Date: Thu, 13 Jul 2017 22:12:41 +0000 (UTC) From: Eric Joyner <erj@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r320972 - head/sys/dev/ixl Message-ID: <201707132212.v6DMCfrC007004@repo.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: erj Date: Thu Jul 13 22:12:41 2017 New Revision: 320972 URL: https://svnweb.freebsd.org/changeset/base/320972 Log: ixl(4)/ixlv(4): Stop leaking every busdma entry in receive path From Brett: In short, busdma maps for received packets were not being unloaded in the interrupt handler before the packets were passed up the network stack. The fix was to add a busdma sync and unload for the two receive maps. This bug is significant for certain busdma providers, for example IOMMUs, where not unloading the maps means that 1) the IOMMU mappings that allow the NIC to DMA the received packets into host memory stay open indefinitely, potentially violating a desired security policy, and 2) resources such as device address space addresses and host memory for bookkeeping are never freed. Without an IOMMU or bounce buffering enabled for the ixl device, I don't think adding these calls will have any significant performance impact. With the IOMMU enabled, I have noticed a performance impact on the receive side, which is expected. Submitted by: Brett Gutstein <bgutstein@rice.edu> Reviewed by: erj@ MFC after: 1 week Modified: head/sys/dev/ixl/ixl_txrx.c Modified: head/sys/dev/ixl/ixl_txrx.c ============================================================================== --- head/sys/dev/ixl/ixl_txrx.c Thu Jul 13 22:01:38 2017 (r320971) +++ head/sys/dev/ixl/ixl_txrx.c Thu Jul 13 22:12:41 2017 (r320972) @@ -1578,6 +1578,18 @@ ixl_rxeof(struct ixl_queue *que, int count) else vtag = 0; + /* Remove device access to the rx buffers. */ + if (rbuf->m_head != NULL) { + bus_dmamap_sync(rxr->htag, rbuf->hmap, + BUS_DMASYNC_POSTREAD); + bus_dmamap_unload(rxr->htag, rbuf->hmap); + } + if (rbuf->m_pack != NULL) { + bus_dmamap_sync(rxr->ptag, rbuf->pmap, + BUS_DMASYNC_POSTREAD); + bus_dmamap_unload(rxr->ptag, rbuf->pmap); + } + /* ** Make sure bad packets are discarded, ** note that only EOP descriptor has valid
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201707132212.v6DMCfrC007004>