Revision ac7cf246dfdbec3d8fed296c7bf30e16f5099dac authored by Joseph Qi on 25 March 2016, 21:21:26 UTC, committed by Linus Torvalds on 25 March 2016, 23:37:42 UTC
There is a race window between dlmconvert_remote and dlm_move_lockres_to_recovery_list, which will cause a lock with OCFS2_LOCK_BUSY in grant list, thus system hangs. dlmconvert_remote { spin_lock(&res->spinlock); list_move_tail(&lock->list, &res->converting); lock->convert_pending = 1; spin_unlock(&res->spinlock); status = dlm_send_remote_convert_request(); >>>>>> race window, master has queued ast and return DLM_NORMAL, and then down before sending ast. this node detects master down and calls dlm_move_lockres_to_recovery_list, which will revert the lock to grant list. Then OCFS2_LOCK_BUSY won't be cleared as new master won't send ast any more because it thinks already be authorized. spin_lock(&res->spinlock); lock->convert_pending = 0; if (status != DLM_NORMAL) dlm_revert_pending_convert(res, lock); spin_unlock(&res->spinlock); } In this case, check if res->state has DLM_LOCK_RES_RECOVERING bit set (res is still in recovering) or res master changed (new master has finished recovery), reset the status to DLM_RECOVERING, then it will retry convert. Signed-off-by: Joseph Qi <joseph.qi@huawei.com> Reported-by: Yiwen Jiang <jiangyiwen@huawei.com> Reviewed-by: Junxiao Bi <junxiao.bi@oracle.com> Cc: Mark Fasheh <mfasheh@suse.de> Cc: Joel Becker <jlbec@evilplan.org> Cc: Tariq Saeed <tariq.x.saeed@oracle.com> Cc: Junxiao Bi <junxiao.bi@oracle.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent 2888868
mtd_test.c
#define pr_fmt(fmt) "mtd_test: " fmt
#include <linux/module.h>
#include <linux/sched.h>
#include <linux/printk.h>
#include "mtd_test.h"
int mtdtest_erase_eraseblock(struct mtd_info *mtd, unsigned int ebnum)
{
int err;
struct erase_info ei;
loff_t addr = (loff_t)ebnum * mtd->erasesize;
memset(&ei, 0, sizeof(struct erase_info));
ei.mtd = mtd;
ei.addr = addr;
ei.len = mtd->erasesize;
err = mtd_erase(mtd, &ei);
if (err) {
pr_info("error %d while erasing EB %d\n", err, ebnum);
return err;
}
if (ei.state == MTD_ERASE_FAILED) {
pr_info("some erase error occurred at EB %d\n", ebnum);
return -EIO;
}
return 0;
}
static int is_block_bad(struct mtd_info *mtd, unsigned int ebnum)
{
int ret;
loff_t addr = (loff_t)ebnum * mtd->erasesize;
ret = mtd_block_isbad(mtd, addr);
if (ret)
pr_info("block %d is bad\n", ebnum);
return ret;
}
int mtdtest_scan_for_bad_eraseblocks(struct mtd_info *mtd, unsigned char *bbt,
unsigned int eb, int ebcnt)
{
int i, bad = 0;
if (!mtd_can_have_bb(mtd))
return 0;
pr_info("scanning for bad eraseblocks\n");
for (i = 0; i < ebcnt; ++i) {
bbt[i] = is_block_bad(mtd, eb + i) ? 1 : 0;
if (bbt[i])
bad += 1;
cond_resched();
}
pr_info("scanned %d eraseblocks, %d are bad\n", i, bad);
return 0;
}
int mtdtest_erase_good_eraseblocks(struct mtd_info *mtd, unsigned char *bbt,
unsigned int eb, int ebcnt)
{
int err;
unsigned int i;
for (i = 0; i < ebcnt; ++i) {
if (bbt[i])
continue;
err = mtdtest_erase_eraseblock(mtd, eb + i);
if (err)
return err;
cond_resched();
}
return 0;
}
int mtdtest_read(struct mtd_info *mtd, loff_t addr, size_t size, void *buf)
{
size_t read;
int err;
err = mtd_read(mtd, addr, size, &read, buf);
/* Ignore corrected ECC errors */
if (mtd_is_bitflip(err))
err = 0;
if (!err && read != size)
err = -EIO;
if (err)
pr_err("error: read failed at %#llx\n", addr);
return err;
}
int mtdtest_write(struct mtd_info *mtd, loff_t addr, size_t size,
const void *buf)
{
size_t written;
int err;
err = mtd_write(mtd, addr, size, &written, buf);
if (!err && written != size)
err = -EIO;
if (err)
pr_err("error: write failed at %#llx\n", addr);
return err;
}
![swh spinner](/static/img/swh-spinner.gif)
Computing file changes ...