From: Yu Kuai <yukuai3@xxxxxxxxxx> match_u64() is called inside ioc->lock, which causes smatch static checker warnings: block/blk-iocost.c:3211 ioc_qos_write() warn: sleeping in atomic context block/blk-iocost.c:3240 ioc_qos_write() warn: sleeping in atomic context Fix the problem by prase params before holding the lock. Fixes: 2c0647988433 ("blk-iocost: don't release 'ioc->lock' while updating params") Reported-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx> Signed-off-by: Yu Kuai <yukuai3@xxxxxxxxxx> --- block/blk-iocost.c | 172 +++++++++++++++++++++++++++++---------------- 1 file changed, 111 insertions(+), 61 deletions(-) diff --git a/block/blk-iocost.c b/block/blk-iocost.c index 2bfecc511dd9..27b41f3f1b07 100644 --- a/block/blk-iocost.c +++ b/block/blk-iocost.c @@ -3165,44 +3165,25 @@ static const match_table_t qos_tokens = { { NR_QOS_PARAMS, NULL }, }; -static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input, - size_t nbytes, loff_t off) -{ - struct block_device *bdev; - struct gendisk *disk; - struct ioc *ioc; +struct ioc_qos_params { u32 qos[NR_QOS_PARAMS]; - bool enable, user; - char *p; - int ret; - - bdev = blkcg_conf_open_bdev(&input); - if (IS_ERR(bdev)) - return PTR_ERR(bdev); - - disk = bdev->bd_disk; - if (!queue_is_mq(disk->queue)) { - ret = -EPERM; - goto out; - } - - ioc = q_to_ioc(disk->queue); - if (!ioc) { - ret = blk_iocost_init(disk); - if (ret) - goto out; - ioc = q_to_ioc(disk->queue); - } + bool enable; + bool user; + bool set_qos[NR_QOS_PARAMS]; + bool set_enable; + bool set_user; +}; - blk_mq_freeze_queue(disk->queue); - blk_mq_quiesce_queue(disk->queue); +static struct ioc_qos_params *ioc_qos_parse_params(char *input) +{ + struct ioc_qos_params *params; + char *p; + int ret = -EINVAL; - spin_lock_irq(&ioc->lock); - memcpy(qos, ioc->params.qos, sizeof(qos)); - enable = ioc->enabled; - user = ioc->user_qos_params; + params = kzalloc(sizeof(*params), GFP_KERNEL); + if (!params) + return ERR_PTR(-ENOMEM); - ret = -EINVAL; while ((p = strsep(&input, " \t\n"))) { substring_t args[MAX_OPT_ARGS]; char buf[32]; @@ -3218,19 +3199,21 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input, err = match_u64(&args[0], &v); if (err) { ret = err; - goto out_unlock; + goto out_err; } - enable = v; + params->enable = v; + params->set_enable = true; continue; case QOS_CTRL: match_strlcpy(buf, &args[0], sizeof(buf)); if (!strcmp(buf, "auto")) - user = false; + params->user = false; else if (!strcmp(buf, "user")) - user = true; + params->user = true; else - goto out_unlock; + goto out_err; + params->set_user = true; continue; } @@ -3240,62 +3223,128 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input, case QOS_WPPM: if (match_strlcpy(buf, &args[0], sizeof(buf)) >= sizeof(buf)) - goto out_unlock; + goto out_err; if (cgroup_parse_float(buf, 2, &v)) - goto out_unlock; + goto out_err; if (v < 0 || v > 10000) - goto out_unlock; - qos[tok] = v * 100; + goto out_err; + params->qos[tok] = v * 100; break; case QOS_RLAT: case QOS_WLAT: err = match_u64(&args[0], &v); if (err) { ret = err; - goto out_unlock; + goto out_err; } - qos[tok] = v; + params->qos[tok] = v; break; case QOS_MIN: case QOS_MAX: if (match_strlcpy(buf, &args[0], sizeof(buf)) >= sizeof(buf)) - goto out_unlock; + goto out_err; if (cgroup_parse_float(buf, 2, &v)) - goto out_unlock; + goto out_err; if (v < 0) - goto out_unlock; - qos[tok] = clamp_t(s64, v * 100, + goto out_err; + params->qos[tok] = clamp_t(s64, v * 100, VRATE_MIN_PPM, VRATE_MAX_PPM); break; default: - goto out_unlock; + goto out_err; } - user = true; + params->user = true; + params->set_user = true; + params->set_qos[tok] = true; } - if (qos[QOS_MIN] > qos[QOS_MAX]) - goto out_unlock; + return params; + +out_err: + kfree(params); + return ERR_PTR(ret); +} + +static int ioc_qos_update_params(struct request_queue *q, struct ioc *ioc, + struct ioc_qos_params *params) +{ + int i; - if (enable) { - blk_stat_enable_accounting(disk->queue); - blk_queue_flag_set(QUEUE_FLAG_RQ_ALLOC_TIME, disk->queue); + for (i = 0; i < NR_QOS_PARAMS; ++i) + if (!params->set_qos[i]) + params->qos[i] = ioc->params.qos[i]; + if (params->qos[QOS_MIN] > params->qos[QOS_MAX]) + return -EINVAL; + + if (!params->set_enable) + params->enable = ioc->enabled; + if (params->enable) { + blk_stat_enable_accounting(q); + blk_queue_flag_set(QUEUE_FLAG_RQ_ALLOC_TIME, q); ioc->enabled = true; - wbt_disable_default(disk->queue); + wbt_disable_default(q); } else { - blk_queue_flag_clear(QUEUE_FLAG_RQ_ALLOC_TIME, disk->queue); + blk_queue_flag_clear(QUEUE_FLAG_RQ_ALLOC_TIME, q); ioc->enabled = false; - wbt_enable_default(disk->queue); + wbt_enable_default(q); } - if (user) { - memcpy(ioc->params.qos, qos, sizeof(qos)); + if (!params->set_user) + params->user = ioc->user_qos_params; + if (params->user) { + memcpy(ioc->params.qos, params->qos, sizeof(params->qos)); ioc->user_qos_params = true; } else { ioc->user_qos_params = false; } + return 0; +} + +static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input, + size_t nbytes, loff_t off) +{ + struct block_device *bdev; + struct gendisk *disk; + struct ioc *ioc; + struct ioc_qos_params *params; + int ret; + + bdev = blkcg_conf_open_bdev(&input); + if (IS_ERR(bdev)) + return PTR_ERR(bdev); + + disk = bdev->bd_disk; + if (!queue_is_mq(disk->queue)) { + ret = -EPERM; + goto out; + } + + ioc = q_to_ioc(disk->queue); + if (!ioc) { + ret = blk_iocost_init(disk); + if (ret) + goto out; + ioc = q_to_ioc(disk->queue); + } + + params = ioc_qos_parse_params(input); + if (IS_ERR(params)) { + ret = PTR_ERR(params); + goto out; + } + + blk_mq_freeze_queue(disk->queue); + blk_mq_quiesce_queue(disk->queue); + + spin_lock_irq(&ioc->lock); + + ret = ioc_qos_update_params(disk->queue, ioc, params); + if (ret) + goto out_unlock; + ioc_refresh_params(ioc, true); ret = nbytes; @@ -3303,6 +3352,7 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input, spin_unlock_irq(&ioc->lock); blk_mq_unquiesce_queue(disk->queue); blk_mq_unfreeze_queue(disk->queue); + kfree(params); out: blkdev_put_no_open(bdev); -- 2.31.1