3 messages, 3 authors, 2007-07-10 · open the first message on its own page
From: Roland Dreier <hidden>
Date: 2007-07-09 21:35:48
Out of curiousity, does this mean that a GRH will be sent on all UD messages (for non-LL QPs)? What decides if a QP is LL or not? - R.
From: Joachim Fenkes <hidden>
Date: 2007-07-10 11:26:23
Roland Dreier [off-list ref] wrote on 09.07.2007 23:35:31:
Out of curiousity, does this mean that a GRH will be sent on all UD messages (for non-LL QPs)?
No - the bit instructs the hardware to fetch the GRH parts of the QP context. The GRH will only be used if the WQE says so. Joachim
From: Christoph Raisch <hidden>
Date: 2007-07-10 13:49:20
What decides if a QP is LL or not? - R.
Currently we use a high bit in the QP type, which is not how we want to keep it permanently. What would you suggest, add two additional LL QP types, or change something more fundamental in libibverbs and kernel ib core? We think we can get along quite well with the existing parameters in the current create QP. The current user-kernel interface is ok for these new QPs for post_send + post_recv, but unfortunately the libibverbs userspace calls don't match exactly how the LL queues are to be used. We would need something like the LL QP interface in libehca in libibverbs to keep that interface generic. We didn't see a usage yet for LL QP in kernel, so maybe we should continue that discussion on general@openfabrics only. We could provide example code in libehca/samples if needed. Gruss / Regards Christoph + Nam