[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [virtio-dev] RFC: Doorbell suppression, packed-ring mode and hardware offload
On 2019/2/13 äå11:20, Michael S. Tsirkin wrote:
On Wed, Feb 13, 2019 at 06:00:41PM +0800, Jason Wang wrote:On 2019/2/12 äå9:44, Michael S. Tsirkin wrote:On Tue, Feb 12, 2019 at 03:28:26PM +0800, Jason Wang wrote:On 2019/2/12 äå1:08, Michael S. Tsirkin wrote:On Mon, Feb 11, 2019 at 02:58:30PM +0000, David Riddoch wrote:This can result in a very high rate of doorbells with some drivers, which can become a severe bottleneck (because x86 CPUs can't emit MMIOs at very high rates).Interesting. Is there any data you could share to help guide the design? E.g. what's the highest rate of MMIO writes supported etc?On an E3-1270 v5 @ 3.60GHz, max rate across all cores is ~18M/s. On an E5-2637 v4 @ 3.50GHz, max rate on PCIe-local socket is ~14M/s. On PCIe-remote socket ~8M/s.Is that mega bytes? Or million writes? With 32 bit writes are we limited to 2M then?Sorry, this is million-writes-per-second. The size of the write doesn't matter.So it's not too bad, you only need one per batch after all. Also I thought some more about this. In fact on x86/intel specifically PCI descriptor reads are atomic with cache line granularity and writes are ordered. So in fact you can safely prefetch descriptors and if you see them valid, you can go ahead and use them.If I understand this correctly, unless driver set avail in order for several descriptors. Device could not assume that if N is avail then [0, N) is also available.Well you don't need that assumption. You check it and discard prefetched data if not.Ok, this is probably ok for hardware implementation. But for software implementation, it may result lots of re-read.That will be on a slow path where you got unlucky.
Yes.
But prefetching indeed improve the performance when I play with vhost kernel implementation. It reduces the overhead of memory accessing, which is similar to PCI.Prefetch as in prefetchnta?Sorry, "prefetch" is probably not accurate. "Speculation" is probably better. What I did is just read more than one descriptor e.g 16. And check whether the last is available, if yes, assume all the rest are available.So how problematic is it to check them all? In fact you can optimize out the rmb (or dependency barrier), just issue one barrier after checking all.
Let me reply in another thread.
This makes the descriptor index merely a hint for performance, which device can play with at will. Other platforms are not like this so you need the data but do they have the same problem?This doesn't just impose a limit on aggregate packet rate: If you hit this bottleneck then the CPU core is back-pressured by the MMIO writes, and so instr/cycle takes a huge hit.The proposed offset/wrap field allows devices to disable doorbells when appropriate, and determine the latest fill level via a PCIe read.This kind of looks like a split ring though, does it not?I don't agree, because the ring isn't split. Split-ring is very painful for hardware because of the potentially random accesses to the descriptor table (including pointer chasing) which inhibit batching of descriptor fetches, as well as causing complexity in implementation.That's different with IN_ORDER, right?Yes, IN_ORDER will give some of the benefit of packed ring, and is also a win because you can merge 'used' writes. (But packed-ring still wins due to not having to fetch ring and descriptor table separately).Note, with IN_ORDER, there's no need to read available ring at all even for split ring. And in some cases, no need to update used ring at all.Yes you said so at some point but it seems that the spec is written ambigously and others do not read it like this. Can you clarify?Yes, it was unclear about how available ring was dealt with in this case. My understanding is, since device use the descriptor table in order, if last_avail_idx is N and avail_idx is M, driver could assume [N, M) is available. In this case, avail_idx points to idx of descriptor ring instead of avail ring. So there's no need for both side to read or write avail ring entries. Does this make sense? 10% PPS was seen through this way when I playing this for vhost.Problem is, if you add a chain of descriptors, index is incremented by 1 while the descriptor buffer advances by X. So how can you find out without reading avail ring?
I think it's as simple as increase the avail idx by X? Since descriptor were used in order, device can just read the next X-1 descriptors in this case.
