Re: [Qemu-devel] [PATCH 2/2] hw/pci-host/x86: extend the 64-bit PCI hole relative to the fw-assigned base

[Eric Blake]

On 9/26/18 3:35 PM, Laszlo Ersek wrote:
> (+Eric)

> This too sounds useful. AIUI, ftruncate() is neither forbidden, nor
> required, to allocate filesystem extents when increasing the size of a
> file. Using one smaller regular temporary file as the common foundation
> for multiple "memory-backend-file" objects will save space on the fs if
> ftruncate() happens to allocate extents.

Correct. Most likely, ftruncate() extends by creating a hole if the file 
system supports sparse files, but POSIX has no way to guarantee holes 
(it merely states that both ftruncate() and write() beyond EOF are the 
two POSIX interfaces that may create holes - but that it is up to the 
implementation whether holes actually get created, then fstat() and 
lseek(SEEK_HOLE) can inform you whether holes were actually created [the 
former when st_blocks < st_size]). posix_fallocate() exists to force 
allocation, and Linux fallocate() offers even more fine-tuning when you 
need specific hole or non-hole behaviors.

--
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org