src/sys/dev/twa, branch releng/8.3

MFC r215818, r216405, r216437, r216448, r216956, r221827, r222975, r223059,

2012-02-15T14:23:01Z

r225343, r225704, r225705, r225706, r225707, r225709, r226029, r220647, r230183, r230587, r230916, r228526, r230879: Bring Xen support in stable/8 up to parity with head. Almost all outstanding Xen support differences between head and stable/8 are included, except for the just added r231743. r215818 | cperciva | 2010-11-25 08:05:21 -0700 (Thu, 25 Nov 2010) | 5 lines Rename HYPERVISOR_multicall (which performs the multicall hypercall) to _HYPERVISOR_multicall, and create a new HYPERVISOR_multicall function which invokes _HYPERVISOR_multicall and checks that the individual hypercalls all succeeded. r216405 | rwatson | 2010-12-13 05:15:46 -0700 (Mon, 13 Dec 2010) | 7 lines Add options NO_ADAPTIVE_SX to the XENHVM kernel configuration, matching its similar disabling of adaptive mutexes and rwlocks. The existing comment on why this is the case also applies to sx locks. MFC after: 3 days Discussed with: attilio r216437 | gibbs | 2010-12-14 10:23:49 -0700 (Tue, 14 Dec 2010) | 2 lines Remove spurious printf left over from debugging our XenStore support. r216448 | gibbs | 2010-12-14 13:57:40 -0700 (Tue, 14 Dec 2010) | 4 lines Fix a typo in a comment. Noticed by: Attila Nagy r216956 | rwatson | 2011-01-04 07:49:54 -0700 (Tue, 04 Jan 2011) | 8 lines Make "options XENHVM" compile for i386, not just amd64 -- a largely mechanical change. This opens the door for using PV device drivers under Xen HVM on i386, as well as more general harmonisation of i386 and amd64 Xen support in FreeBSD. Reviewed by: cperciva MFC after: 3 weeks r221827 | mav | 2011-05-12 21:40:16 -0600 (Thu, 12 May 2011) | 2 lines Fix msleep() usage in Xen balloon driver to not wake up on every HZ tick. r222975 | gibbs | 2011-06-10 22:59:01 -0600 (Fri, 10 Jun 2011) | 63 lines Monitor and emit events for XenStore changes to XenBus trees of the devices we manage. These changes can be due to writes we make ourselves or due to changes made by the control domain. The goal of these changes is to insure that all state transitions can be detected regardless of their source and to allow common device policies (e.g. "onlined" backend devices) to be centralized in the XenBus bus code. sys/xen/xenbus/xenbusvar.h: sys/xen/xenbus/xenbus.c: sys/xen/xenbus/xenbus_if.m: Add a new method for XenBus drivers "localend_changed". This method is invoked whenever a write is detected to a device's XenBus tree. The default implementation of this method is a no-op. sys/xen/xenbus/xenbus_if.m: sys/dev/xen/netfront/netfront.c: sys/dev/xen/blkfront/blkfront.c: sys/dev/xen/blkback/blkback.c: Change the signature of the "otherend_changed" method. This notification cannot fail, so it should return void. sys/xen/xenbus/xenbusb_back.c: Add "online" device handling to the XenBus Back Bus support code. An online backend device remains active after a front-end detaches as a reconnect is expected to occur in the near future. sys/xen/interface/io/xenbus.h: Add comment block further explaining the meaning and driver responsibilities associated with the XenBus Closed state. sys/xen/xenbus/xenbusb.c: sys/xen/xenbus/xenbusb.h: sys/xen/xenbus/xenbusb_back.c: sys/xen/xenbus/xenbusb_front.c: sys/xen/xenbus/xenbusb_if.m: o Register a XenStore watch against the local XenBus tree for all devices. o Cache the string length of the path to our local tree. o Allow the xenbus front and back drivers to hook/filter both local and otherend watch processing. o Update the device ivar version of "state" when we detect a XenStore update of that node. sys/dev/xen/control/control.c: sys/xen/xenbus/xenbus.c: sys/xen/xenbus/xenbusb.c: sys/xen/xenbus/xenbusb.h: sys/xen/xenbus/xenbusvar.h: sys/xen/xenstore/xenstorevar.h: Allow clients of the XenStore watch mechanism to attach a single uintptr_t worth of client data to the watch. This removes the need to carefully place client watch data within enclosing objects so that a cast or offsetof calculation can be used to convert from watch to enclosing object. Sponsored by: Spectra Logic Corporation MFC after: 1 week r223059 | gibbs | 2011-06-13 14:36:29 -0600 (Mon, 13 Jun 2011) | 36 lines Several enhancements to the Xen block back driver. sys/dev/xen/blkback/blkback.c: o Implement front-end request coalescing. This greatly improves the performance of front-end clients that are unaware of the dynamic request-size/number of requests negotiation available in the FreeBSD backend driver. This required a large restructuring in how this driver records in-flight transactions and how those transactions are mapped into kernel KVA. For example, the driver now includes a mini "KVA manager" that allocates ranges of contiguous KVA to patches of requests that are physically contiguous in the backing store so that a single bio or UIO segment can be used to represent the I/O. o Refuse to open any backend files or devices if the system has yet to mount root. This avoids a panic. o Properly handle "onlined" devices. An "onlined" backend device stays attached to its backing store across front-end disconnections. This feature is intended to reduce latency when a front-end does a hand-off to another driver (e.g. PV aware bootloader to OS kernel) or during a VM reboot. o Harden the driver against a pathological/buggy front-end by carefully vetting front-end XenStore data such as the front-end state. o Add sysctls that report the negotiated number of segments per-request and the number of requests that can be concurrently in flight. Submitted by: kdm Reviewed by: gibbs Sponsored by: Spectra Logic Corporation MFC after: 1 week r225343 | rwatson | 2011-09-02 11:36:01 -0600 (Fri, 02 Sep 2011) | 7 lines Add support for alternative break-to-debugger support on the Xen console. This should help debug boot-time hangs experienced in 9.0-BETA. MFC after: 3 weeks Tested by: sbruno Approved by: re (kib) r225704 | gibbs | 2011-09-20 17:44:34 -0600 (Tue, 20 Sep 2011) | 29 lines Properly handle suspend/resume events in the Xen device framework. Sponsored by: BQ Internet sys/xen/xenbus/xenbusb.c: o In xenbusb_resume(), publish the state transition of the resuming device into XenbusStateIntiailising so that the remote peer can see it. Recording the state locally is not sufficient