On 15.04.25 18:22, Eric Blake wrote:
On Tue, Apr 15, 2025 at 03:37:39PM +0300, Vladimir Sementsov-Ogievskiy wrote:
On 14.04.25 21:12, Eric Blake wrote:
On Mon, Apr 14, 2025 at 08:05:21PM +0300, Vladimir Sementsov-Ogievskiy wrote:
On 11.04.25 04:04, Eric Blake wrote:
The 'want_zero' parameter
On Tue, Apr 15, 2025 at 03:37:39PM +0300, Vladimir Sementsov-Ogievskiy wrote:
> On 14.04.25 21:12, Eric Blake wrote:
> > On Mon, Apr 14, 2025 at 08:05:21PM +0300, Vladimir Sementsov-Ogievskiy
> > wrote:
> > > On 11.04.25 04:04, Eric Blake wrote:
> > > > The 'want_zero' parameter to raw_co_block_st
On 14.04.25 21:12, Eric Blake wrote:
On Mon, Apr 14, 2025 at 08:05:21PM +0300, Vladimir Sementsov-Ogievskiy wrote:
On 11.04.25 04:04, Eric Blake wrote:
The 'want_zero' parameter to raw_co_block_status() was added so that
we can avoid potentially time-consuming lseek(SEEK_DATA) calls
throughout
On Mon, Apr 14, 2025 at 08:05:21PM +0300, Vladimir Sementsov-Ogievskiy wrote:
> On 11.04.25 04:04, Eric Blake wrote:
> > The 'want_zero' parameter to raw_co_block_status() was added so that
> > we can avoid potentially time-consuming lseek(SEEK_DATA) calls
> > throughout the file (working around po
On 11.04.25 04:04, Eric Blake wrote:
The 'want_zero' parameter to raw_co_block_status() was added so that
we can avoid potentially time-consuming lseek(SEEK_DATA) calls
throughout the file (working around poor filesystems that have O(n)
rather than O(1) extent probing). But when it comes to lear
The 'want_zero' parameter to raw_co_block_status() was added so that
we can avoid potentially time-consuming lseek(SEEK_DATA) calls
throughout the file (working around poor filesystems that have O(n)
rather than O(1) extent probing). But when it comes to learning if a
file is completely sparse (fo
