On Wed, Jul 24, 2019 at 10:02:04AM -0500, Eric Blake wrote:
On 7/24/19 7:17 AM, Richard W.M. Jones wrote:
In preparation for closure lifetimes, split up the Closure so it no longer describes a list of closures, but a single callback.
This changes the API because functions which take 2 or more closures now pass a separate user_data for each one. --- docs/libnbd.pod | 3 +- examples/strict-structured-reads.c | 2 +- generator/generator | 760 ++++++++++++---------------- generator/states-reply-simple.c | 2 +- generator/states-reply-structured.c | 8 +- lib/internal.h | 3 +- lib/rw.c | 32 +- 7 files changed, 364 insertions(+), 446 deletions(-)
diff --git a/docs/libnbd.pod b/docs/libnbd.pod index 5608e63..631bb3b 100644 --- a/docs/libnbd.pod +++ b/docs/libnbd.pod @@ -487,8 +487,7 @@ C<nbd_set_debug_callback>, C<nbd_pread_callback>). Libnbd can call these functions while processing.
Callbacks have an opaque C<void *user_data> pointer. This is passed
Should we tweak this to explicitly call out "Callbacks in the C language", to differentiate it from bindings to other languages that do not do this?
Yes, the documentation is rather C-specific. I'll add it (but as another patch on top).
-as the first parameter to the callback. libnbd functions that take -two callback pointers share the same opaque data for both calls. +as the first parameter to the callback.
+++ b/generator/generator @@ -849,10 +849,10 @@ and arg = written by the function *) | BytesPersistIn of string * string (* same as above, but buffer persists *) | BytesPersistOut of string * string -| Closure of bool * closure list (* void *opaque + one or more closures - flag if true means callbacks persist - in the handle, false means they only - exist during the function call *) +| Closure of bool * closure (* function pointer + void *opaque + flag if true means callbacks persist + in the handle, false means they only + exist during the function call *)
Do we still need the closure record type, or can/should we simplify further to:
| Closure of bool * string * arg list
(and later in the series, even further by dropping 'bool *')
There's actually no difference at the machine code level - both tuple and struct have the same representation (same as it would be in C in fact). So it comes down to which is nicer, and I think it's slightly nicer to have the named fields.
| Flags of string (* NBD_CMD_FLAG_* flags *) | Int of string (* small int *) | Int64 of string (* 64 bit signed int *) @@ -921,8 +921,8 @@ Return the state of the debug flag on this handle."; "set_debug_callback", { default_call with args = [ Closure (true, - [{ cbname="debug_fn"; - cbargs=[String "context"; String "msg"] }]) ]; + { cbname="debug_fn"; + cbargs=[String "context"; String "msg"] }) ];
Keeping the record type means that you still have to provide names, vs. directly using tuples where we'd write:
args = [ Closure (true, "debug_fn", [String "context"; String "msg"]) ];
Indeed.
@@ -3811,154 +3795,140 @@ let print_python_binding name { args; ret } = *) List.iter ( function - | Closure (persistent, cls) -> - pr "struct %s_user_data {\n" name; - List.iter ( - fun { cbname } -> - pr " PyObject *%s;\n" cbname - ) cls; - pr "};\n"; - pr "\n"; - + | Closure (persistent, { cbname; cbargs }) -> (* Persistent closures need an explicit function to decrement * the closure refcounts and free the user_data struct.
Comment needs a touchup now that there is no user_data struct.
This comment goes away in the following commit :-)
+ pr " py_args = Py_BuildValue (\"(\""; + List.iter ( + function + | ArrayAndLen (UInt32 n, len) -> pr " \"O\"" + | BytesIn (n, len) -> pr " \"y#\"" + | Int n -> pr " \"i\"" + | Int64 n -> pr " \"L\"" + | Mutable (Int n) -> pr " \"O\"" + | String n -> pr " \"s\"" + | UInt64 n -> pr " \"K\"" + (* The following not yet implemented for callbacks XXX *) + | ArrayAndLen _ | Bool _ | BytesOut _ + | BytesPersistIn _ | BytesPersistOut _
Not your usual indentation style.
I think at some point I gave up fighting tuareg mode :-( I need to ask Martin for the right settings to make it indent in the preferred way. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://people.redhat.com/~rjones/virt-top