[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [RFC] Generating Go bindings for libxl


  • To: Nicholas Rosbrook <rosbrookn@xxxxxxxxxxxx>, "xen-devel@xxxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxxx>
  • From: George Dunlap <george.dunlap@xxxxxxxxxx>
  • Date: Wed, 4 Sep 2019 17:52:16 +0100
  • Authentication-results: esa4.hc3370-68.iphmx.com; dkim=none (message not signed) header.i=none; spf=None smtp.pra=george.dunlap@xxxxxxxxxx; spf=Pass smtp.mailfrom=George.Dunlap@xxxxxxxxxx; spf=None smtp.helo=postmaster@xxxxxxxxxxxxxxx
  • Autocrypt: addr=george.dunlap@xxxxxxxxxx; prefer-encrypt=mutual; keydata= mQINBFPqG+MBEACwPYTQpHepyshcufo0dVmqxDo917iWPslB8lauFxVf4WZtGvQSsKStHJSj 92Qkxp4CH2DwudI8qpVbnWCXsZxodDWac9c3PordLwz5/XL41LevEoM3NWRm5TNgJ3ckPA+J K5OfSK04QtmwSHFP3G/SXDJpGs+oDJgASta2AOl9vPV+t3xG6xyfa2NMGn9wmEvvVMD44Z7R W3RhZPn/NEZ5gaJhIUMgTChGwwWDOX0YPY19vcy5fT4bTIxvoZsLOkLSGoZb/jHIzkAAznug Q7PPeZJ1kXpbW9EHHaUHiCD9C87dMyty0N3TmWfp0VvBCaw32yFtM9jUgB7UVneoZUMUKeHA fgIXhJ7I7JFmw3J0PjGLxCLHf2Q5JOD8jeEXpdxugqF7B/fWYYmyIgwKutiGZeoPhl9c/7RE Bf6f9Qv4AtQoJwtLw6+5pDXsTD5q/GwhPjt7ohF7aQZTMMHhZuS52/izKhDzIufl6uiqUBge 0lqG+/ViLKwCkxHDREuSUTtfjRc9/AoAt2V2HOfgKORSCjFC1eI0+8UMxlfdq2z1AAchinU0 eSkRpX2An3CPEjgGFmu2Je4a/R/Kd6nGU8AFaE8ta0oq5BSFDRYdcKchw4TSxetkG6iUtqOO ZFS7VAdF00eqFJNQpi6IUQryhnrOByw+zSobqlOPUO7XC5fjnwARAQABtCRHZW9yZ2UgVy4g RHVubGFwIDxkdW5sYXBnQHVtaWNoLmVkdT6JAlcEEwEKAEECGwMFCwkIBwMFFQoJCAsFFgID AQACHgECF4ACGQEWIQTXqBy2bTNXPzpOYFimNjwxBZC0bQUCXEowWQUJDCJ7dgAKCRCmNjwx BZC0beKvEACJ75YlJXd7TnNHgFyiCJkm/qPeoQ3sFGSDZuZh7SKcdt9+3V2bFEb0Mii1hQaz 3hRqZb8sYPHJrGP0ljK09k3wf8k3OuNxziLQBJyzvn7WNlE4wBEcy/Ejo9TVBdA4ph5D0YaZ nqdsPmxe/xlTFuSkgu4ep1v9dfVP1TQR0e+JIBa/Ss+cKC5intKm+8JxpOploAHuzaPu0L/X FapzsIXqgT9eIQeBEgO2hge6h9Jov3WeED/vh8kA7f8c6zQ/gs5E7VGALwsiLrhr0LZFcKcw kI3oCCrB/C/wyPZv789Ra8EXbeRSJmTjcnBwHRPjnjwQmetRDD1t+VyrkC6uujT5jmgOBzaj