diff options
Diffstat (limited to 'sys')
57 files changed, 1155 insertions, 227 deletions
diff --git a/sys/conf/dtb.build.mk b/sys/conf/dtb.build.mk index 327d69106244..7eb0db5e8b80 100644 --- a/sys/conf/dtb.build.mk +++ b/sys/conf/dtb.build.mk @@ -1,7 +1,3 @@ - -.include <bsd.init.mk> -# Grab all the options for a kernel build. For backwards compat, we need to -# do this after bsd.own.mk. .include "kern.opts.mk" DTC?= dtc diff --git a/sys/conf/kern.opts.mk b/sys/conf/kern.opts.mk index 045e55d1b19a..cef4dd11ba58 100644 --- a/sys/conf/kern.opts.mk +++ b/sys/conf/kern.opts.mk @@ -4,6 +4,7 @@ # parts to omit (eg CDDL or SOURCELESS_HOST). Some of these will cause # config.mk to define symbols in various opt_*.h files. + # # Define MK_* variables (which are either "yes" or "no") for users # to set via WITH_*/WITHOUT_* in /etc/src.conf and override in the @@ -13,17 +14,12 @@ # that haven't been converted over. # -# Note: bsd.own.mk must be included before the rest of kern.opts.mk to make -# building on 10.x and earlier work. This should be removed when that's no -# longer supported since it confounds the defaults (since it uses the host's -# notion of defaults rather than what's default in current when building -# within sys/modules). -.include <bsd.own.mk> - # These options are used by the kernel build process (kern.mk and kmod.mk) # They have to be listed here so we can build modules outside of the # src tree. +.include <bsd.init.mk> + KLDXREF_CMD?= kldxref __DEFAULT_YES_OPTIONS = \ diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/sysctl_os.c b/sys/contrib/openzfs/module/os/freebsd/zfs/sysctl_os.c index 393bfaa65ff5..ace2360c032d 100644 --- a/sys/contrib/openzfs/module/os/freebsd/zfs/sysctl_os.c +++ b/sys/contrib/openzfs/module/os/freebsd/zfs/sysctl_os.c @@ -188,6 +188,11 @@ param_set_arc_max(SYSCTL_HANDLER_ARGS) return (0); } +SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_max, + CTLTYPE_ULONG | CTLFLAG_RWTUN | CTLFLAG_MPSAFE, + NULL, 0, param_set_arc_max, "LU", + "Maximum ARC size in bytes (LEGACY)"); + int param_set_arc_min(SYSCTL_HANDLER_ARGS) { @@ -212,6 +217,11 @@ param_set_arc_min(SYSCTL_HANDLER_ARGS) return (0); } +SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_min, + CTLTYPE_ULONG | CTLFLAG_RWTUN | CTLFLAG_MPSAFE, + NULL, 0, param_set_arc_min, "LU", + "Minimum ARC size in bytes (LEGACY)"); + extern uint_t zfs_arc_free_target; int @@ -235,6 +245,16 @@ param_set_arc_free_target(SYSCTL_HANDLER_ARGS) return (0); } +/* + * NOTE: This sysctl is CTLFLAG_RW not CTLFLAG_RWTUN due to its dependency on + * pagedaemon initialization. + */ +SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_free_target, + CTLTYPE_UINT | CTLFLAG_RW | CTLFLAG_MPSAFE, + NULL, 0, param_set_arc_free_target, "IU", + "Desired number of free pages below which ARC triggers reclaim" + " (LEGACY)"); + int param_set_arc_no_grow_shift(SYSCTL_HANDLER_ARGS) { @@ -253,6 +273,187 @@ param_set_arc_no_grow_shift(SYSCTL_HANDLER_ARGS) return (0); } +SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_no_grow_shift, + CTLTYPE_INT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE, + NULL, 0, param_set_arc_no_grow_shift, "I", + "log2(fraction of ARC which must be free to allow growing) (LEGACY)"); + +extern uint64_t l2arc_write_max; + +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_max, + CTLFLAG_RWTUN, &l2arc_write_max, 0, + "Max write bytes per interval (LEGACY)"); + +extern uint64_t l2arc_write_boost; + +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_boost, + CTLFLAG_RWTUN, &l2arc_write_boost, 0, + "Extra write bytes during device warmup (LEGACY)"); + +extern uint64_t l2arc_headroom; + +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_headroom, + CTLFLAG_RWTUN, &l2arc_headroom, 0, + "Number of max device writes to precache (LEGACY)"); + +extern uint64_t l2arc_headroom_boost; + +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_headroom_boost, + CTLFLAG_RWTUN, &l2arc_headroom_boost, 0, + "Compressed l2arc_headroom multiplier (LEGACY)"); + +extern uint64_t l2arc_feed_secs; + +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_secs, + CTLFLAG_RWTUN, &l2arc_feed_secs, 0, + "Seconds between L2ARC writing (LEGACY)"); + +extern uint64_t l2arc_feed_min_ms; + +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_min_ms, + CTLFLAG_RWTUN, &l2arc_feed_min_ms, 0, + "Min feed interval in milliseconds (LEGACY)"); + +extern int l2arc_noprefetch; + +SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_noprefetch, + CTLFLAG_RWTUN, &l2arc_noprefetch, 0, + "Skip caching prefetched buffers (LEGACY)"); + +extern int l2arc_feed_again; + +SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_feed_again, + CTLFLAG_RWTUN, &l2arc_feed_again, 0, + "Turbo L2ARC warmup (LEGACY)"); + +extern int l2arc_norw; + +SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_norw, + CTLFLAG_RWTUN, &l2arc_norw, 0, + "No reads during writes (LEGACY)"); + +static int +param_get_arc_state_size(SYSCTL_HANDLER_ARGS) +{ + arc_state_t *state = (arc_state_t *)arg1; + int64_t val; + + val = zfs_refcount_count(&state->arcs_size[ARC_BUFC_DATA]) + + zfs_refcount_count(&state->arcs_size[ARC_BUFC_METADATA]); + return (sysctl_handle_64(oidp, &val, 0, req)); +} + +extern arc_state_t ARC_anon; + +SYSCTL_PROC(_vfs_zfs, OID_AUTO, anon_size, + CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE, + &ARC_anon, 0, param_get_arc_state_size, "Q", + "size of anonymous state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_metadata_esize, CTLFLAG_RD, + &ARC_anon.arcs_esize[ARC_BUFC_METADATA].rc_count, 0, + "size of evictable metadata in anonymous state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_data_esize, CTLFLAG_RD, + &ARC_anon.arcs_esize[ARC_BUFC_DATA].rc_count, 0, + "size of evictable data in anonymous state"); + +extern arc_state_t ARC_mru; + +SYSCTL_PROC(_vfs_zfs, OID_AUTO, mru_size, + CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE, + &ARC_mru, 0, param_get_arc_state_size, "Q", + "size of mru state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_metadata_esize, CTLFLAG_RD, + &ARC_mru.arcs_esize[ARC_BUFC_METADATA].rc_count, 0, + "size of evictable metadata in mru state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_data_esize, CTLFLAG_RD, + &ARC_mru.arcs_esize[ARC_BUFC_DATA].rc_count, 0, + "size of evictable data in mru state"); + +extern arc_state_t ARC_mru_ghost; + +SYSCTL_PROC(_vfs_zfs, OID_AUTO, mru_ghost_size, + CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE, + &ARC_mru_ghost, 0, param_get_arc_state_size, "Q", + "size of mru ghost state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_metadata_esize, CTLFLAG_RD, + &ARC_mru_ghost.arcs_esize[ARC_BUFC_METADATA].rc_count, 0, + "size of evictable metadata in mru ghost state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_data_esize, CTLFLAG_RD, + &ARC_mru_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0, + "size of evictable data in mru ghost state"); + +extern arc_state_t ARC_mfu; + +SYSCTL_PROC(_vfs_zfs, OID_AUTO, mfu_size, + CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE, + &ARC_mfu, 0, param_get_arc_state_size, "Q", + "size of mfu state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_metadata_esize, CTLFLAG_RD, + &ARC_mfu.arcs_esize[ARC_BUFC_METADATA].rc_count, 0, + "size of evictable metadata in mfu state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_data_esize, CTLFLAG_RD, + &ARC_mfu.arcs_esize[ARC_BUFC_DATA].rc_count, 0, + "size of evictable data in mfu state"); + +extern arc_state_t ARC_mfu_ghost; + +SYSCTL_PROC(_vfs_zfs, OID_AUTO, mfu_ghost_size, + CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE, + &ARC_mfu_ghost, 0, param_get_arc_state_size, "Q", + "size of mfu ghost state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_metadata_esize, CTLFLAG_RD, + &ARC_mfu_ghost.arcs_esize[ARC_BUFC_METADATA].rc_count, 0, + "size of evictable metadata in mfu ghost state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_data_esize, CTLFLAG_RD, + &ARC_mfu_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0, + "size of evictable data in mfu ghost state"); + +extern arc_state_t ARC_uncached; + +SYSCTL_PROC(_vfs_zfs, OID_AUTO, uncached_size, + CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE, + &ARC_uncached, 0, param_get_arc_state_size, "Q", + "size of uncached state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, uncached_metadata_esize, CTLFLAG_RD, + &ARC_uncached.arcs_esize[ARC_BUFC_METADATA].rc_count, 0, + "size