Date: Thu, 28 Jan 2016 19:06:50 +0100 From: Harry Schmalzbauer <freebsd@omnilan.de> To: Marius Strobl <marius@freebsd.org> Cc: FreeBSD Stable <freebsd-stable@freebsd.org> Subject: em(4) watchdog timeout redemption [Was: Re: svn commit: r294958 - in stable/10: share/man/man4 sys/dev/e1000 sys/dev/ixgb sys/dev/netmap] Message-ID: <56AA58BA.8070404@omnilan.de> In-Reply-To: <201601272231.u0RMV8LW019394@repo.freebsd.org> References: <201601272231.u0RMV8LW019394@repo.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Bez=FCglich Marius Strobl's Nachricht vom 27.01.2016 23:31 (localtime): > Author: marius > Date: Wed Jan 27 22:31:08 2016 > New Revision: 294958 > URL: https://svnweb.freebsd.org/changeset/base/294958 > > Log: > Sync the e1000 drivers with what's in head as of r294327, modulo part= s > that don't apply to stable/10 (driver API, if_inc_counter(), RSS chan= ges > etc.) and modulo r287465 (which reportedly breaks igb(4)), i. e. asso= rted > fixes and improvements only: > =20 > o MFC r267385 (partial): > - Don't compare bus_dma map pointers for static DMA allocations aga= inst > NULL to determine if bus_dmamap_unload() or bus_dmamem_free() sho= uld be > called. Instead, check the associated bus and virtual addresses. > - Don't clear static DMA maps to NULL. > o MFC r284933: > Delete the refernce to VLAN handling being disabled by default. Thi= s is > no longer the case. [1] > o MFC r285639: > Add an adapter CORE lock in the DDB hook em_dump_queue to avoid WIT= NESS > panic in em_init_locked() while debugging. > o MFC r285879: > - Remove unused txd_saved. > - Intialize txd_upper, txd_lower and txd_used at declaration. > o MFC r286162: > Free mbufs when busdma loading fails. > o MFC r286829: > Add capability to disable CRC stripping as it breaks IPMI/BMC capab= ilities > on certain adatpers. [2] > o MFC r286831: [3] > - Increase EM_MAX_SCATTER to 64 such that the size of em_xmit():: > segs[EM_MAX_SCATTER] doesn't get overrun by things like NFS that = can > and do shove more than 32 segs when being used with em(4) and TSO= 4. > - Update tso handling code in em_xmit() with update from jhb@ > - Set if_hw_tsomax, if_hw_tsomaxsegcount and if_hw_tsomaxsegsize to= > appropriate values. > - Define a TSO workaround "magic" number of 4 that is used to avoid= an > alignment issue in hardware. > - Change a couple of integer values that were used as booleans to a= ctual > bool types. > - Ensure that em_enable_intr() enables the appropriate mask of inte= rrupts > and not just a hardcoded define of values. > o MFC r286832: > e1000/if_lem.c bump to 1.1.0 > o MFC r286833: > Bump all copywrite dates to 2015. > o MFC r287112: > Style/whitespace cleanup in shared/common code. > o MFC r293331: > - Switch em(4) to the extended RX descriptor format. > - Split rxbuffer and txbuffer apart to support the new RX descripto= r > format structures. Move rxbuffer manipulation to em_setup_rxdesc(= ) to > unify the new behavior changes. > - Add a RSSKEYLEN macro for help in generating the RSSKEY data stru= ctures > in the card. > - Change em_receive_checksum() to process the new rxdescriptor form= at > status bit. > o MFC r293332: > Disable the reuse of checksum offload context descriptors in the ca= se > of multiple queues in em(4). Document errata in the code. > o MFC r293854: > Given that em(4), lem(4) and igb(4) hardware doesn't require the > alignment guarantees provided by m_defrag(9), use m_collapse(9) > instead for performance reasons. > While at it, sanitize the statistics softc members, i. e. retire > unused ones and add SYSCTL nodes missing for actually used ones. > =20 > PR: 118693 [1], 161277 [2], 195078 [3], 199174 [3], 200221 [3] Thanks, especially to sbruno@ I'd like to confirm r294958 fixes multiple em(4) problems I observed up to r294156, especially EM_MULTIQUEUE support on hartwell (82574) (haven't filed a bug report since I haven't had time to analyze, seems 199174 and 200221 match well). Glad to see 10.3 will ship with em(4) able to sustain GbE with one NFS transfer (111,3MiB/s), while keeping low latency for additional (low-trhoughput) connections without having unrecoverably watchdog timeouts anymore (adding 2nd queue to em(4) reduces latency from 10ms to ~3ms on new sockets). For the records, this kind of watchdog timeouts with unsuccessful interface resets are fixed for me: em0: Watchdog timeout Queue[0]-- resetting Interface is RUNNING and ACTIVE em0: TX Queue 0 ------ em0: hw tdh =3D 210, hw tdt =3D 674 em0: Tx Queue Status =3D -2147483648 em0: TX descriptors avail =3D 3632 em0: Tx Descriptors avail failure =3D 0 em0: RX Queue 0 ------ em0: hw rdh =3D 896, hw rdt =3D 895 em0: RX discarded packets =3D 0 em0: RX Next to Check =3D 896 em0: RX Next to Refresh =3D 895 em0: TX Queue 1 ------ em0: hw tdh =3D 575, hw tdt =3D 716 em0: Tx Queue Status =3D -2147483648 em0: TX descriptors avail =3D 3937 em0: Tx Descriptors avail failure =3D 0 em0: RX Queue 1 ------ em0: hw rdh =3D 192, hw rdt =3D 191 em0: RX discarded packets =3D 0 em0: RX Next to Check =3D 192 em0: RX Next to Refresh =3D 191 em0: link state changed to DOWN em0: link state changed to UP -Harry
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?56AA58BA.8070404>