On 8/5/20 8:32 AM, Richard W.M. Jones wrote:
On Wed, Aug 05, 2020 at 08:12:26AM -0500, Eric Blake wrote:
> On 8/5/20 7:48 AM, Richard W.M. Jones wrote:
>> See comment in code and
>>
https://www.redhat.com/archives/libguestfs/2020-August/msg00023.html
>> ---
>> plugins/vddk/vddk.c | 32 ++++++++++++++++++++++++++++----
>> 1 file changed, 28 insertions(+), 4 deletions(-)
>
> I'll have to trust your testing, but code-wise, this looks safe.
I'm in a bit two minds about this one. It appears safe. OTOH it's
somewhat risky in theory, and it doesn't give us very much performance
improvement.
I agree that we are NOT strictly following the VDDK documentation, in
that this risks 2 different threads calling VDiskLib_Open (even though
we have serialized it so that only one open is ever active at a time).
So it is indeed risky.
Let's discuss alternate implementations with Nir in the other thread.
I'll respond there.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3226
Virtualization:
qemu.org |
libvirt.org