Now that we have a SpiceRefcountedChunk, we can make use of it to
make the code handling the agent data pipe item much simpler and
readable.
---
server/main_channel.c | 51
1 files changed, 13 insertions(+), 38 deletions(-)
diff --git a/server/ma
It's a SpiceChunk with an additional reference count so that the
chunk data can be freed when the chunk is no longer needed.
---
common/mem.c | 57 +
common/mem.h |7 +++
2 files changed, 64 insertions(+), 0 deletions(-)
diff --git
---
common/mem.c | 11 +++
common/mem.h |2 ++
2 files changed, 13 insertions(+), 0 deletions(-)
diff --git a/common/mem.c b/common/mem.c
index 7236cf0..cfc3ef1 100644
--- a/common/mem.c
+++ b/common/mem.c
@@ -186,6 +186,17 @@ void *spice_realloc_n(void *mem, size_t n_blocks, size_t
Hey,
I've been looking at the multiclient branch this morning, in particular
commit c1b3e13ee3549 server/main_channel: support multiple clients.
The AgentDataPipeItemRefs stuff got me really confused, after a while I think
I figured it out, ie, AgentDataPipeItemRefs is used to reference count the
On Wed, Aug 17, 2011 at 05:32:49PM -0700, Paolo Bonzini wrote:
> On 08/17/2011 06:17 AM, Christophe Fergeau wrote:
> >ACK series, but I haven't read the winfiber code really carefully (since I
> >know nothing about it).
>
> I compared it with the implementation in QEMU and it seems good.
Cool, th
Hey,
On Sun, Aug 21, 2011 at 09:28:58AM +0300, Yonit Halperin wrote:
> resending the 8th patch Christophe sent for review (and omitted debug prints)
This patch works for me, but I have no idea whether this is the right way
of fixing it or not :-/
Christophe
pgpdOBkrJhtul.pgp
Description: PGP s
On Wed, Aug 17, 2011 at 10:57:26AM -0700, Alon Levy wrote:
>
> ACK. Actually also the previous patch. I can rebase on top I guess, and
> it might make sure I notice if the multiclient patchset breaks it or not.
Oh this can wait after the multiclient work lands if that makes things
easier for you.