of evictable metadata in uncached state"); +SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, uncached_data_esize, CTLFLAG_RD, + &ARC_uncached.arcs_esize[ARC_BUFC_DATA].rc_count, 0, + "size of evictable data in uncached state"); + +extern arc_state_t ARC_l2c_only; + +SYSCTL_PROC(_vfs_zfs, OID_AUTO, l2c_only_size, + CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE, + &ARC_l2c_only, 0, param_get_arc_state_size, "Q", + "size of l2c_only state"); + +/* dbuf.c */ + +/* dmu.c */ + +/* dmu_zfetch.c */ + +SYSCTL_NODE(_vfs_zfs, OID_AUTO, zfetch, CTLFLAG_RW, 0, "ZFS ZFETCH (LEGACY)"); + +extern uint32_t zfetch_max_distance; + +SYSCTL_UINT(_vfs_zfs_zfetch, OID_AUTO, max_distance, + CTLFLAG_RWTUN, &zfetch_max_distance, 0, + "Max bytes to prefetch per stream (LEGACY)"); + +extern uint32_t zfetch_max_idistance; + +SYSCTL_UINT(_vfs_zfs_zfetch, OID_AUTO, max_idistance, + CTLFLAG_RWTUN, &zfetch_max_idistance, 0, + "Max bytes to prefetch indirects for per stream (LEGACY)"); + +/* dsl_pool.c */ + +/* dnode.c */ + +/* dsl_scan.c */ + /* metaslab.c */ int @@ -313,6 +514,19 @@ SYSCTL_UINT(_vfs_zfs, OID_AUTO, condense_pct, "Condense on-disk spacemap when it is more than this many percents" " of in-memory counterpart"); +extern uint_t zfs_remove_max_segment; + +SYSCTL_UINT(_vfs_zfs, OID_AUTO, remove_max_segment, + CTLFLAG_RWTUN, &zfs_remove_max_segment, 0, + "Largest contiguous segment ZFS will attempt to allocate when removing" + " a device"); + +extern int zfs_removal_suspend_progress; + +SYSCTL_INT(_vfs_zfs, OID_AUTO, removal_suspend_progress, + CTLFLAG_RWTUN, &zfs_removal_suspend_progress, 0, + "Ensures certain actions can happen while in the middle of a removal"); + /* * Minimum size which forces the dynamic allocator to change * it's allocation strategy. Once the space map cannot satisfy @@ -535,6 +749,12 @@ param_set_min_auto_ashift(SYSCTL_HANDLER_ARGS) return (0); } +SYSCTL_PROC(_vfs_zfs, OID_AUTO, min_auto_ashift, + CTLTYPE_UINT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE, + &zfs_vdev_min_auto_ashift, sizeof (zfs_vdev_min_auto_ashift), + param_set_min_auto_ashift, "IU", + "Min ashift used when creating new top-level vdev. (LEGACY)"); + int param_set_max_auto_ashift(SYSCTL_HANDLER_ARGS) { @@ -554,6 +774,13 @@ param_set_max_auto_ashift(SYSCTL_HANDLER_ARGS) return (0); } +SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift, + CTLTYPE_UINT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE, + &zfs_vdev_max_auto_ashift, sizeof (zfs_vdev_max_auto_ashift), + param_set_max_auto_ashift, "IU", + "Max ashift used when optimizing for logical -> physical sector size on" + " new top-level vdevs. (LEGACY)"); + /* * Since the DTL space map of a vdev is not expected to have a lot of * entries, we default its block size to 4K. @@ -575,6 +802,23 @@ SYSCTL_INT(_vfs_zfs, OID_AUTO, standard_sm_blksz, CTLFLAG_RDTUN, &zfs_vdev_standard_sm_blksz, 0, "Block size for standard space map. Power of 2 greater than 4096."); +extern int vdev_validate_skip; + +SYSCTL_INT(_vfs_zfs, OID_AUTO, validate_skip, + CTLFLAG_RDTUN, &vdev_validate_skip, 0, + "Enable to bypass vdev_validate()."); + +/* vdev_mirror.c */ + +/* vdev_queue.c */ + +extern uint_t zfs_vdev_max_active; + +SYSCTL_UINT(_vfs_zfs, OID_AUTO, top_maxinflight, + CTLFLAG_RWTUN, &zfs_vdev_max_active, 0, + "The maximum number of I/Os of all types active for each device." + " (LEGACY)"); + /* zio.c */ SYSCTL_INT(_vfs_zfs_zio, OID_AUTO, exclude_metadata, diff --git a/sys/contrib/openzfs/module/zfs/arc.c b/sys/contrib/openzfs/module/zfs/arc.c index 591e2dade59e..b677f90280d7 100644 --- a/sys/contrib/openzfs/module/zfs/arc.c +++ b/sys/contrib/openzfs/module/zfs/arc.c @@ -486,13 +486,13 @@ static taskq_t *arc_flush_taskq; static uint_t zfs_arc_evict_threads = 0; /* The 7 states: */ -static arc_state_t ARC_anon; -/* */ arc_state_t ARC_mru; -static arc_state_t ARC_mru_ghost; -/* */ arc_state_t ARC_mfu; -static arc_state_t ARC_mfu_ghost; -static arc_state_t ARC_l2c_only; -static arc_state_t ARC_uncached; +arc_state_t ARC_anon; +arc_state_t ARC_mru; +arc_state_t ARC_mru_ghost; +arc_state_t ARC_mfu; +arc_state_t ARC_mfu_ghost; +arc_state_t ARC_l2c_only; +arc_state_t ARC_uncached; arc_stats_t arc_stats = { { "hits", KSTAT_DATA_UINT64 }, @@ -832,15 +832,15 @@ typedef struct arc_async_flush { #define L2ARC_FEED_TYPES 4 /* L2ARC Performance Tunables */ -static uint64_t l2arc_write_max = L2ARC_WRITE_SIZE; /* def max write size */ -static uint64_t l2arc_write_boost = L2ARC_WRITE_SIZE; /* extra warmup write */ -static uint64_t l2arc_headroom = L2ARC_HEADROOM; /* # of dev writes */ -static uint64_t l2arc_headroom_boost = L2ARC_HEADROOM_BOOST; -static uint64_t l2arc_feed_secs = L2ARC_FEED_SECS; /* interval seconds */ -static uint64_t l2arc_feed_min_ms = L2ARC_FEED_MIN_MS; /* min interval msecs */ -static int l2arc_noprefetch = B_TRUE; /* don't cache prefetch bufs */ -static int l2arc_feed_again = B_TRUE; /* turbo warmup */ -static int l2arc_norw = B_FALSE; /* no reads during writes */ +uint64_t l2arc_write_max = L2ARC_WRITE_SIZE; /* def max write size */ +uint64_t l2arc_write_boost = L2ARC_WRITE_SIZE; /* extra warmup write */ +uint64_t l2arc_headroom = L2ARC_HEADROOM; /* # of dev writes */ +uint64_t l2arc_headroom_boost = L2ARC_HEADROOM_BOOST; +uint64_t l2arc_feed_secs = L2ARC_FEED_SECS; /* interval seconds */ +uint64_t l2arc_feed_min_ms = L2ARC_FEED_MIN_MS; /* min interval msecs */ +int l2arc_noprefetch = B_TRUE; /* don't cache prefetch bufs */ +int l2arc_feed_again = B_TRUE; /* turbo warmup */ +int l2arc_norw = B_FALSE; /* no reads during writes */ static uint_t l2arc_meta_percent = 33; /* limit on headers size */ /* diff --git a/sys/contrib/openzfs/module/zfs/dmu_zfetch.c b/sys/contrib/openzfs/module/zfs/dmu_zfetch.c index 3d3a9c713568..51165d0bf723 100644 --- a/sys/contrib/openzfs/module/zfs/dmu_zfetch.c +++ b/sys/contrib/openzfs/module/zfs/dmu_zfetch.c @@ -57,19 +57,19 @@ static unsigned int zfetch_max_sec_reap = 2; /* min bytes to prefetch per stream (default 2MB) */ static unsigned int zfetch_min_distance = 2 * 1024 * 1024; /* max bytes to prefetch per stream (default 8MB) */ -static unsigned int zfetch_max_distance = 8 * 1024 * 1024; +unsigned int zfetch_max_distance = 8 * 1024 * 1024; #else /* min bytes to prefetch per stream (default 4MB) */ static unsigned int zfetch_min_distance = 4 * 1024 * 1024; /* max bytes to prefetch per stream (default 64MB) */ -static unsigned int zfetch_max_distance = 64 * 1024 * 1024; +unsigned int zfetch_max_distance = 64 * 1024 * 1024; #endif /* max bytes to prefetch indirects for per stream (default 128MB) */ -static unsigned int zfetch_max_idistance = 128 * 1024 * 1024; +unsigned int zfetch_max_idistance = 128 * 1024 * 1024; /* max request reorder distance within a stream (default 16MB) */ -static unsigned int zfetch_max_reorder = 16 * 1024 * 1024; +unsigned int zfetch_max_reorder = 16 * 1024 * 1024; /* Max log2 fraction of holes in a stream */ -static unsigned int zfetch_hole_shift = 2; +unsigned int zfetch_hole_shift = 2; typedef struct zfetch_stats { kstat_named_t zfetchstat_hits; diff --git a/sys/contrib/openzfs/module/zfs/vdev.c b/sys/contrib/openzfs/module/zfs/vdev.c index 654e034de9e1..c8d7280387a2 100644 --- a/sys/contrib/openzfs/module/zfs/vdev.c +++ b/sys/contrib/openzfs/module/zfs/vdev.c @@ -100,7 +100,7 @@ static uint_t zfs_vdev_default_ms_shift = 29; /* upper limit for metaslab size (16G) */ static uint_t zfs_vdev_max_ms_shift = 34; -static int vdev_validate_skip = B_FALSE; +int vdev_validate_skip = B_FALSE; /* * Since the DTL space map of a vdev is not expected to have a lot of diff --git a/sys/contrib/openzfs/module/zfs/vdev_queue.c b/sys/contrib/openzfs/module/zfs/vdev_queue.c index e69e5598939e..c12713b107bf 100644 --- a/sys/contrib/openzfs/module/zfs/vdev_queue.c +++ b/sys/contrib/openzfs/module/zfs/vdev_queue.c @@ -122,7 +122,7 @@ * The maximum number of i/os active to each device. Ideally, this will be >= * the sum of each queue's max_active. */ -static uint_t zfs_vdev_max_active = 1000; +uint_t zfs_vdev_max_active = 1000; /* * Per-queue limits on the number of i/os active to each device. If the diff --git a/sys/contrib/openzfs/module/zfs/vdev_removal.c b/sys/contrib/openzfs/module/zfs/vdev_removal.c index 2ce0121324ad..2f7a739da241 100644 --- a/sys/contrib/openzfs/module/zfs/vdev_removal.c +++ b/sys/contrib/openzfs/module/zfs/vdev_removal.c @@ -105,7 +105,7 @@ static const uint_t zfs_remove_max_copy_bytes = 64 * 1024 * 1024; * * See also the