to trigger a re-connect sequence. o In xenbusb_resume(), defer new-bus resume processing until after the remote peer's XenStore address has been updated. The drivers may need to refer to this information during resume processing. sys/xen/xenbus/xenbusb_back.c: sys/xen/xenbus/xenbusb_front.c: Register xenbusb_resume() rather than bus_generic_resume() as the handler for device_resume events. sys/xen/xenstore/xenstore.c: o Fix grammer in a comment. o In xs_suspend(), pass suspend events on to the child devices (e.g. xenbusb_front/back, that are attached to the XenStore. Approved by: re MFC after: 1 week r225705 | gibbs | 2011-09-20 18:02:44 -0600 (Tue, 20 Sep 2011) | 35 lines Add suspend/resume support to the Xen blkfront driver. Sponsored by: BQ Internet sys/dev/xen/blkfront/block.h: sys/dev/xen/blkfront/blkfront.c: Remove now unused blkif_vdev_t from the blkfront soft. sys/dev/xen/blkfront/blkfront.c: o In blkfront_suspend(), indicate the desire to suspend by changing the softc connected state to SUSPENDED, and then wait for any I/O pending on the remote peer to drain. Cancel suspend processing if I/O does not drain within 30 seconds. o Enable and update blkfront_resume(). Since I/O is drained prior to the suspension of the VM, the complicated recovery process performed by other Xen blkfront implementations is avoided. We simply tear down the connection to our old peer, and then re-connect. o In blkif_initialize(), fix a resource leak and botched return if we cannot allocate shadow memory for our requests. o In blkfront_backend_changed(), correct our response to the XenbusStateInitialised state. This state indicates that our backend peer has published sufficient data for blkfront to publish ring information and other XenStore data, not that a connection can occur. Blkfront now will only perform connection processing in response to the XenbusStateConnected state. This corrects an issue where blkfront connected before the backend was ready during resume processing. Approved by: re MFC after: 1 week r225706 | gibbs | 2011-09-20 18:06:02 -0600 (Tue, 20 Sep 2011) | 11 lines [ Forced commit. Actual changes accidentally included in r225704 ] sys/dev/xen/control/control.c: Fix locking violations in Xen HVM suspend processing and have it perform similar actions to those performed during an ACPI triggered suspend. Sponsored by: BQ Internet Approved by: re MFC after: 1 week r225707 | gibbs | 2011-09-20 18:08:25 -0600 (Tue, 20 Sep 2011) | 21 lines Correct suspend/resume support in the Netfront driver. Sponsored by: BQ Internet sys/dev/xen/netfront/netfront.c: o Implement netfront_suspend(), a specialized suspend handler for the netfront driver. This routine simply disables the carrier so the driver is idle during system suspend processing. o Fix a leak when re-initializing LRO during a link reset. o In netif_release_tx_bufs(), when cleaning up the grant references for our TX ring, use gnttab_end_foreign_access_ref instead of attempting to grant the page again. o In netif_release_tx_bufs(), we do not track mbufs associated with mbuf chains, but instead just free each mbuf directly. Use m_free(), not m_freem(), to avoid double frees of mbufs. o Refactor some code to enhance clarity. Approved by: re MFC after: 1 week r225709 | gibbs | 2011-09-20 18:15:29 -0600 (Tue, 20 Sep 2011) | 19 lines Update netfront so that it queries and honors published back-end features. sys/dev/xen/netfront/netfront.c: o Add xn_query_features() which reads the XenStore and records the TSO, LRO, and chained ring-request support of the backend. o Rename xn_configure_lro() to xn_configure_features() and use this routine to manage the setup of TSO, LRO, and checksum offload. o In create_netdev(), initialize if_capabilities and if_hwassist to the capabilities found on all backends. Delegate configuration of if_capenable and the TSO flag if if_hwassist to xn_configure_features(). Reported by: Hugo Silva (fix inspired by patch provided) Approved by: re MFC after: 1 week r226029 | jkim | 2011-10-04 17:53:47 -0600 (Tue, 04 Oct 2011) | 2 lines Add strnlen() to libkern. r220647 | jkim | 2011-04-14 16:17:39 -0600 (Thu, 14 Apr 2011) | 4 lines Add event handlers for (ACPI) suspend/resume events. Suspend event handlers are invoked right before device drivers go into sleep state and resume event handlers are invoked right after all device drivers are waken up. r230183 | cperciva | 2012-01-15 19:38:45 -0700 (Sun, 15 Jan 2012) | 3 lines Make XENHVM work on i386. The __ffs() function counts bits starting from zero, unlike ffs(3), which starts counting from 1. r230587 | ken | 2012-01-26 09:35:09 -0700 (Thu, 26 Jan 2012) | 38 lines Xen netback driver rewrite. share/man/man4/Makefile, share/man/man4/xnb.4, sys/dev/xen/netback/netback.c, sys/dev/xen/netback/netback_unit_tests.c: Rewrote the netback driver for xen to attach properly via newbus and work properly in both HVM and PVM mode (only HVM is tested). Works with the in-tree FreeBSD netfront driver or the Windows netfront driver from SuSE. Has not been extensively tested with a Linux netfront driver. Does not implement LRO, TSO, or polling. Includes unit tests that may be run through sysctl after compiling with XNB_DEBUG defined. sys/dev/xen/blkback/blkback.c, sys/xen/interface/io/netif.h: Comment elaboration. sys/kern/uipc_mbuf.c: Fix page fault in kernel mode when calling m_print() on a null mbuf. Since m_print() is only used for debugging, there are no performance concerns for extra error checking code. sys/kern/subr_scanf.c: Add the "hh" and "ll" width specifiers from C99 to scanf(). A few callers were already using "ll" even though scanf() was handling it as "l". Submitted by: Alan Somers Submitted by: John Suykerbuyk Sponsored by: Spectra Logic MFC after: 1 week Reviewed by: ken r230916 | ken | 2012-02-02 10:54:35 -0700 (Thu, 02 Feb 2012) | 13 lines Fix the netback driver build for i386. netback.c: Add missing VM includes. xen/xenvar.h, xen/xenpmap.h: Move some XENHVM macros from to on i386 to match the amd64 headers. conf/files: Add netback to the build. Submitted by: jhb MFC after: 3 days r228526 | kevlo | 2011-12-14 23:29:13 -0700 (Wed, 14 Dec 2011) | 2 lines s/timout/timeout r230879 | ken | 2012-02-01 13:19:33 -0700 (Wed, 01 Feb 2012) | 4 lines Add the GSO prefix descriptor define. MFC after: 3 days

