Roland Dreier wrote:
Vu, you pointed out that the current SRP code might look at an IU that it sent after that IU has been reused for a different command. I realized that a simple fix for this is just to keep the DMA address (the only thing we look at in the IU) in the request structure.
Just thinking about it and you beat me with the patch
To add FMR support, we can just put all the FMR stuff in the request structure instead of the IU structure. This saves bloating the IUs we use for receives and task management, so it seems like a win anyway.
Since all the tuned parameter are target-centralized (passing in when add new target) I think about moving FMR resources (size, max_page...) ie. fmr_pool into srp_target_port struct. Each newly added target will have their own customized FMR pool.
Have you reviewed the FMR? What your take on Christoph's point about the high bit of dma_address_ts are used by some platforms IOMMU - I think that it's OK since FMR code only touch the lower bit of dma_address_ts
Does this patch seem OK and work for you? It works for me in my setup.
Yes the patch without FMR integration works for me - I'll integrate FMR in, test and let you know later
vu _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