accessor function spa_remove_max_segment(). */ -static uint_t zfs_remove_max_segment = SPA_MAXBLOCKSIZE; +uint_t zfs_remove_max_segment = SPA_MAXBLOCKSIZE; /* * Ignore hard IO errors during device removal. When set if a device @@ -137,7 +137,7 @@ uint_t vdev_removal_max_span = 32 * 1024; * This is used by the test suite so that it can ensure that certain * actions happen while in the middle of a removal. */ -static int zfs_removal_suspend_progress = 0; +int zfs_removal_suspend_progress = 0; #define VDEV_REMOVAL_ZAP_OBJS "lzap" diff --git a/sys/contrib/openzfs/tests/zfs-tests/include/tunables.cfg b/sys/contrib/openzfs/tests/zfs-tests/include/tunables.cfg index 54b50c9dba77..127ea188f17f 100644 --- a/sys/contrib/openzfs/tests/zfs-tests/include/tunables.cfg +++ b/sys/contrib/openzfs/tests/zfs-tests/include/tunables.cfg @@ -76,8 +76,8 @@ READ_SIT_OUT_SECS vdev.read_sit_out_secs vdev_read_sit_out_secs SIT_OUT_CHECK_INTERVAL vdev.raidz_outlier_check_interval_ms vdev_raidz_outlier_check_interval_ms SIT_OUT_INSENSITIVITY vdev.raidz_outlier_insensitivity vdev_raidz_outlier_insensitivity REBUILD_SCRUB_ENABLED rebuild_scrub_enabled zfs_rebuild_scrub_enabled -REMOVAL_SUSPEND_PROGRESS vdev.removal_suspend_progress zfs_removal_suspend_progress -REMOVE_MAX_SEGMENT vdev.remove_max_segment zfs_remove_max_segment +REMOVAL_SUSPEND_PROGRESS removal_suspend_progress zfs_removal_suspend_progress +REMOVE_MAX_SEGMENT remove_max_segment zfs_remove_max_segment RESILVER_MIN_TIME_MS resilver_min_time_ms zfs_resilver_min_time_ms RESILVER_DEFER_PERCENT resilver_defer_percent zfs_resilver_defer_percent SCAN_LEGACY scan_legacy zfs_scan_legacy diff --git a/sys/dev/ice/ice_common.c b/sys/dev/ice/ice_common.c index ad4ea4c8e7a1..b895f661bc46 100644 --- a/sys/dev/ice/ice_common.c +++ b/sys/dev/ice/ice_common.c @@ -213,6 +213,15 @@ int ice_set_mac_type(struct ice_hw *hw) case ICE_DEV_ID_E830_L_QSFP: case ICE_DEV_ID_E830C_SFP: case ICE_DEV_ID_E830_L_SFP: + case ICE_DEV_ID_E835CC_BACKPLANE: + case ICE_DEV_ID_E835CC_QSFP56: + case ICE_DEV_ID_E835CC_SFP: + case ICE_DEV_ID_E835C_BACKPLANE: + case ICE_DEV_ID_E835C_QSFP: + case ICE_DEV_ID_E835C_SFP: + case ICE_DEV_ID_E835_L_BACKPLANE: + case ICE_DEV_ID_E835_L_QSFP: + case ICE_DEV_ID_E835_L_SFP: hw->mac_type = ICE_MAC_E830; break; default: diff --git a/sys/dev/ice/ice_devids.h b/sys/dev/ice/ice_devids.h index 3f91e9dfbcaf..74712c61ae8e 100644 --- a/sys/dev/ice/ice_devids.h +++ b/sys/dev/ice/ice_devids.h @@ -62,6 +62,24 @@ #define ICE_DEV_ID_E830C_SFP 0x12DA /* Intel(R) Ethernet Controller E830-L for SFP */ #define ICE_DEV_ID_E830_L_SFP 0x12DE +/* Intel(R) Ethernet Controller E835-CC for backplane */ +#define ICE_DEV_ID_E835CC_BACKPLANE 0x1248 +/* Intel(R) Ethernet Controller E835-CC for QSFP */ +#define ICE_DEV_ID_E835CC_QSFP56 0x1249 +/* Intel(R) Ethernet Controller E835-CC for SFP */ +#define ICE_DEV_ID_E835CC_SFP 0x124A +/* Intel(R) Ethernet Controller E835-C for backplane */ +#define ICE_DEV_ID_E835C_BACKPLANE 0x1261 +/* Intel(R) Ethernet Controller E835-C for QSFP */ +#define ICE_DEV_ID_E835C_QSFP 0x1262 +/* Intel(R) Ethernet Controller E835-C for SFP */ +#define ICE_DEV_ID_E835C_SFP 0x1263 +/* Intel(R) Ethernet Controller E835-L for backplane */ +#define ICE_DEV_ID_E835_L_BACKPLANE 0x1265 +/* Intel(R) Ethernet Controller E835-L for QSFP */ +#define ICE_DEV_ID_E835_L_QSFP 0x1266 +/* Intel(R) Ethernet Controller E835-L for SFP */ +#define ICE_DEV_ID_E835_L_SFP 0x1267 /* Intel(R) Ethernet Controller E810-C for backplane */ #define ICE_DEV_ID_E810C_BACKPLANE 0x1591 /* Intel(R) Ethernet Controller E810-C for QSFP */ diff --git a/sys/dev/ice/ice_drv_info.h b/sys/dev/ice/ice_drv_info.h index 2a51a7394424..46965f4124bc 100644 --- a/sys/dev/ice/ice_drv_info.h +++ b/sys/dev/ice/ice_drv_info.h @@ -218,6 +218,45 @@ static const pci_vendor_info_t ice_vendor_info_array[] = { "Intel(R) Ethernet Network Adapter E830-XXV-2"), PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E830_L_SFP, "Intel(R) Ethernet Connection E830-L for SFP"), + PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_BACKPLANE, + "Intel(R) Ethernet Connection E835-CC for backplane"), + PVIDV_OEM(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_QSFP56, + ICE_INTEL_VENDOR_ID, 0x0001, 0, + "Intel(R) Ethernet Network Adapter E835-C-Q2"), + PVIDV_OEM(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_QSFP56, + ICE_INTEL_VENDOR_ID, 0x0002, 0, + "Intel(R) Ethernet Network Adapter E835-C-Q2 for OCP 3.0"), + PVIDV_OEM(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_QSFP56, + ICE_INTEL_VENDOR_ID, 0x0003, 0, + "Intel(R) Ethernet Network Adapter E835-CC-Q1"), + PVIDV_OEM(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_QSFP56, + ICE_INTEL_VENDOR_ID, 0x0004, 0, + "Intel(R) Ethernet Network Adapter E835-CC-Q1 for OCP 3.0"), + PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_QSFP56, + "Intel(R) Ethernet Connection E835-CC for QSFP56"), + PVIDV_OEM(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_SFP, + ICE_INTEL_VENDOR_ID, 0x0001, 0, + "Intel(R) Ethernet Network Adapter E835-XXV-2 for OCP 3.0"), + PVIDV_OEM(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_SFP, + ICE_INTEL_VENDOR_ID, 0x0003, 0, + "Intel(R) Ethernet Network Adapter E835-XXV-2"), + PVIDV_OEM(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_SFP, + ICE_INTEL_VENDOR_ID, 0x0004, 0, + "Intel(R) Ethernet Network Adapter E835-XXV-4 for OCP 3.0"), + PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835CC_SFP, + "Intel(R) Ethernet Connection E835-CC for SFP"), + PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835C_BACKPLANE, + "Intel(R) Ethernet Connection E835-C for backplane"), + PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835C_QSFP, + "Intel(R) Ethernet Connection E835-C for QSFP"), + PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835C_SFP, + "Intel(R) Ethernet Connection E835-C for SFP"), + PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835_L_BACKPLANE, + "Intel(R) Ethernet Connection E835-L for backplane"), + PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835_L_QSFP, + "Intel(R) Ethernet Connection E835-L for QSFP"), + PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E835_L_SFP, + "Intel(R) Ethernet Connection E835-L for SFP"), PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E825C_BACKPLANE, "Intel(R) Ethernet Connection E825-C for backplane"), PVIDV(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E825C_QSFP, diff --git a/sys/dev/ixgbe/if_ix.c b/sys/dev/ixgbe/if_ix.c index 6d08bd49bc04..1d36fd11f368 100644 --- a/sys/dev/ixgbe/if_ix.c +++ b/sys/dev/ixgbe/if_ix.c @@ -192,6 +192,8 @@ static int ixgbe_if_i2c_req(if_ctx_t, struct ifi2creq *); static bool ixgbe_if_needs_restart(if_ctx_t, enum iflib_restart_event); int ixgbe_intr(void *); +static int ixgbe_if_priv_ioctl(if_ctx_t ctx, u_long command, caddr_t data); + /************************************************************************ * Function prototypes ************************************************************************/ @@ -239,6 +241,13 @@ static void ixgbe_setup_vlan_hw_support(if_ctx_t); static void ixgbe_config_gpie(struct ixgbe_softc *); static void ixgbe_config_delay_values(struct ixgbe_softc *); +static void ixgbe_add_debug_sysctls(struct ixgbe_softc *sc); +static void ixgbe_add_debug_dump_sysctls(struct ixgbe_softc *sc); +static int ixgbe_debug_dump_ioctl(struct ixgbe_softc *sc, struct ifdrv *ifd); +static u8 ixgbe_debug_dump_print_cluster(struct ixgbe_softc *sc, + struct sbuf *sbuf, u8 cluster_id); +static int ixgbe_nvm_access_ioctl(struct ixgbe_softc *sc, struct ifdrv *ifd); + /* Sysctl handlers */ static int ixgbe_sysctl_flowcntl(SYSCTL_HANDLER_ARGS); static int ixgbe_sysctl_advertise(SYSCTL_HANDLER_ARGS); @@ -260,6 +269,9 @@ static int ixgbe_sysctl_wol_enable(SYSCTL_HANDLER_ARGS); static int ixgbe_sysctl_wufc(SYSCTL_HANDLER_ARGS); static int ixgbe_sysctl_tso_tcp_flags_mask(SYSCTL_HANDLER_ARGS); +static int ixgbe_sysctl_debug_dump_set_clusters(SYSCTL_HANDLER_ARGS); +static int ixgbe_sysctl_dump_debug_dump(SYSCTL_HANDLER_ARGS); + /* Deferred interrupt tasklets */ static void ixgbe_handle_msf(void *); static void ixgbe_handle_mod(void *); @@ -330,6 +342,7 @@ static device_method_t ixgbe_if_methods[] = { DEVMETHOD(ifdi_get_counter, ixgbe_if_get_counter), DEVMETHOD(ifdi_i2c_req, ixgbe_if_i2c_req), DEVMETHOD(ifdi_needs_restart, ixgbe_if_needs_restart), + DEVMETHOD(ifdi_priv_ioctl, ixgbe_if_priv_ioctl), #ifdef PCI_IOV