> I'm not convinced it needs to be a PERF_RECORD_SAMPLE; but some > PERF_RECORD_* type for sure.
Adding a header shouldn't be a problem, it's merely wasting 4K. > Also it must allow interleaving with other > events. But can you describe a concrete use case where interleaving is better? I'm not aware of any. Anything that could be usefully interleaved can just be in the side band stream, and if you want a unified uncompressed stream you just run perf inject. The standard tools don't care for it as they have to reorder everything anyways to deal with multi CPU reordering. Your scheme is very complex and adds a lot of use restrictions over the current code, so there should be a good reason for it at least. Especially the TLB hac^wproposal sounds horrible to me, compared to the straight forward zero copy ring buffer used today. -Andi -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/