KCqZ8PcMAssOzdzQtKmjUQ2b3ICPs2X13xZ5M5/OVs1W3TG5gkvMh4YoHi4ilFnOk+v3/j7q 65FG6N0JLb94Ndi80HkIOQQ1XVGTyu6bUPaBg3rWK91Csp1682kD/dNVF3FKHrRLmSVtmEQR 5rK0+VGc/FmR6vd4haKGWIRuPxzg+pBR77avIZpU7C7+UXGuZ5CbHwIdY8LojJg2TuUdqaVj yxmEZLOA8rVHipCGrslRNthVbJrGN/pqtKjCClFZHIAYJQ9EGLHXLG9Pj76opfjHij3MpR3o pCGAh6KsCrfrsvjnpDwqSbngGyEVH030irSk4SwIqZ7FwLkBDQRUWmc6AQgAzpc8Ng5Opbrh iZrn69Xr3js28p+b4a+0BOvC48NfrNovZw4eFeKIzmI/t6EkJkSqBIxobWRpBkwGweENsqnd 0qigmsDw4N7J9Xx0h9ARDqiWxX4jr7u9xauI+CRJ1rBNO3VV30QdACwQ4LqhR/WA+IjdhyMH wj3EJGE61NdP/h0zfaLYAbvEg47/TPThFsm4m8Rd6bX7RkrrOgBbL/AOnYOMEivyfZZKX1vv iEemAvLfdk2lZt7Vm6X/fbKbV8tPUuZELzNedJvTTBS3/l1FVz9OUcLDeWhGEdlxqXH0sYWh E9+PXTAfz5JxKH+LMetwEM8DbuOoDIpmIGZKrZ+2fQARAQABiQNbBBgBCgAmAhsCFiEE16gc tm0zVz86TmBYpjY8MQWQtG0FAlxKMJ4FCQnQ/OQBKcBdIAQZAQoABgUCVFpnOgAKCRCyFcen x4Qb7cXrCAC0qQeEWmLa9oEAPa+5U6wvG1t/mi22gZN6uzQXH1faIOoDehr7PPESE6tuR/vI CTTnaSrd4UDPNeqOqVF07YexWD1LDcQG6PnRqC5DIX1RGE3BaSaMl2pFJP8y+chews11yP8G DBbxaIsTcHZI1iVIC9XLhoeegWi84vYc8F4ziADVfowbmbvcVw11gE8tmALCwTeBeZVteXjh 0OELHwrc1/4j4yvENjIXRO+QLIgk43kB57Upr4tP2MEcs0odgPM+Q+oETOJ00xzLgkTnLPim C1FIW2bOZdTj+Uq6ezRS2LKsNmW+PRRvNyA5ojEbA/faxmAjMZtLdSSSeFK8y4SoCRCmNjwx BZC0bevWEACRu+GyQgrdGmorUptniIeO1jQlpTiP5WpVnk9Oe8SiLoXUhXXNj6EtzyLGpYmf kEAbki+S6WAKnzZd3shL58AuMyDxtFNNjNeKJOcl6FL7JPBIIgIp3wR401Ep+/s5pl3Nw8Ii 157f0T7o8CPb54w6S1WsMkU78WzTxIs/1lLblSMcvyz1Jq64g4OqiWI85JfkzPLlloVf1rzy ebIBLrrmjhCE2tL1RONpE/KRVb+Q+PIs5+YcZ+Q1e0vXWA7NhTWFbWx3+N6WW6gaGpbFbopo FkYRpj+2TA5cX5zW148/xU5/ATEb5vdUkFLUFVy5YNUSyeBHuaf6fGmBrDc47rQjAOt1rmyD 56MUBHpLUbvA6NkPezb7T6bQpupyzGRkMUmSwHiLyQNJQhVe+9NiJJvtEE3jol0JVJoQ9WVn FAzPNCgHQyvbsIF3gYkCYKI0w8EhEoH5FHYLoKS6Jg880IY5rXzoAEfPvLXegy6mhYl+mNVN QUBD4h9XtOvcdzR559lZuC0Ksy7Xqw3BMolmKsRO3gWKhXSna3zKl4UuheyZtubVWoNWP/bn vbyiYnLwuiKDfNAinEWERC8nPKlv3PkZw5d3t46F1Dx0TMf16NmP+azsRpnMZyzpY8BL2eur