On 2019/4/23 3:08 下午, Hannes Reinecke wrote:
> On 4/19/19 6:05 PM, Coly Li wrote:
>> If a bcache device is in dirty state and its cache set is not
>> registered, this bcache deivce will not appear in /dev/bcache<N>,
>> and there is no way to stop it or remove the bcache kernel module.
>>
>> This is an as-designed behavior, but sometimes people has to reboot
>> whole system to release or stop the pending backing device.
>>
>> This sysfs interface may remove such pending bcache devices when
>> write anything into the sysfs file manually.
>>
>> Signed-off-by: Coly Li <colyli@xxxxxxx>
>> ---
>> drivers/md/bcache/super.c | 55
>> +++++++++++++++++++++++++++++++++++++++++++++++
>> 1 file changed, 55 insertions(+)
>>
>> diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
>> index 9b41e0b62cc0..e988e46a6479 100644
>> --- a/drivers/md/bcache/super.c
>> +++ b/drivers/md/bcache/super.c
>> @@ -2246,9 +2246,13 @@ static int register_cache(struct cache_sb *sb,
>> struct page *sb_page,
>> static ssize_t register_bcache(struct kobject *k, struct
>> kobj_attribute *attr,
>> const char *buffer, size_t size);
>> +static ssize_t bch_pending_bdevs_cleanup(struct kobject *k,
>> + struct kobj_attribute *attr,
>> + const char *buffer, size_t size);
>> kobj_attribute_write(register, register_bcache);
>> kobj_attribute_write(register_quiet, register_bcache);
>> +kobj_attribute_write(pendings_cleanup, bch_pending_bdevs_cleanup);
>> static bool bch_is_open_backing(struct block_device *bdev)
>> {
>> @@ -2373,6 +2377,56 @@ static ssize_t register_bcache(struct kobject
>> *k, struct kobj_attribute *attr,
>> goto out;
>> }
>> +
>> +struct pdev {
>> + struct list_head list;
>> + struct cached_dev *dc;
>> +};
>> +
>> +static ssize_t bch_pending_bdevs_cleanup(struct kobject *k,
>> + struct kobj_attribute *attr,
>> + const char *buffer,
>> + size_t size)
>> +{
>> + LIST_HEAD(pending_devs);
>> + ssize_t ret = size;
>> + struct cached_dev *dc, *tdc;
>> + struct pdev *pdev, *tpdev;
>> + struct cache_set *c, *tc;
>> +
>> + mutex_lock(&bch_register_lock);
>> + list_for_each_entry_safe(dc, tdc, &uncached_devices, list) {
>> + pdev = kmalloc(sizeof(struct pdev), GFP_KERNEL);
>> + if (!pdev)
>> + break;
>> + pdev->dc = dc;
>> + list_add(&pdev->list, &pending_devs);
>> + }
>> +
>> + list_for_each_entry_safe(pdev, tpdev, &pending_devs, list) {
>> + list_for_each_entry_safe(c, tc, &bch_cache_sets, list) {
>> + char *pdev_set_uuid = pdev->dc->sb.set_uuid;
>> + char *set_uuid = c->sb.uuid;
>> +
>> + if (!memcmp(pdev_set_uuid, set_uuid, 16)) {
>> + list_del(&pdev->list);
>> + kfree(pdev);
>> + break;
>> + }
>> + }
>> + }
>> + mutex_unlock(&bch_register_lock);
>> +
>> + list_for_each_entry_safe(pdev, tpdev, &pending_devs, list) {
>> + pr_info("delete pdev %p", pdev);
>> + list_del(&pdev->list);
>> + bcache_device_stop(&pdev->dc->disk);
>> + kfree(pdev);
>> + }
>> +
>> + return ret;
>> +}
>> +
>> static int bcache_reboot(struct notifier_block *n, unsigned long
>> code, void *x)
>> {
>> if (code == SYS_DOWN ||
>> @@ -2483,6 +2537,7 @@ static int __init bcache_init(void)
>> static const struct attribute *files[] = {
>> &ksysfs_register.attr,
>> &ksysfs_register_quiet.attr,
>> + &ksysfs_pendings_cleanup.attr,
>> NULL
>> };
>>
> _Actually_ I would like it better if the bcache device would be present
> in sysfs for these cases, too, albeit in a disabled state.
> That would allow us to remove the device like normal and we wouldn't
> need to worry about yet another interface.
Hi Hannes,
I see. The awkward condition is,
- If only show up the bcache device in /sys/block/bcache<?>/ directory
but not show up in /dev/bcache<?> node, it might be more confused for users.
- If show up both bcache devices in /sys/block/bcache<?> and
/dev/bcache<?>, it means such device can be read/written, but it is
staled (there are dirty data on cache device), which may corrupt
existing data.
So I choose to not change current code behavior and just add a sysfs
interface.
--
Coly Li
