On 4/11/19 5:44 AM, Song Liu wrote:
On Wed, Apr 10, 2019 at 2:10 AM Guoqing Jiang <gqjiang@xxxxxxxx> wrote:
The line[200] may not big enough to create raid0 with
lots of members, then the calltrace appears.
[ 42.081732] WARNING: CPU: 1 PID: 282 at lib/vsprintf.c:2334 vsnprintf+0x485/0x520
[ 42.082318] Modules linked in: raid0(E) md_mod(E) zram(E) configfs(E) dm_mod(E)
[ 42.082318] CPU: 1 PID: 282 Comm: mdadm Tainted: G E 5.1.0-rc4-5-default+ #30
[ 42.082318] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.0.0-prebuilt.qemu-project.org 04/01/2014
[ 42.082318] RIP: 0010:vsnprintf+0x485/0x520
[ 42.082318] Code: 08 48 8d 42 08 49 89 44 24 08 e9 1a fe ff ff 83 f8 2f 77 49 89 c2 83 c0 08 49 03 54 24 10 41 89 04 24 48 63 12 e9 02 fe ff ff <0f> 0b e9 26 fc ff ff 83 f8 2f 77 17 89 c2 83 c0 08 49 03 54 24 10
[ 42.082318] RSP: 0018:ffffc9000095fa88 EFLAGS: 00010292
[ 42.082318] RAX: 0000000000000000 RBX: 00000000000000cc RCX: ffffc9000095fae0
[ 42.082318] RDX: ffffffffa007f649 RSI: fffffffffffffffc RDI: ffffc9000095fc44
[ 42.082318] RBP: ffffc9000095fb30 R08: ffffc9000095fb58 R09: 000000008e97e7f4
[ 42.082318] R10: ffffc9000095fb78 R11: 0000000000000000 R12: 0000000000000000
[ 42.082318] R13: ffffc9000095fb78 R14: 000000000000001f R15: 0000000000000000
[ 42.090354] FS: 00007fc0b3cc4700(0000) GS:ffff88807b400000(0000) knlGS:0000000000000000
[ 42.090354] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 42.090354] CR2: 000000000121cd88 CR3: 000000007df96000 CR4: 00000000000006e0
[ 42.090354] Call Trace:
[ 42.090354] snprintf+0x49/0x60
[ 42.090354] ? disk_name+0x8b/0xa0
[ 42.090354] dump_zones+0x104/0x220 [raid0]
[ 42.090354] ? validate_chain.isra.43+0x6ec/0xad0
[ 42.090354] raid0_run+0x197/0x240 [raid0]
[ 42.090354] md_run+0x45a/0xae0 [md_mod]
[ 42.090354] ? security_capable+0x38/0x50
[ 42.090354] do_md_run+0xf/0xc0 [md_mod]
[ 42.090354] md_ioctl+0x1169/0x1b00 [md_mod]
[ 42.090354] ? cyc2ns_read_begin+0x20/0x50
[ 42.090354] ? sched_clock+0x5/0x10
[ 42.090354] blkdev_ioctl+0x506/0xa30
[ 42.090354] ? selinux_file_ioctl+0x190/0x230
[ 42.090354] block_ioctl+0x3f/0x50
[ 42.090354] do_vfs_ioctl+0xa5/0x6e0
[ 42.098359] ksys_ioctl+0x60/0x90
[ 42.098359] __x64_sys_ioctl+0x16/0x20
[ 42.098359] do_syscall_64+0x60/0x210
[ 42.098359] entry_SYSCALL_64_after_hwframe+0x49/0xbe
So revert it to use previous way.
Fixes: 766038846e875 ("md/raid0: replace printk() with pr_*()")
Signed-off-by: Guoqing Jiang <gqjiang@xxxxxxxx>
---
drivers/md/raid0.c | 11 +++++------
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
index f3fb5bb8c82a..edd17025faf4 100644
--- a/drivers/md/raid0.c
+++ b/drivers/md/raid0.c
@@ -63,14 +63,13 @@ static void dump_zones(struct mddev *mddev)
mdname(mddev),
conf->nr_strip_zones, conf->nr_strip_zones==1?"":"s");
for (j = 0; j < conf->nr_strip_zones; j++) {
- char line[200];
- int len = 0;
+ printk(KERN_INFO "md: zone%d=[", j);
for (k = 0; k < conf->strip_zone[j].nb_dev; k++)
- len += snprintf(line+len, 200-len, "%s%s", k?"/":"",
- bdevname(conf->devlist[j*raid_disks
- + k]->bdev, b));
- pr_debug("md: zone%d=[%s]\n", j, line);
+ printk(KERN_CONT "%s%s", k?"/":"",
+ bdevname(conf->devlist[j*raid_disks
+ + k]->bdev, b));
+ printk(KERN_CONT "]\n");
I think it is not recommended to use KERN_CONT here. How about we
trim the line like:
diff --git i/drivers/md/raid0.c w/drivers/md/raid0.c
index f3fb5bb8c82a..cb996aa9f02b 100644
--- i/drivers/md/raid0.c
+++ w/drivers/md/raid0.c
@@ -66,7 +66,7 @@ static void dump_zones(struct mddev *mddev)
char line[200];
int len = 0;
- for (k = 0; k < conf->strip_zone[j].nb_dev; k++)
+ for (k = 0; k < conf->strip_zone[j].nb_dev && len < 200; k++)
len += snprintf(line+len, 200-len, "%s%s", k?"/":"",
bdevname(conf->devlist[j*raid_disks
+ k]->bdev, b));
Then we can't get the all the member of zone with "pr_debug("md:
zone%d=[%s]\n", j, line)",
eg, if create raid0 with zram0-zram31, the above only can show:
zone0=[zram0/zram1/zram10/zram11/zram12/zram13/zram14/zram15/zram16/zram17/zram18/zram19/zram2/zram20/zram21/zram22/zram23/zram24/zram25/zram26/zram27/zram28/zram29/zram3/zram30/zram31/zram4/zram5/zram6/zra]
Thanks,
Guoqing