feSGAOB9qjZNyzbo5nEKHldKWCKE7Ye0EPEjECS1gjKDwbkBDQRUWrq9AQgA7aJ0i1pQSmUR 6ZXZD2YEDxia2ByR0uZoTS7N0NYv1OjU8v6p017u0Fco5+Qoju/fZ97ScHhp5xGVAk5kxZBF DT4ovJd0nIeSr3bbWwfNzGx1waztfdzXt6n3MBKr7AhioB1m+vuk31redUdnhbtvN7O40MC+ fgSk5/+jRGxY3IOVPooQKzUO7M51GoOg4wl9ia3H2EzOoGhN2vpTbT8qCcL92ZZZwkBRldoA Wn7c1hEKSTuT3f1VpSmhjnX0J4uvKZ1V2R7rooKJYFBcySC0wa8aTmAtAvLgfcpe+legOtgq DKzLuN45xzEjyjCiI521t8zxNMPJY9FiCPNv0sCkDwARAQABiQI8BBgBCgAmAhsMFiEE16gc tm0zVz86TmBYpjY8MQWQtG0FAlxKNJYFCQnQrVkACgkQpjY8MQWQtG2Xxg//RrRP+PFYuNXt 9C5hec/JoY24TkGPPd2tMC9usWZVImIk7VlHlAeqHeE0lWU0LRGIvOBITbS9izw6fOVQBvCA Fni56S12fKLusWgWhgu03toT9ZGxZ9W22yfw5uThSHQ4y09wRWAIYvhJsKnPGGC2KDxFvtz5 4pYYNe8Icy4bwsxcgbaSFaRh+mYtts6wE9VzyJvyfTqbe8VrvE+3InG5rrlNn51AO6M4Wv20 iFEgYanJXfhicl0WCQrHyTLfdB5p1w+072CL8uryHQVfD0FcDe+J/wl3bmYze+aD1SlPzFoI MaSIXKejC6oh6DAT4rvU8kMAbX90T834Mvbc3jplaWorNJEwjAH/r+v877AI9Vsmptis+rni JwUissjRbcdlkKBisoUZRPmxQeUifxUpqgulZcYwbEC/a49+WvbaYUriaDLHzg9xisijHwD2 yWV8igBeg+cmwnk0mPz8tIVvwi4lICAgXob7HZiaqKnwaDXs4LiS4vdG5s/ElnE3rIc87yru 24n3ypeDZ6f5LkdqL1UNp5/0Aqbr3EiN7/ina4YVyscy9754l944kyHnnMRLVykg0v+kakj0 h0RJ5LbfLAMM8M52KIA3y14g0Fb7kHLcOUMVcgfQ3PrN6chtC+5l6ouDIlSLR3toxH8Aam7E rIFfe2Dk+lD9A9BVd2rfoHA=
  • Cc: "anthony.perard@xxxxxxxxxx" <anthony.perard@xxxxxxxxxx>, Brendan Kerrigan <kerriganb@xxxxxxxxxxxx>, "ian.jackson@xxxxxxxxxxxxx" <ian.jackson@xxxxxxxxxxxxx>, Nicolas Belouin <nicolas.belouin@xxxxxxxxx>, "wl@xxxxxxx" <wl@xxxxxxx>
  • Delivery-date: Wed, 04 Sep 2019 16:52:26 +0000
  • Ironport-sdr: a8i7fX02H1N9zFmSyGga8PjytLWhKx/FpVGssahZE/VANRnJAaO2iVjpb9NbWPVQkKTmF415+5 rW2Qo+DofJxIq7vPnWGDXYHSSRRC/5Ds4Jr1cH92/PvQHBC3Cywv6fvB5RelyRkKhvoX4TWdjO BXHWGDBB9q7/Ljr2e1HJ52MDIblAZ65errdA8ZpEbvkzbG09ZWC6xVrcR1gPlekD9dStMHqBTn gCXWpyiBWvHY0iRq8ikqrCINgW2vsAwVjdytyL7xnj2rkGCq5a15Ywibh81j8MDK70wwbhdj1h s6o=
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
  • Openpgp: preference=signencrypt

