Re: [PATCH v3] mkfs: Break block discard into chunks of 2 GB

[Eric Sandeen <sandeen@sandeen.net>]

On 12/9/19 5:34 PM, Darrick J. Wong wrote:
> On Mon, Dec 09, 2019 at 04:00:17PM -0600, Eric Sandeen wrote:
>> On 11/28/19 12:21 AM, Pavel Reichl wrote:
>>> Some users are not happy about the BLKDISCARD taking too long and at the same
>>> time not being informed about that - so they think that the command actually
>>> hung.
>>>
>>> This commit changes code so that progress reporting is possible and also typing
>>> the ^C will cancel the ongoing BLKDISCARD.
>> Ok I'm going to nitpick this just a little bit more...
>>
>>> Signed-off-by: Pavel Reichl <preichl@redhat.com>
>>> ---
>>>  mkfs/xfs_mkfs.c | 50 ++++++++++++++++++++++++++++++++++++-------------
>>>  1 file changed, 37 insertions(+), 13 deletions(-)
>>>
>>> diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
>>> index 18338a61..0defadb2 100644
>>> --- a/mkfs/xfs_mkfs.c
>>> +++ b/mkfs/xfs_mkfs.c
>>> @@ -1240,17 +1240,40 @@ done:
>>>  }
>>>  
>>>  static void
>>> -discard_blocks(dev_t dev, uint64_t nsectors)
>>> +discard_blocks(dev_t dev, uint64_t nsectors, int quiet)
>>>  {
>>> -	int fd;
>>> +	int		fd;
>>> +	uint64_t	offset = 0;
>>> +	/* Discard the device 2G at a time */
>>> +	const uint64_t	step = 2ULL << 30;
>>> +	const uint64_t	count = BBTOB(nsectors);
>>>  
>>> -	/*
>>> -	 * We intentionally ignore errors from the discard ioctl.  It is
>>> -	 * not necessary for the mkfs functionality but just an optimization.
>>> -	 */
>>>  	fd = libxfs_device_to_fd(dev);
>>> -	if (fd > 0)
>>> -		platform_discard_blocks(fd, 0, nsectors << 9);
>>> +	if (fd <= 0)
>>> +		return;
>>> +	if (!quiet) {
>>> +		printf("Discarding blocks...\n");
>>> +		printf("...\n");
>>> +	}
>> Let's change this output so that it prints only a single line, i.e.
>>
>> +	if (!quiet) {
>> +		printf("Discarding blocks... ");
> This needs fflush(stdout); here to force the message out of stdio output
> buffering.
> 
> --D

I thought of that, then forgot.  ;)  Maybe we do need a formal v4 on the list.