DEVMETHOD(ifdi_iov_init, ixgbe_if_iov_init), DEVMETHOD(ifdi_iov_uninit, ixgbe_if_iov_uninit), @@ -1015,6 +1028,8 @@ ixgbe_if_attach_pre(if_ctx_t ctx) if (hw->mac.type == ixgbe_mac_E610) ixgbe_init_aci(hw); + sc->do_debug_dump = false; + if (hw->mac.ops.fw_recovery_mode && hw->mac.ops.fw_recovery_mode(hw)) { device_printf(dev, @@ -1395,6 +1410,248 @@ ixgbe_if_needs_restart(if_ctx_t ctx __unused, enum iflib_restart_event event) } /************************************************************************ + * ixgbe_if_priv_ioctl - Ioctl handler for driver + * + * Handler for custom driver specific ioctls + * + * return 0 on success, positive on failure + ************************************************************************/ +static int +ixgbe_if_priv_ioctl(if_ctx_t ctx, u_long command, caddr_t data) +{ + struct ixgbe_softc *sc = iflib_get_softc(ctx); + struct ifdrv *ifd; + device_t dev = sc->dev; + + /* Make sure the command type is valid */ + switch (command) { + case SIOCSDRVSPEC: + case SIOCGDRVSPEC: + /* Accepted commands */ + break; + case SIOCGPRIVATE_0: + /* + * Although we do not support this ioctl command, it's expected + * that iflib will forward it to the IFDI_PRIV_IOCTL handler. + * Do not print a message in this case. + */ + return (ENOTSUP); + default: + /* + * If we get a different command for this function, it's + * definitely unexpected, so log a message indicating what + * command we got for debugging purposes. + */ + device_printf(dev, + "%s: unexpected ioctl command %08lx\n", + __func__, command); + return (EINVAL); + } + + ifd = (struct ifdrv *)data; + + switch (ifd->ifd_cmd) { + case IXGBE_NVM_ACCESS: + IOCTL_DEBUGOUT("ioctl: NVM ACCESS"); + return (ixgbe_nvm_access_ioctl(sc, ifd)); + case IXGBE_DEBUG_DUMP: + IOCTL_DEBUGOUT("ioctl: DEBUG DUMP"); + return (ixgbe_debug_dump_ioctl(sc, ifd)); + default: + IOCTL_DEBUGOUT1( + "ioctl: UNKNOWN SIOC(S|G)DRVSPEC (0x%X) command\n", + (int)ifd->ifd_cmd); + return (EINVAL); + } + + return (0); +} + +/************************************************************************ + * ixgbe_nvm_access_ioctl + * + * Handles an NVM access ioctl request + ************************************************************************/ +static int +ixgbe_nvm_access_ioctl(struct ixgbe_softc *sc, struct ifdrv *ifd) +{ + struct ixgbe_nvm_access_data *data; + struct ixgbe_nvm_access_cmd *cmd; + struct ixgbe_hw *hw = &sc->hw; + size_t ifd_len = ifd->ifd_len; + size_t malloc_len; + device_t dev = sc->dev; + u8 *nvm_buffer; + s32 error = 0; + + /* + * ifioctl forwards SIOCxDRVSPEC to iflib without conducting + * a privilege check. Subsequently, iflib passes the ioctl to the driver + * without verifying privileges. To prevent non-privileged threads from + * accessing this interface, perform a privilege check at this point. + */ + error = priv_check(curthread, PRIV_DRIVER); + if (error) + return (error); + + if (ifd_len < sizeof(*cmd)) { + device_printf(dev, + "%s: ifdrv length is too small. Got %zu, " + "but expected %zu\n", + __func__, ifd_len, sizeof(*cmd)); + return (EINVAL); + } + + if (ifd->ifd_data == NULL) { + device_printf(dev, "%s: No ifd data buffer.\n", + __func__); + return (EINVAL); + } + + malloc_len = max(ifd_len, sizeof(*data) + sizeof(*cmd)); + + nvm_buffer = (u8 *)malloc(malloc_len, M_IXGBE, M_ZERO | M_NOWAIT); + if (!nvm_buffer) + return (ENOMEM); + + /* Copy the NVM access command and data in from user space */ + error = copyin(ifd->ifd_data, nvm_buffer, ifd_len); + if (error) { + device_printf(dev, "%s: Failed to copy data in, error: %d\n", + __func__, error); + goto cleanup_free_nvm_buffer; + } + + /* + * The NVM command structure is immediately followed by data which + * varies in size based on the command. + */ + cmd = (struct ixgbe_nvm_access_cmd *)nvm_buffer; + data = (struct ixgbe_nvm_access_data *) + (nvm_buffer + sizeof(struct ixgbe_nvm_access_cmd)); + + /* Handle the NVM access request */ + error = ixgbe_handle_nvm_access(hw, cmd, data); + if (error) { + device_printf(dev, "%s: NVM access request failed, error %d\n", + __func__, error); + } + + /* Copy the possibly modified contents of the handled request out */ + error = copyout(nvm_buffer, ifd->ifd_data, ifd_len); + if (error) { + device_printf(dev, "%s: Copying response back to " + "user space failed, error %d\n", + __func__, error); + goto cleanup_free_nvm_buffer; + } + +cleanup_free_nvm_buffer: + free(nvm_buffer, M_IXGBE); + return (error); +} + +/************************************************************************ + * ixgbe_debug_dump_ioctl + * + * Makes debug dump of internal FW/HW data. + ************************************************************************/ +static int +ixgbe_debug_dump_ioctl(struct ixgbe_softc *sc, struct ifdrv *ifd) +{ + struct ixgbe_debug_dump_cmd *dd_cmd; + struct ixgbe_hw *hw = &sc->hw; + size_t ifd_len = ifd->ifd_len; + device_t dev = sc->dev; + s32 error = 0; + + if (!(sc->feat_en & IXGBE_FEATURE_DBG_DUMP)) + return (ENODEV); + + /* Data returned from ACI command */ + u16 ret_buf_size = 0; + u16 ret_next_cluster = 0; + u16 ret_next_table = 0; + u32 ret_next_index = 0; + + /* + * ifioctl forwards SIOCxDRVSPEC to iflib without conducting + * a privilege check. Subsequently, iflib passes the ioctl to the driver + * without verifying privileges. To prevent non-privileged threads from + * accessing this interface, perform a privilege check at this point. + */ + error = priv_check(curthread, PRIV_DRIVER); + if (error) + return (error); + + if (ifd_len < sizeof(*dd_cmd)) { + device_printf(dev, + "%s: ifdrv length is too small. Got %zu, " + "but expected %zu\n", + __func__, ifd_len, sizeof(*dd_cmd)); + return (EINVAL); + } + + if (ifd->ifd_data == NULL) { + device_printf(dev, "%s: No ifd data buffer.\n", + __func__); + return (EINVAL); + } + + dd_cmd = (struct ixgbe_debug_dump_cmd *)malloc(ifd_len, M_IXGBE, + M_NOWAIT | M_ZERO); + if (!dd_cmd) { + error = -ENOMEM; + goto out; + } + /* copy data from userspace */ + error = copyin(ifd->ifd_data, dd_cmd, ifd_len); + if (error) { + device_printf(dev, "%s: Failed to copy data in, error: %d\n", + __func__, error); + goto out; + } + + /* ACI command requires buf_size arg to be grater than 0 */ + if (dd_cmd->data_size == 0) { + device_printf(dev, "%s: data_size must be greater than 0\n", + __func__); + error = EINVAL; + goto out; + } + + /* Zero the data buffer memory space */ + memset(dd_cmd->data, 0, ifd_len - sizeof(*dd_cmd)); + + error = ixgbe_aci_get_internal_data(hw, dd_cmd->cluster_id, + dd_cmd->table_id, dd_cmd->offset, dd_cmd->data, dd_cmd->data_size, + &ret_buf_size, &ret_next_cluster, &ret_next_table, &ret_next_index); + if (error) { + device_printf(dev, + "%s: Failed to get internal FW/HW data, error: %d\n", + __func__, error); + goto out; + } + + dd_cmd->cluster_id = ret_next_cluster; + dd_cmd->table_id = ret_next_table; + dd_cmd->offset = ret_next_index; + dd_cmd->data_size = ret_buf_size; + + error = copyout(dd_cmd, ifd->ifd_data, ifd->ifd_len); + if (error) { + device_printf(dev, + "%s: Failed to copy data out, error: %d\n", + __func__, error); + } + +out: + free(dd_cmd, M_IXGBE); + + return (error); +} + +/************************************************************************ * ixgbe_add_media_types ************************************************************************/ static void @@ -2883,6 +3140,264 @@ ixgbe_sysctl_interrupt_rate_handler(SYSCTL_HANDLER_ARGS) } /* ixgbe_sysctl_interrupt_rate_handler */ /************************************************************************ + * ixgbe_debug_dump_print_cluster + ************************************************************************/ +static u8 +ixgbe_debug_dump_print_cluster(struct ixgbe_softc *sc, struct sbuf *sbuf, + u8 cluster_id) +{ + u16 data_buf_size = IXGBE_ACI_MAX_BUFFER_SIZE; + device_t dev = sc->dev; + struct ixgbe_hw *hw = &sc->hw; + const u8 reserved_buf[8] = {}; + int max_aci_calls = 1000; + int error, counter = 0; + u8 *data_buf; + + /* Input parameters / loop variables */ + u16 table_id = 0; + u32 offset = 0; + + /* Data returned from ACI command */ + u16 ret_buf_size = 0; + u16 ret_next_cluster = 0; + u16 ret_next_table = 0; + u32 ret_next_index = 0; + + data_buf = (u8 *)malloc(data_buf_size, M_IXGBE, M_NOWAIT | M_ZERO); + if (!data_buf) + return (0); + + DEBUGOUT2("%s: dumping cluster id (relative) %d\n", + __func__, cluster_id); + + do { + DEBUGOUT3("table_id 0x%04x offset 0x%08x buf_size %d\n", + table_id, offset, data_buf_size); + + error = ixgbe_aci_get_internal_data(hw, cluster_id, table_id, + offset, data_buf, data_buf_size, &ret_buf_size, + &ret_next_cluster, &ret_next_table, &ret_next_index); + if (error) { + device_printf(dev, + "%s: Failed to get internal FW/HW data, error: %d, " + "last aci status: %d\n", + __func__, error, hw->aci.last_status); + break; + } + + DEBUGOUT3("ret_table_id 0x%04x ret_offset 0x%08x " + "ret_buf_size %d\n", + ret_next_table, ret_next_index, ret_buf_size); + + /* Print cluster id */ + u32 print_cluster_id = (u32)cluster_id; + sbuf_bcat(sbuf, &print_cluster_id, sizeof(print_cluster_id)); + /* Print table id */ + u32 print_table_id = (u32)table_id; + sbuf_bcat(sbuf, &print_table_id, sizeof(print_table_id)); + /* Print table length */ + u32 print_table_length = (u32)ret_buf_size; + sbuf_bcat(sbuf, &print_table_length, + sizeof(print_table_length)); + /* Print current offset */ + u32 print_curr_offset = offset; + sbuf_bcat(sbuf, &print_curr_offset, sizeof(print_curr_offset)); + /* Print reserved bytes */ + sbuf_bcat(sbuf, reserved_buf, sizeof(reserved_buf)); + /* Print data */ + sbuf_bcat(sbuf, data_buf, ret_buf_size); + + /* Prepare for the next loop spin */ + memset(data_buf, 0, data_buf_size); + + bool last_index = (ret_next_index == 0xffffffff); + bool last_table = ((ret_next_table == 0xff || + ret_next_table == 0xffff) && + last_index); + + if (last_table) { + /* End of the cluster */ + DEBUGOUT1("End of the cluster ID %d\n", cluster_id); + break; + } else if (last_index) { + /* End of the table */ + table_id = ret_next_table; + offset = 0; + } else { + /* More data left in the table */ + offset = ret_next_index; + } + } while (++counter < max_aci_calls); + + if (counter >= max_aci_calls) + device_printf(dev, "Exceeded nr of ACI calls for cluster %d\n", + cluster_id); + + free(data_buf, M_IXGBE); + + return (++cluster_id); +} /* ixgbe_print_debug_dump_cluster */ + +/************************************************************************ + * ixgbe_sysctl_debug_dump_set_clusters + * + * Sets the cluster to dump from FW when Debug Dump requested. + ************************************************************************/ +static int +ixgbe_sysctl_debug_dump_set_clusters(SYSCTL_HANDLER_ARGS) +{ + struct ixgbe_softc *sc = (struct ixgbe_softc *)arg1; + u32 clusters = sc->debug_dump_cluster_mask; + device_t dev = sc->dev; + int error; + + error = sysctl_handle_32(oidp, &clusters, 0, req); + if ((error) || !req->newptr) + return (error); + + if (clusters & ~(IXGBE_DBG_DUMP_VALID_CLUSTERS_MASK)) { + device_printf(dev, + "%s: Unrecognized parameter: %u\n", + __func__, clusters); + sc->debug_dump_cluster_mask = + IXGBE_ACI_DBG_DUMP_CLUSTER_ID_INVALID; + return (EINVAL); + } + + sc->debug_dump_cluster_mask = clusters; + + return (0); +} /* ixgbe_sysctl_debug_dump_set_clusters */ + +/************************************************************************ + * ixgbe_sysctl_dump_debug_dump + ************************************************************************/ +static int +ixgbe_sysctl_dump_debug_dump(SYSCTL_HANDLER_ARGS) +{ + struct ixgbe_softc *sc = (struct ixgbe_softc *)arg1; + device_t dev = sc->dev; + struct sbuf *sbuf; + int error = 0; + + UNREFERENCED_PARAMETER(arg2); + + if (!sc->do_debug_dump) { + if (req->oldptr == NULL && req->newptr == NULL) { + error = SYSCTL_OUT(req, 0, 0); + return (error); + } + + char input_buf[2] = ""; + error = sysctl_handle_string(oidp, input_buf, + sizeof(input_buf), req); + if ((error) || (req->newptr == NULL)) + return (error); + + if (input_buf[0] == '1') { + if (sc->debug_dump_cluster_mask == + IXGBE_ACI_DBG_DUMP_CLUSTER_ID_INVALID) { + device_printf(dev, + "Debug Dump failed because an invalid " + "cluster was specified.\n"); + return (EINVAL); + } + + sc->do_debug_dump = true; + return (0); + } + + return (EINVAL); + } + + /* Caller just wants the upper bound for size */ + if (req->oldptr == NULL && req->newptr == NULL) { + size_t est_output_len = IXGBE_DBG_DUMP_BASE_SIZE; + if (sc->debug_dump_cluster_mask & 0x2) + est_output_len += IXGBE_DBG_DUMP_BASE_SIZE; + error = SYSCTL_OUT(req, 0, est_output_len); + return (error); + } + + sbuf = sbuf_new_for_sysctl(NULL, NULL, 128, req); + sbuf_clear_flags(sbuf, SBUF_INCLUDENUL); + + DEBUGOUT("FW Debug Dump running...\n"); + + if (sc->debug_dump_cluster_mask) { + for (u8 id = 0; id <= IXGBE_ACI_DBG_DUMP_CLUSTER_ID_MAX; id++) { + if (sc->debug_dump_cluster_mask & BIT(id)) { + DEBUGOUT1("Dumping cluster ID %u...\n", id); + ixgbe_debug_dump_print_cluster(sc, sbuf, id); + } + } + } else { + u8 next_cluster_id = 0; + do { + DEBUGOUT1("Dumping cluster ID %u...\n", + next_cluster_id); + next_cluster_id = ixgbe_debug_dump_print_cluster(sc, + sbuf, next_cluster_id); + } while (next_cluster_id != 0 && + next_cluster_id <= IXGBE_ACI_DBG_DUMP_CLUSTER_ID_MAX); + } + + sbuf_finish(sbuf); + sbuf_delete(sbuf); + + sc->do_debug_dump = false; + + return (error); +} /* ixgbe_sysctl_dump_debug_dump */ + +/************************************************************************ + * ixgbe_add_debug_dump_sysctls + ************************************************************************/ +static void +ixgbe_add_debug_dump_sysctls(struct ixgbe_softc *sc) +{ + struct sysctl_oid_list *debug_list, *dump_list; + struct sysctl_oid *dump_node; + struct sysctl_ctx_list *ctx; + device_t dev = sc->dev; + + ctx = device_get_sysctl_ctx(dev); + debug_list = SYSCTL_CHILDREN(sc->debug_sysctls); + + dump_node = SYSCTL_ADD_NODE(ctx, debug_list, OID_AUTO, "dump", + CTLFLAG_RD, NULL, "Internal FW/HW Dump"); + dump_list = SYSCTL_CHILDREN(dump_node); + + SYSCTL_ADD_PROC(ctx, dump_list, OID_AUTO, "clusters", + CTLTYPE_U32 | CTLFLAG_RW, sc, 0, + ixgbe_sysctl_debug_dump_set_clusters, "SU", + IXGBE_SYSCTL_DESC_DEBUG_DUMP_SET_CLUSTER); + + SYSCTL_ADD_PROC(ctx, dump_list, OID_AUTO, "dump", + CTLTYPE_STRING | CTLFLAG_RW | CTLFLAG_MPSAFE, sc, 0, + ixgbe_sysctl_dump_debug_dump, "", + IXGBE_SYSCTL_DESC_DUMP_DEBUG_DUMP); +} /* ixgbe_add_debug_dump_sysctls */ + +static void +ixgbe_add_debug_sysctls(struct ixgbe_softc *sc) +{ + struct sysctl_oid_list *ctx_list; + struct sysctl_ctx_list *ctx; + device_t dev = sc->dev; + + ctx = device_get_sysctl_ctx(dev); + ctx_list = SYSCTL_CHILDREN(device_get_sysctl_tree(dev)); + + sc->debug_sysctls = SYSCTL_ADD_NODE(ctx, ctx_list, OID_AUTO, "debug", + CTLFLAG_RD, NULL, "Debug Sysctls"); + + if (sc->feat_en & IXGBE_FEATURE_DBG_DUMP) + ixgbe_add_debug_dump_sysctls(sc); +} /* ixgbe_add_debug_sysctls */ + +/************************************************************************ * ixgbe_add_device_sysctls ************************************************************************/ static void @@ -2992,6 +3507,8 @@ ixgbe_add_device_sysctls(if_ctx_t ctx) CTLTYPE_INT | CTLFLAG_RW, sc, 0, ixgbe_sysctl_eee_state, "I", "EEE Power Save State"); } + + ixgbe_add_debug_sysctls(sc); } /* ixgbe_add_device_sysctls */ /************************************************************************ @@ -5182,6 +5699,7 @@ ixgbe_init_device_features(struct ixgbe_softc *sc) break; case ixgbe_mac_E610: sc->feat_cap |= IXGBE_FEATURE_RECOVERY_MODE; + sc->feat_cap |= IXGBE_FEATURE_DBG_DUMP; break; default: break; @@ -5203,6 +5721,9 @@ ixgbe_init_device_features(struct ixgbe_softc *sc) /* Recovery mode */ if (sc->feat_cap & IXGBE_FEATURE_RECOVERY_MODE) sc->feat_en |= IXGBE_FEATURE_RECOVERY_MODE; + /* FW Debug Dump */ + if (sc->feat_cap & IXGBE_FEATURE_DBG_DUMP) + sc->feat_en |= IXGBE_FEATURE_DBG_DUMP; /* Enabled via global sysctl... */ /* Flow Director */ diff --git a/sys/dev/ixgbe/ixgbe.h b/sys/dev/ixgbe/ixgbe.h index 844064bf8543..624b71acabea 100644 --- a/sys/dev/ixgbe/ixgbe.h +++ b/sys/dev/ixgbe/ixgbe.h @@ -46,6 +46,7 @@ #include <sys/module.h> #include <sys/sockio.h> #include <sys/eventhandler.h> +#include <sys/priv.h> #include <net/if.h> #include <net/if_var.h> @@ -475,6 +476,20 @@ struct ixgbe_softc { u32 feat_cap; u32 feat_en; u16 lse_mask; + + struct sysctl_oid *debug_sysctls; + u32 debug_dump_cluster_mask; + bool do_debug_dump; +}; + +struct ixgbe_debug_dump_cmd { + u32 offset; /* offset to read/write from table, in bytes */ + u8 cluster_id; /* also used to get next cluster id */ + u16 table_id; + u16 data_size; /* size of data field, in bytes */ + u16 reserved1; + u32 reserved2; + u8 data[]; }; /* Precision Time Sync (IEEE 1588) defines */ @@ -499,6 +514,43 @@ struct ixgbe_softc { #define IXGBE_PHY_CURRENT_TEMP 0xC820 #define IXGBE_PHY_OVERTEMP_STATUS 0xC830 +/** + * The ioctl command number used by NVM update for accessing the driver for + * NVM