On 9/4/19 1:36 AM, Nicholas Rosbrook wrote:
> George,
> 
>> Are you up for taking a stab at something like `gengotypes.py`?
> 
> I have was able to make a bit of progress on this over the weekend. I've 
> started
> `gengotypes.py` in my branch[1]; the portion which generates 
> `xenlight_types.go`
> (the counterpart to _libxl_types.h) is mostly working. 
> 
> The main exception is that I am not certain how the `KeyedUnion` type from 
> the IDL
> should be translated for Go. One option is to do something similar to the 
> `oneof` field 
> in gRPC's protobuf messages[2]. Essentially, we would define a separate 
> struct for each
> of the structs that belong to the union. Then, where a union would be used in 
> C, we use
> an interface type which the previously defined structs implement.

Yes, I think that's really the only option.  Poking around, it looks
like a lot of different people have recommended it; and the fact that
it's in use by gRPC means it can't be too terrible a solution.

The really annoying thing is that with the "interface-as-union", we
can't use anonymous types: we'll have to explicitly define the
{parent-struct} x {union-key} as a distinct type, and the is$TYPE()
method on each one.

> E.g.
> 
> type isDomainTypeStruct interface {
>         isDomainTypeStruct()
> }

The interface type itself will need to be exported, right?  (Obviously
we don't want to export the defining method.)

> type domainTypeHVMStruct struct {
>         ...
> }

So you've named the struct after the name of the key (libxl_domain_type)
and the key value (hvm); but I don't think that's sufficient.  Already
there are two different structures indexed by libxl_channel_connection:

typedef struct libxl_device_channel {
    libxl_domid backend_domid;
    char * backend_domname;
    libxl_devid devid;
    char * name;
    libxl_channel_connection connection;
    union {
        struct {
            char * path;
        } socket;
    } u;
} libxl_device_channel;

typedef struct libxl_channelinfo {
    char * backend;
    uint32_t backend_id;
    char * frontend;
    uint32_t frontend_id;
    libxl_devid devid;
    int state;
    int evtch;
    int rref;
    libxl_channel_connection connection;
    union {
        struct {
            char * path;
        } pty;
    } u;
} libxl_channelinfo;


(Note that in one, `socket` is defined, and in the other, `pty` is
defined.  I'm not sure that's not a bug, but anyway, that's what the IDL
allows.)

And there's no reason, theoretically, we couldn't have the following:

    ("u1", KeyedUnion(None, libxl_channel_connection, "connection",
           [/* One set of types */,
           ])),
    ("u2", KeyedUnion(None, libxl_channel_connection, "connection2",
           [/* A second set of types set of types */,
           ])),

So we need to include the element name as well.  But actually, I suppose
that means we don't actually need to include the type, since the element
name will be unique.

> func (d domainTypeHVMStruct) isDomainTypeStruct() {}
> 
> type DomainBuildInfo struct {
>         ...
>         Type DomainType
>         dts    isDomainTypeStruct
> }

...and then I'm afraid you'd need to have 'Dts' (should be exported,
right?) instead by the element specified by the IDL; so 'U' in all the
current cases.

This gives us:

type DomainBuildInfoU interface {
    isDomainBuildInfoU()
}

type DomainBuildInfoUHvmstruct {
  // ...
}

func (s DomainBuildInfoUHvm) isDomainBuildInfoU() { }

...

struct DomainBuildInfo {
    // ...
    Type DomainType
    U    DomainBuildInfoU
    // ...
}

Alternately, since the "key" element is also unique, we could use that
instead:

type DomainBuildInfoTypeUnion {
 // ...
}

struct DomainBuildInfo {
    // ...
    Type      DomainType
    TypeUnion DomainBuildInfoTypeUnion
    // ...
}

(And in the example given above where there are two keyed unions, one
would be "ConnectionUnion ParentTypeConnectionUnion" and
"Connection2Union ParentTypeConnection2Union".)

I think the second one looks prettier.  (Actually I think using 'u' as
the element name for the union was kind of a bad idea in the first
place.)  But that does mean we're 'overriding' the instructions of the
IDL (which prescribe both the key element name and the union element name).

What do you think?  If like me, you prefer the second one, I'll try to
ping Ian Jackson to make sure he doesn't have any objections to it.

 -George

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.