On 9/16/2016 12:16 PM, Jason Gunthorpe wrote: > On Fri, Sep 16, 2016 at 11:12:17AM -0400, Doug Ledford wrote: > >> I would. Mainly because it is tied to both the kernel uAPI (via >> netlink)and to librdmacm > > With what little I know about ibacm I'm inclined to agree. It seems > like a utility for librdmacm, and is consuming a kernel interface. > > Reconsider Sean? > >> (such that librdmacm changes how it builds when ibacm is present, so >> if we don't include it, then ibacm becomes a requirement of the >> larger package if you want librdmacm to support the ibacm cache >> service). > > I double checked and I couldn't find anything like > this. librdmacm/src/amc.c builds unconditionally and contains no > ifdefs, configure does not key on anything to do with acm. Sorry, must have changed. It used to do so (configure used to test for presence of ibacm header file and change compile settings accordingly as I recall). -- Doug Ledford <dledford@xxxxxxxxxx> GPG Key ID: 0E572FDD
Attachment:
signature.asc
Description: OpenPGP digital signature