cpu_ids)
 		return &cpu_data(*pos);
 	return NULL;
 }
 
 static void *c_next(struct seq_file *m, void *v, loff_t *pos)
 {
-	*pos = next_cpu(*pos, cpu_online_map);
+	(*pos)++;
 	return c_start(m, pos);
 }