access commands. + */ +#define IXGBE_NVM_ACCESS \ + (((((((('E' << 4) + '1') << 4) + 'K') << 4) + 'G') << 4) | 5) + +/* + * The ioctl command number used by a userspace tool for accessing the driver + * for getting debug dump data from the firmware. + */ +#define IXGBE_DEBUG_DUMP \ + (((((((('E' << 4) + '1') << 4) + 'K') << 4) + 'G') << 4) | 6) + +/* Debug Dump related definitions */ +#define IXGBE_ACI_DBG_DUMP_CLUSTER_ID_INVALID 0xFFFFFF +#define IXGBE_ACI_DBG_DUMP_CLUSTER_ID_BASE 50 +#define IXGBE_ACI_DBG_DUMP_CLUSTER_ID_MAX 1 + +#define IXGBE_DBG_DUMP_VALID_CLUSTERS_MASK 0x3 +#define IXGBE_DBG_DUMP_BASE_SIZE (2 * 1024 * 1024) + +#define IXGBE_SYSCTL_DESC_DEBUG_DUMP_SET_CLUSTER \ +"\nSelect clusters to dump with \"dump\" sysctl" \ +"\nFlags:" \ +"\n\t 0x1 - Link" \ +"\n\t 0x2 - Full CSR Space, excluding RCW registers" \ +"\n\t" \ +"\nUse \"sysctl -x\" to view flags properly." + +#define IXGBE_SYSCTL_DESC_DUMP_DEBUG_DUMP \ +"\nWrite 1 to output a FW debug dump containing the clusters " \ +"specified by the \"clusters\" sysctl" \ +"\nThe \"-b\" flag must be used in order to dump this data " \ +"as binary data because" \ +"\nthis data is opaque and not a string." + /* Sysctl help messages; displayed with sysctl -d */ #define IXGBE_SYSCTL_DESC_ADV_SPEED \ "\nControl advertised link speed using these flags:\n" \ diff --git a/sys/dev/ixgbe/ixgbe_features.h b/sys/dev/ixgbe/ixgbe_features.h index 0cef334a185f..bee9040319d8 100644 --- a/sys/dev/ixgbe/ixgbe_features.h +++ b/sys/dev/ixgbe/ixgbe_features.h @@ -57,6 +57,7 @@ #define IXGBE_FEATURE_LEGACY_IRQ (u32)(1 << 12) #define IXGBE_FEATURE_NEEDS_CTXD (u32)(1 << 13) #define IXGBE_FEATURE_RECOVERY_MODE (u32)(1 << 15) +#define IXGBE_FEATURE_DBG_DUMP (u32)(1 << 16) /* Check for OS support. Undefine features if not included in the OS */ #ifndef PCI_IOV diff --git a/sys/dev/nvme/nvme.h b/sys/dev/nvme/nvme.h index 17c5cdb4db87..57cb37907e65 100644 --- a/sys/dev/nvme/nvme.h +++ b/sys/dev/nvme/nvme.h @@ -1507,9 +1507,7 @@ struct nvme_namespace_data { uint8_t eui64[8]; /** lba format support */ - uint32_t lbaf[16]; - - uint8_t reserved7[192]; + uint32_t lbaf[64]; uint8_t vendor_specific[3712]; } __packed __aligned(4); @@ -2175,7 +2173,7 @@ void nvme_namespace_data_swapbytes(struct nvme_namespace_data *s __unused) s->anagrpid = le32toh(s->anagrpid); s->nvmsetid = le16toh(s->nvmsetid); s->endgid = le16toh(s->endgid); - for (i = 0; i < 16; i++) + for (i = 0; i < 64; i++) s->lbaf[i] = le32toh(s->lbaf[i]); #endif } diff --git a/sys/dev/nvme/nvme_private.h b/sys/dev/nvme/nvme_private.h index 52f9e12f8f9a..52e9fcbbebcd 100644 --- a/sys/dev/nvme/nvme_private.h +++ b/sys/dev/nvme/nvme_private.h @@ -463,13 +463,13 @@ static __inline void nvme_completion_poll(struct nvme_completion_poll_status *status) { int timeout = ticks + 10 * hz; - sbintime_t delta_t = SBT_1US; + sbintime_t delta = SBT_1US; while (!atomic_load_acq_int(&status->done)) { if (timeout - ticks < 0) panic("NVME polled command failed to complete within 10s."); - pause_sbt("nvme", delta_t, 0, C_PREL(1)); - delta_t = min(SBT_1MS, delta_t * 3 / 2); + pause_sbt("nvme", delta, 0, C_PREL(1)); + delta = min(SBT_1MS, delta + delta / 2); } } diff --git a/sys/fs/nullfs/null.h b/sys/fs/nullfs/null.h index ad3f7779e108..7bfdc20a3f67 100644 --- a/sys/fs/nullfs/null.h +++ b/sys/fs/nullfs/null.h @@ -35,11 +35,12 @@ #ifndef FS_NULL_H #define FS_NULL_H -#define NULLM_CACHE 0x0001 - #include <sys/ck.h> #include <vm/uma.h> +#define NULLM_CACHE 0x0001 +#define NULLM_NOUNPBYPASS 0x0002 + struct null_mount { struct mount *nullm_vfs; struct vnode *nullm_lowerrootvp; /* Ref to lower root vnode */ @@ -82,6 +83,16 @@ struct vnode *null_checkvp(struct vnode *vp, char *fil, int lno); #endif extern struct vop_vector null_vnodeops; +extern struct vop_vector null_vnodeops_no_unp_bypass; + +static inline bool +null_is_nullfs_vnode(struct vnode *vp) +{ + const struct vop_vector *op; + + op = vp->v_op; + return (op == &null_vnodeops || op == &null_vnodeops_no_unp_bypass); +} extern uma_zone_t null_node_zone; diff --git a/sys/fs/nullfs/null_subr.c b/sys/fs/nullfs/null_subr.c index d7f847d449d0..a843ae44f121 100644 --- a/sys/fs/nullfs/null_subr.c +++ b/sys/fs/nullfs/null_subr.c @@ -240,7 +240,9 @@ null_nodeget(struct mount *mp, struct vnode *lowervp, struct vnode **vpp) */ xp = uma_zalloc_smr(null_node_zone, M_WAITOK); - error = getnewvnode("nullfs", mp, &null_vnodeops, &vp); + error = getnewvnode("nullfs", mp, (MOUNTTONULLMOUNT(mp)->nullm_flags & + NULLM_NOUNPBYPASS) != 0 ? &null_vnodeops_no_unp_bypass : + &null_vnodeops, &vp); if (error) { vput(lowervp); uma_zfree_smr(null_node_zone, xp); diff --git a/sys/fs/nullfs/null_vfsops.c b/sys/fs/nullfs/null_vfsops.c index 4cddf24a5745..170a3dd51cd8 100644 --- a/sys/fs/nullfs/null_vfsops.c +++ b/sys/fs/nullfs/null_vfsops.c @@ -85,6 +85,10 @@ nullfs_mount(struct mount *mp) char *target; int error, len; bool isvnunlocked; + static const char cache_opt_name[] = "cache"; + static const char nocache_opt_name[] = "nocache"; + static const char unixbypass_opt_name[] = "unixbypass"; + static const char nounixbypass_opt_name[] = "nounixbypass"; NULLFSDEBUG("nullfs_mount(mp = %p)\n", (void *)mp); @@ -116,7 +120,7 @@ nullfs_mount(struct mount *mp) /* * Unlock lower node to avoid possible deadlock. */ - if (mp->mnt_vnodecovered->v_op == &null_vnodeops && + if (null_is_nullfs_vnode(mp->mnt_vnodecovered) && VOP_ISLOCKED(mp->mnt_vnodecovered) == LK_EXCLUSIVE) { VOP_UNLOCK(mp->mnt_vnodecovered); isvnunlocked = true; @@ -150,7 +154,7 @@ nullfs_mount(struct mount *mp) /* * Check multi null mount to avoid `lock against myself' panic. */ - if (mp->mnt_vnodecovered->v_op == &null_vnodeops) { + if (null_is_nullfs_vnode(mp->mnt_vnodecovered)) { nn = VTONULL(mp->mnt_vnodecovered); if (nn == NULL || lowerrootvp == nn->null_lowervp) { NULLFSDEBUG("nullfs_mount: multi null mount?\n"); @@ -205,9 +209,10 @@ nullfs_mount(struct mount *mp) MNT_IUNLOCK(mp); } - if (vfs_getopt(mp->mnt_optnew, "cache", NULL, NULL) == 0) { + if (vfs_getopt(mp->mnt_optnew, cache_opt_name, NULL, NULL) == 0) { xmp->nullm_flags |= NULLM_CACHE; - } else if (vfs_getopt(mp->mnt_optnew, "nocache", NULL, NULL) == 0) { + } else if (vfs_getopt(mp->mnt_optnew, nocache_opt_name, NULL, + NULL) == 0) { ; } else if (null_cache_vnodes && (xmp->nullm_vfs->mnt_kern_flag & MNTK_NULL_NOCACHE) == 0) { @@ -219,6 +224,13 @@ nullfs_mount(struct mount *mp) &xmp->notify_node); } + if (vfs_getopt(mp->mnt_optnew, unixbypass_opt_name, NULL, NULL) == 0) { + ; + } else if (vfs_getopt(mp->mnt_optnew, nounixbypass_opt_name, NULL, + NULL) == 0) { + xmp->nullm_flags |= NULLM_NOUNPBYPASS; + } + if (lowerrootvp == mp->mnt_vnodecovered) { vn_lock(lowerrootvp, LK_EXCLUSIVE | LK_RETRY | LK_CANRECURSE); lowerrootvp->v_vflag |= VV_CROSSLOCK; diff --git a/sys/fs/nullfs/null_vnops.c b/sys/fs/nullfs/null_vnops.c index ec8a6b10b13f..d4baabeb40ab 100644 --- a/sys/fs/nullfs/null_vnops.c +++ b/sys/fs/nullfs/null_vnops.c @@ -278,7 +278,7 @@ null_bypass(struct vop_generic_args *ap) * that aren't. (We must always map first vp or vclean fails.) */ if (i != 0 && (*this_vp_p == NULL || - (*this_vp_p)->v_op != &null_vnodeops)) { + !null_is_nullfs_vnode(*this_vp_p))) { old_vps[i] = NULL; } else { old_vps[i] = *this_vp_p; @@ -1256,3 +1256,11 @@ struct vop_vector null_vnodeops = { .vop_copy_file_range = VOP_PANIC, }; VFS_VOP_VECTOR_REGISTER(null_vnodeops); + +struct vop_vector null_vnodeops_no_unp_bypass = { + .vop_default = &null_vnodeops, + .vop_unp_bind = vop_stdunp_bind, + .vop_unp_connect = vop_stdunp_connect, + .vop_unp_detach = vop_stdunp_detach, +}; +VFS_VOP_VECTOR_REGISTER(null_vnodeops_no_unp_bypass); diff --git a/sys/geom/part/g_part.c b/sys/geom/part/g_part.c index 4c0d0c3aa902..1e4236507fa4 100644 --- a/sys/geom/part/g_part.c +++ b/sys/geom/part/g_part.c @@ -122,13 +122,13 @@ struct g_part_alias_list { { "ntfs", G_PART_ALIAS_MS_NTFS }, { "openbsd-data", G_PART_ALIAS_OPENBSD_DATA }, { "prep-boot", G_PART_ALIAS_PREP_BOOT }, - { "solaris-boot", G_PART_ALIAS_SOLARIS_BOOT }, - { "solaris-root", G_PART_ALIAS_SOLARIS_ROOT }, - { "solaris-swap", G_PART_ALIAS_SOLARIS_SWAP }, - { "solaris-backup", G_PART_ALIAS_SOLARIS_BACKUP }, - { "solaris-var", G_PART_ALIAS_SOLARIS_VAR }, - { "solaris-home", G_PART_ALIAS_SOLARIS_HOME }, - { "solaris-altsec", G_PART_ALIAS_SOLARIS_ALTSEC }, + { "solaris-boot", G_PART_ALIAS_SOLARIS_BOOT }, + { "solaris-root", G_PART_ALIAS_SOLARIS_ROOT }, + { "solaris-swap", G_PART_ALIAS_SOLARIS_SWAP }, + { "solaris-backup", G_PART_ALIAS_SOLARIS_BACKUP }, + { "solaris-var", G_PART_ALIAS_SOLARIS_VAR }, + { "solaris-home", G_PART_ALIAS_SOLARIS_HOME }, + { "solaris-altsec", G_PART_ALIAS_SOLARIS_ALTSEC }, { "solaris-reserved", G_PART_ALIAS_SOLARIS_RESERVED }, { "u-boot-env", G_PART_ALIAS_U_BOOT_ENV }, { "vmware-reserved", G_PART_ALIAS_VMRESERVED }, diff --git a/sys/modules/aic7xxx/ahc/Makefile b/sys/modules/aic7xxx/ahc/Makefile index 3741d4fb666f..6f9bdcb1d8bd 100644 --- a/sys/modules/aic7xxx/ahc/Makefile +++ b/sys/modules/aic7xxx/ahc/Makefile @@ -1,6 +1,4 @@ SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${SYSDIR}/dev/aic7xxx KMOD= ahc SUBDIR+= ahc_isa ahc_pci diff --git a/sys/modules/cxgb/Makefile b/sys/modules/cxgb/Makefile index 2989ad580b97..7ebdc1d51945 100644 --- a/sys/modules/cxgb/Makefile +++ b/sys/modules/cxgb/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - SUBDIR= cxgb SUBDIR+= cxgb_t3fw diff --git a/sys/modules/dpdk_lpm4/Makefile b/sys/modules/dpdk_lpm4/Makefile index ff68fac78915..9bc2693aeffb 100644 --- a/sys/modules/dpdk_lpm4/Makefile +++ b/sys/modules/dpdk_lpm4/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${SYSDIR}/contrib/dpdk_rte_lpm KMOD= dpdk_lpm4 diff --git a/sys/modules/dpdk_lpm6/Makefile b/sys/modules/dpdk_lpm6/Makefile index f2248e5d1c1c..9de2c6650422 100644 --- a/sys/modules/dpdk_lpm6/Makefile +++ b/sys/modules/dpdk_lpm6/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${SYSDIR}/contrib/dpdk_rte_lpm KMOD= dpdk_lpm6 diff --git a/sys/modules/fib_dxr/Makefile b/sys/modules/fib_dxr/Makefile index 7d1996ba510f..f8a28abe957a 100644 --- a/sys/modules/fib_dxr/Makefile +++ b/sys/modules/fib_dxr/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${SYSDIR}/netinet KMOD= fib_dxr diff --git a/sys/modules/if_enc/Makefile b/sys/modules/if_enc/Makefile index 449d869d6a21..bd865a0216a4 100644 --- a/sys/modules/if_enc/Makefile +++ b/sys/modules/if_enc/Makefile @@ -1,6 +1,4 @@ SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${SYSDIR}/net KMOD= if_enc diff --git a/sys/modules/if_gif/Makefile b/sys/modules/if_gif/Makefile index efcd6952a8ac..5e3fda3a51c6 100644 --- a/sys/modules/if_gif/Makefile +++ b/sys/modules/if_gif/Makefile @@ -1,6 +1,4 @@ SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${SYSDIR}/net ${SYSDIR}/netinet ${SYSDIR}/netinet6 KMOD= if_gif diff --git a/sys/modules/if_gre/Makefile b/sys/modules/if_gre/Makefile index 9f50708a14d7..58bd03c23785 100644 --- a/sys/modules/if_gre/Makefile +++ b/sys/modules/if_gre/Makefile @@ -1,6 +1,5 @@ SYSDIR?=${SRCTOP}/sys .PATH: ${SYSDIR}/net ${SYSDIR}/netinet ${SYSDIR}/netinet6 -.include "${SYSDIR}/conf/kern.opts.mk" KMOD= if_gre SRCS= if_gre.c opt_inet.h opt_inet6.h opt_rss.h diff --git a/sys/modules/iser/Makefile b/sys/modules/iser/Makefile index 615199ec97a3..ff08ae6f346a 100644 --- a/sys/modules/iser/Makefile +++ b/sys/modules/iser/Makefile @@ -1,6 +1,4 @@ SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${SYSDIR}/dev/iser/ KMOD= iser diff --git a/sys/modules/ktest/Makefile b/sys/modules/ktest/Makefile index 151db53417df..a3052efa9ed9 100644 --- a/sys/modules/ktest/Makefile +++ b/sys/modules/ktest/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - SUBDIR= ktest \ ktest_example \ ktest_netlink_message_writer diff --git a/sys/modules/ktest/ktest/Makefile b/sys/modules/ktest/ktest/Makefile index 3d4f1a8c2cc0..9741662ef709 100644 --- a/sys/modules/ktest/ktest/Makefile +++ b/sys/modules/ktest/ktest/Makefile @@ -1,9 +1,5 @@ PACKAGE= tests - -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - -.PATH: ${SYSDIR}/tests +.PATH: ${SRCTOP}/sys/tests KMOD= ktest SRCS= ktest.c diff --git a/sys/modules/ktest/ktest_example/Makefile b/sys/modules/ktest/ktest_example/Makefile index 2b572d867aa5..aacc8f0e4ca5 100644 --- a/sys/modules/ktest/ktest_example/Makefile +++ b/sys/modules/ktest/ktest_example/Makefile @@ -1,9 +1,8 @@ PACKAGE= tests -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" +.include "${SRCTOP}/sys/conf/kern.opts.mk" -.PATH: ${SYSDIR}/tests +.PATH: ${SRCTOP}/sys/tests KMOD= ktest_example SRCS= ktest_example.c diff --git a/sys/modules/ktest/ktest_netlink_message_writer/Makefile b/sys/modules/ktest/ktest_netlink_message_writer/Makefile index a91c45755d0d..3f05f9b26785 100644 --- a/sys/modules/ktest/ktest_netlink_message_writer/Makefile +++ b/sys/modules/ktest/ktest_netlink_message_writer/Makefile @@ -1,8 +1,6 @@ PACKAGE= tests SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${SYSDIR}/netlink KMOD= ktest_netlink_message_writer diff --git a/sys/modules/miiproxy/Makefile b/sys/modules/miiproxy/Makefile index 730bef4220cd..ab92ebe71b43 100644 --- a/sys/modules/miiproxy/Makefile +++ b/sys/modules/miiproxy/Makefile @@ -3,7 +3,7 @@ KMOD = miiproxy SRCS= miiproxy.c -SRCS+= bus_if.h mdio_if.h miibus_if.h opt_platform.h +SRCS+= bus_if.h device_if.h mdio_if.h miibus_if.h opt_platform.h CFLAGS+= -I${SRCTOP}/sys/dev/etherswitch .include <bsd.kmod.mk> diff --git a/sys/modules/netgraph/Makefile b/sys/modules/netgraph/Makefile index 94560d5c51d7..b2d65af16e7f 100644 --- a/sys/modules/netgraph/Makefile +++ b/sys/modules/netgraph/Makefile @@ -1,5 +1,3 @@ -# $Whistle: Makefile,v 1.5 1999/01/24 06:48:37 archie Exp $ - SYSDIR?=${SRCTOP}/sys .include "${SYSDIR}/conf/kern.opts.mk" diff --git a/sys/modules/netgraph/checksum/Makefile b/sys/modules/netgraph/checksum/Makefile index 4e2b1f547a40..bbbc7363d045 100644 --- a/sys/modules/netgraph/checksum/Makefile +++ b/sys/modules/netgraph/checksum/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - KMOD= ng_checksum SRCS= ng_checksum.c opt_inet.h opt_inet6.h diff --git a/sys/modules/netmap/Makefile b/sys/modules/netmap/Makefile index 17b52aec1893..8c114ac51538 100644 --- a/sys/modules/netmap/Makefile +++ b/sys/modules/netmap/Makefile @@ -2,9 +2,6 @@ # Compile netmap as a module, useful if you want a netmap bridge # or loadable drivers. -.include <bsd.own.mk> # FreeBSD 10 and earlier -# .include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${.CURDIR}/../../dev/netmap .PATH.h: ${.CURDIR}/../../net CFLAGS += -I${.CURDIR}/../../ -D INET -D VIMAGE diff --git a/sys/modules/opensolaris/Makefile b/sys/modules/opensolaris/Makefile index 98f52057e45e..7e2d5f9101ad 100644 --- a/sys/modules/opensolaris/Makefile +++ b/sys/modules/opensolaris/Makefile @@ -1,4 +1,4 @@ -SYSDIR?= ${SRCTOP}/sys +SYSDIR?=${SRCTOP}/sys .PATH: ${SYSDIR}/cddl/compat/opensolaris/kern .PATH: ${SYSDIR}/contrib/openzfs/module/os/freebsd/spl diff --git a/sys/modules/ow/Makefile b/sys/modules/ow/Makefile index 76fefe3e63be..7aa9d2de8183 100644 --- a/sys/modules/ow/Makefile +++ b/sys/modules/ow/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - SUBDIR = ow owc ow_temp .include <bsd.subdir.mk> diff --git a/sys/modules/qlnx/Makefile b/sys/modules/qlnx/Makefile index 2121f9d586a6..291b681c809e 100644 --- a/sys/modules/qlnx/Makefile +++ b/sys/modules/qlnx/Makefile @@ -31,9 +31,6 @@ # # -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - SUBDIR=qlnxe SUBDIR+=qlnxev SUBDIR+=qlnxr diff --git a/sys/modules/rtwn/Makefile b/sys/modules/rtwn/Makefile index 9afdd2084ecb..f15cbbe8236b 100644 --- a/sys/modules/rtwn/Makefile +++ b/sys/modules/rtwn/Makefile @@ -1,7 +1,5 @@ .PATH: ${SRCTOP}/sys/dev/rtwn - -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" +.include "${SRCTOP}/sys/conf/kern.opts.mk" KMOD = rtwn SRCS = if_rtwn.c if_rtwn_tx.c if_rtwn_rx.c if_rtwn_beacon.c \ diff --git a/sys/modules/rtwn_pci/Makefile b/sys/modules/rtwn_pci/Makefile index ce2144121e88..3fea80d7d256 100644 --- a/sys/modules/rtwn_pci/Makefile +++ b/sys/modules/rtwn_pci/Makefile @@ -1,7 +1,5 @@ .PATH: ${SRCTOP}/sys/dev/rtwn/pci - -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" +.include "${SRCTOP}/sys/conf/kern.opts.mk" KMOD = if_rtwn_pci SRCS = rtwn_pci_attach.c rtwn_pci_reg.c rtwn_pci_rx.c rtwn_pci_tx.c \ diff --git a/sys/modules/rtwn_usb/Makefile b/sys/modules/rtwn_usb/Makefile index 16899b8a8c49..6a73276d088c 100644 --- a/sys/modules/rtwn_usb/Makefile +++ b/sys/modules/rtwn_usb/Makefile @@ -1,7 +1,5 @@ .PATH: ${SRCTOP}/sys/dev/rtwn/usb - -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" +.include "${SRCTOP}/sys/conf/kern.opts.mk" KMOD = if_rtwn_usb SRCS = rtwn_usb_attach.c rtwn_usb_ep.c rtwn_usb_reg.c rtwn_usb_rx.c \ diff --git a/sys/modules/sound/driver/Makefile b/sys/modules/sound/driver/Makefile index ff9499fdf841..02703d4b591a 100644 --- a/sys/modules/sound/driver/Makefile +++ b/sys/modules/sound/driver/Makefile @@ -1,5 +1,4 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" +.include "${SRCTOP}/sys/conf/kern.opts.mk" # Modules that include binary-only blobs of microcode should be selectable by # MK_SOURCELESS_UCODE option (see below). diff --git a/sys/modules/sound/sound/Makefile b/sys/modules/sound/sound/Makefile index f3978e9bd9cc..169b1a2730ec 100644 --- a/sys/modules/sound/sound/Makefile +++ b/sys/modules/sound/sound/Makefile @@ -1,5 +1,4 @@ SYSDIR?