[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <f8c68944-57b6-18a6-aa50-c019baa2a63b@gmail.com>
Date: Tue, 19 Jun 2018 14:27:37 -0600
From: David Ahern <dsahern@...il.com>
To: Stephen Hemminger <stephen@...workplumber.org>,
Steve Wise <swise@...ngridcomputing.com>
Cc: netdev <netdev@...r.kernel.org>
Subject: Re: iproute2 won't compile without AF_VSOCK
On 6/19/18 9:47 AM, Stephen Hemminger wrote:
> On Tue, 19 Jun 2018 10:17:45 -0500
> Steve Wise <swise@...ngridcomputing.com> wrote:
>
>> Hey David,
>>
>> I'm trying to compile the latest iproute2 on an RHEL-7.3 distro, and it
>> fails to compile because AF_VSOCK is not defined. Should this
>> functionality be a configure option to disable it on older distros?
>>
>>
>> Thanks,
>>
>> Steve.
>>
>> ----
>>
>> misc
>> CC ss.o
>> ss.c:301:27: error: ‘AF_VSOCK’ undeclared here (not in a function)
>> .families = FAMILY_MASK(AF_VSOCK),
>> ^
>> ss.c:252:46: note: in definition of macro ‘FAMILY_MASK’
>> #define FAMILY_MASK(family) ((uint64_t)1 << (family))
>> ^
>> ss.c:334:2: error: array index in initializer not of integer type
>> [AF_VSOCK] = {
>> ^
>> ss.c:334:2: error: (near initialization for ‘default_afs’)
>> make[1]: *** [ss.o] Error 1
>> make: *** [all] Error 2
>>
>
> Probably should just add an #ifdef to takeout that if not present
>
Most userspace tools have a compat header for cases like this.
#ifndef AF_VSOCK
#define AF_VSOCK 40
#endif
Powered by blists - more mailing lists