MFC: r227843 (partial)

2012-01-29T01:22:48Z

- There's no need to overwrite the default device method with the default one. Interestingly, these are actually the default for quite some time (bus_generic_driver_added(9) since r52045 and bus_generic_print_child(9) since r52045) but even recently added device drivers do this unnecessarily. Discussed with: jhb, marcel - While at it, use DEVMETHOD_END. Discussed with: jhb - Also while at it, use __FBSDID.

MFC r212008, 212028, 212210: Vendor update to version 3.80.06.003 to

2010-09-12T06:19:35Z

fix panic with ZFS under heavy I/O load. PR: kern/149968 Submitted by: LSI (Tom Couch) Reported by: Kai Kockro Tested by: Kai Kockro, jpaetzel

MFC r209860:

2010-07-09T20:44:14Z

Don't use pack() for structures that is used purely for software state. Otherwise the resulting, unaligned mutex structure would trigger panic. Submitted by: Tom Cough Reported/Tested by: jhb

MFC r208969,209268:

2010-06-21T22:00:57Z

Driver update of twa(4) from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch PR: kern/147695

MFC r203108:

2010-02-14T19:38:27Z

Large set of CAM improvements: - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.

We no longer need to use d_thread_t, migrate to struct thread *.

2009-05-20T17:29:21Z

Use si_drv1 instead of dev2unit() in twe(4) and twa(4)

2009-04-14T14:02:43Z

Reviewed by: scottl

Reenable 64-bit DMA for twa(4) controllers, but use a boundary of 4GB to

2009-03-23T19:27:23Z

prevent individual transactions from crossing a 4GB address boundary. Due to bus_size_t type limitations, the driver uses a 2GB boundary in PAE kernels. Reviewed by: scottl MFC after: 1 week

Limit DMA memory to lower addressable 4GB, without this patch, we

2008-10-06T22:34:17Z

can reliably provoke data corruption on systems equipped with a plenty of memory during high load. Reported by: gnn via iXsystems MFC candidate: RELENG_7_1, RELENG_7