Thanks
Maybe you can start with fetching based on index (so 1 above), then fetch an extra descriptor each time you see a NEXT bit. And given PV tends to use INDIRECT aggressively, that won't happen too often. Is that the idea?Packed ring is a huge improvement on both dimensions, but introduces a specific pain point for hardware offload.The issue is we will again need to bounce more cache lines to communicate.You'll only bounce cache lines if the device chooses to read the idx. A PV device need not offer this feature. A PCIe device will, but the cost to the driver of writing an in-memory idx is much smaller than the cost of an MMIO, which is always a strongly ordered barrier on x86/64. With vDPA you ideally would have this feature enabled, and the device would sometimes be PV and sometimes PCIe. The PV device would ignore the new idx field and so cache lines would not bounce then either. Ie. The only time cache lines are shared is when sharing with a PCIe device, which is the scenario when this is a win.True. OTOH on non-x86 you will need more write barriers :( It would be natural to say driver can reuse the barrier before flags update, but note that that would mean hardware still needs to support re-fetching if flags value is wrong. Such re-fetch is probably rare so fine by me, but it does add a bit more complexity.I would prefer to have the write barrier before writing the idx.Well that's driver overhead for something device might never utilise in a given workload. If we are optimizing let's optimize for speed.Note that drivers could take advantage of this feature to avoid read barriers when consuming descriptors: At the moment there is a virtio_rmb() per descriptor read. With the proposed feature the driver can do just one barrier after reading idx.Oh you want device to write a used index too? Hmm. Isn't this contrary to your efforts to consume PCIe BW?I expect that on platforms where write barriers have a cost read barriers likely have a significant cost too, so this might be a win with PV devices too.I'm not sure. It is pretty easy to replace an rmb with a dependency which is generally quite cheap in my testing. But since it's supposed to benefit PV, at this point we already have implementations so rather than speculate (pun intended), people can write patches and experiment. For the proposed avail idx I think Jason (CC'd) was interested in adding something like this for vhost.Yes, it's kind of IN_ORDER + split ring I believe. ThanksWell IN_ORDER is only good for e.g. network. Storage wants out of order so it could benefit from a new feature. I am however worried about the extra wmb implied. Still thinking about ways to avoid adding that.Ok. ThanksSo I wonder: what if we made a change to spec that would allow prefetch of ring entries? E.g. you would be able to read at random and if you guessed right then you can just use what you have read, no need to re-fetch?Unless I've misunderstood I think this would imply that the driver would have to ensure strict ordering for each descriptor it wrote, which would impose a cost to the driver. At the moment a write barrier is only needed before writing the flags of the first descriptor in a batch.On non-x86 right? OTOH the extra data structure also adds more ordering requirements.Yes on non-x86. The extra data structure only adds an ordering once per request (when enabled) whereas allowing prefetch would add an ordering per descriptor. The number of descriptors is always >= the number of requests, and can be much larger (esp. for a net rx virt-queue).Question is, is MMIO also slow on these non-x86 platforms? Prefetch if successful just drastically lowers latency. You can't beat not needing a read at all.I suggest the best place to put this would be in the driver area, immediately after the event suppression structure.Could you comment on why is that a good place though?The new field is written by the driver, as are the other fields in the driver area. Also I expect that devices might be able to read this new idx together with the interrupt suppression fields in a single read, reducing PCIe overheads.OK then you need it in the same aligned 256 byte boundary.The event suppression structs currently require 4byte align. Are you suggesting we increase the align required when VIRTIO_F_RING_PACKED_AVAIL_IDX is negotiated? Sounds good to me, but 8bytes would presumably suffice.OK. I am also wondering: what would the analog of this feature be for split rings? We are burning a feature bit, might as well find a good use for it.Thanks for the feedback. David -- David Riddoch <driddoch@solarflare.com> -- Chief Architect, Solarflare --------------------------------------------------------------------- To unsubscribe, e-mail: virtio-dev-unsubscribe@lists.oasis-open.org For additional commands, e-mail: virtio-dev-help@lists.oasis-open.org
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]