=${SRCTOP}/sys - .PATH: ${SYSDIR}/dev/sound .PATH: ${SYSDIR}/dev/sound/pcm .PATH: ${SYSDIR}/dev/sound/midi diff --git a/sys/modules/tests/fib_lookup/Makefile b/sys/modules/tests/fib_lookup/Makefile index 7d6198396911..b78d4309f145 100644 --- a/sys/modules/tests/fib_lookup/Makefile +++ b/sys/modules/tests/fib_lookup/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - .PATH: ${SYSDIR}/tests/fib_lookup KMOD= test_lookup diff --git a/sys/modules/vnic/Makefile b/sys/modules/vnic/Makefile index 7b975bfebe81..53e208328159 100644 --- a/sys/modules/vnic/Makefile +++ b/sys/modules/vnic/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - CFLAGS+= -DFDT SUBDIR = mrmlbus thunder_mdio thunder_bgx vnicpf vnicvf diff --git a/sys/modules/vnic/mrmlbus/Makefile b/sys/modules/vnic/mrmlbus/Makefile index a3581b7a79a5..a8fe9e5474e1 100644 --- a/sys/modules/vnic/mrmlbus/Makefile +++ b/sys/modules/vnic/mrmlbus/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - S= ${SRCTOP}/sys .PATH: $S/dev/vnic diff --git a/sys/modules/vnic/thunder_bgx/Makefile b/sys/modules/vnic/thunder_bgx/Makefile index 90df4b25df90..bf46c3194493 100644 --- a/sys/modules/vnic/thunder_bgx/Makefile +++ b/sys/modules/vnic/thunder_bgx/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - S= ${SRCTOP}/sys .PATH: $S/dev/vnic diff --git a/sys/modules/vnic/thunder_mdio/Makefile b/sys/modules/vnic/thunder_mdio/Makefile index 37032516f3ca..07cc583bfaf8 100644 --- a/sys/modules/vnic/thunder_mdio/Makefile +++ b/sys/modules/vnic/thunder_mdio/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - S= ${SRCTOP}/sys .PATH: $S/dev/vnic diff --git a/sys/modules/vnic/vnicpf/Makefile b/sys/modules/vnic/vnicpf/Makefile index 37cd29e6fdd8..3cd64d08a788 100644 --- a/sys/modules/vnic/vnicpf/Makefile +++ b/sys/modules/vnic/vnicpf/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - S= ${SRCTOP}/sys .PATH: $S/dev/vnic diff --git a/sys/modules/vnic/vnicvf/Makefile b/sys/modules/vnic/vnicvf/Makefile index c6ffaaa2c302..da938b7fd073 100644 --- a/sys/modules/vnic/vnicvf/Makefile +++ b/sys/modules/vnic/vnicvf/Makefile @@ -1,6 +1,3 @@ -SYSDIR?=${SRCTOP}/sys -.include "${SYSDIR}/conf/kern.opts.mk" - S= ${SRCTOP}/sys .PATH: $S/dev/vnic diff --git a/sys/netinet/tcp_lro.c b/sys/netinet/tcp_lro.c index 64efa4bf060f..9b5baf115855 100644 --- a/sys/netinet/tcp_lro.c +++ b/sys/netinet/tcp_lro.c @@ -1475,10 +1475,11 @@ tcp_lro_queue_mbuf(struct lro_ctrl *lc, struct mbuf *mb) } /* create sequence number */ - lc->lro_mbuf_data[lc->lro_mbuf_count].seq = - (((uint64_t)M_HASHTYPE_GET(mb)) << 56) | - (((uint64_t)mb->m_pkthdr.flowid) << 24) | - ((uint64_t)lc->lro_mbuf_count); + lc->lro_mbuf_data[lc->lro_mbuf_count].seq = lc->lro_mbuf_count; + if (M_HASHTYPE_ISHASH(mb)) + lc->lro_mbuf_data[lc->lro_mbuf_count].seq |= + (((uint64_t)M_HASHTYPE_GET(mb)) << 56) | + (((uint64_t)mb->m_pkthdr.flowid) << 24); /* enter mbuf */ lc->lro_mbuf_data[lc->lro_mbuf_count].mb = mb; diff --git a/sys/vm/uma_core.c b/sys/vm/uma_core.c index 679b2e20e88b..b80b5cc781f7 100644 --- a/sys/vm/uma_core.c +++ b/sys/vm/uma_core.c @@ -4009,21 +4009,15 @@ restart: /* * Use the keg's policy if upper layers haven't already specified a * domain (as happens with first-touch zones). - * - * To avoid races we run the iterator with the keg lock held, but that - * means that we cannot allow the vm_domainset layer to sleep. Thus, - * clear M_WAITOK and handle low memory conditions locally. */ rr = rdomain == UMA_ANYDOMAIN; + aflags = flags; if (rr) { - aflags = (flags & ~M_WAITOK) | M_NOWAIT; if (vm_domainset_iter_policy_ref_init(&di, &keg->uk_dr, &domain, &aflags) != 0) return (NULL); - } else { - aflags = flags; + } else domain = rdomain; - } for (;;) { slab = keg_fetch_free_slab(keg, domain, rr, flags); @@ -4053,13 +4047,8 @@ restart: if ((flags & M_WAITOK) == 0) break; vm_wait_domain(domain); - } else if (vm_domainset_iter_policy(&di, &domain) != 0) { - if ((flags & M_WAITOK) != 0) { - vm_wait_doms(&keg->uk_dr.dr_policy->ds_mask, 0); - goto restart; - } + } else if (vm_domainset_iter_policy(&di, &domain) != 0) break; - } } /* @@ -5245,7 +5234,7 @@ uma_prealloc(uma_zone_t zone, int items) KEG_GET(zone, keg); slabs = howmany(items, keg->uk_ipers); while (slabs-- > 0) { - aflags = M_NOWAIT; + aflags = M_WAITOK; if (vm_domainset_iter_policy_ref_init(&di, &keg->uk_dr, &domain, &aflags) != 0) panic("%s: Domainset is empty", __func__); @@ -5266,7 +5255,8 @@ uma_prealloc(uma_zone_t zone, int items) break; } if (vm_domainset_iter_policy(&di, &domain) != 0) - vm_wait_doms(&keg->uk_dr.dr_policy->ds_mask, 0); + panic("%s: Cannot allocate from any domain", + __func__); } } } diff --git a/sys/x86/x86/mca.c b/sys/x86/x86/mca.c index 4b40f343ac90..735efe307215 100644 --- a/sys/x86/x86/mca.c +++ b/sys/x86/x86/mca.c @@ -46,9 +46,11 @@ #include <sys/malloc.h> #include <sys/mutex.h> #include <sys/proc.h> +#include <sys/sbuf.h> #include <sys/sched.h> #include <sys/smp.h> #include <sys/sysctl.h> +#include <sys/syslog.h> #include <sys/systm.h> #include <sys/taskqueue.h> #include <machine/intr_machdep.h> @@ -135,6 +137,11 @@ SYSCTL_INT(_hw_mca, OID_AUTO, fake_bank, CTLFLAG_RW, "Bank to use for artificial MCAs (testing purpose only)"); #endif +static bool mca_uselog = false; +SYSCTL_BOOL(_hw_mca, OID_AUTO, uselog, CTLFLAG_RWTUN, &mca_uselog, 0, + "Should the system send non-fatal machine check errors to the log " + "(instead of the console)?"); + static STAILQ_HEAD(, mca_internal) mca_freelist; static int mca_freecount; static STAILQ_HEAD(, mca_internal) mca_records; @@ -147,12 +154,40 @@ static struct timeout_task mca_scan_task; static struct mtx mca_lock; static bool mca_startup_done = false; -/* Statistics on number of MCA events by type, updated atomically. */ +/* Static buffer to compose messages while in an interrupt context. */ +static char mca_msg_buf[1024]; +static struct mtx mca_msg_buf_lock; + +/* Statistics on number of MCA events by type, updated with the mca_lock. */ static uint64_t mca_stats[MCA_T_COUNT]; SYSCTL_OPAQUE(_hw_mca, OID_AUTO, stats, CTLFLAG_RD | CTLFLAG_SKIP, mca_stats, MCA_T_COUNT * sizeof(mca_stats[0]), "S", "Array of MCA events by type"); +/* Variables to track and control message rate limiting. */ +static struct timeval mca_last_log_time; +static struct timeval mca_log_interval; +static int mca_log_skipped; + +static int +sysctl_mca_log_interval(SYSCTL_HANDLER_ARGS) +{ + int error; + u_int val; + + val = mca_log_interval.tv_sec; + error = sysctl_handle_int(oidp, &val, 0, req); + if (error != 0 || req->newptr == NULL) + return (error); + mca_log_interval.tv_sec = val; + return (0); +} +SYSCTL_PROC(_hw_mca, OID_AUTO, log_interval, + CTLTYPE_UINT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE, &mca_log_interval, 0, + sysctl_mca_log_interval, "IU", + "Minimum number of seconds between logging correctable MCAs" + " (0 = no limit)"); + static unsigned int mca_ia32_ctl_reg(int bank) { @@ -448,98 +483,111 @@ mca_mute(const struct mca_record *rec) /* Dump details about a single machine check. */ static void -mca_log(const struct mca_record *rec) +mca_log(enum scan_mode mode, const struct mca_record *rec, bool fatal) { + int error, numskipped; uint16_t mca_error; enum mca_stat_types event_type; + struct sbuf sb; + bool uncor, using_shared_buf; if (mca_mute(rec)) return; - if (!log_corrected && (rec->mr_status & MC_STATUS_UC) == 0 && - (!tes_supported(rec->mr_mcg_cap) || + uncor = (rec->mr_status & MC_STATUS_UC) != 0; + + if (!log_corrected && !uncor && (!tes_supported(rec->mr_mcg_cap) || ((rec->mr_status & MC_STATUS_TES_STATUS) >> 53) != 0x2)) return; - printf("MCA: Bank %d, Status 0x%016llx\n", rec->mr_bank, + /* Try to use an allocated buffer when not in an interrupt context. */ + if (mode == POLLED && sbuf_new(&sb, NULL, 512, SBUF_AUTOEXTEND) != NULL) + using_shared_buf = false; + else { + using_shared_buf = true; + mtx_lock_spin(&mca_msg_buf_lock); + sbuf_new(&sb, mca_msg_buf, sizeof(mca_msg_buf), SBUF_FIXEDLEN); + } + + sbuf_printf(&sb, "MCA: Bank %d, Status 0x%016llx\n", rec->mr_bank, (long long)rec->mr_status); - printf("MCA: Global Cap 0x%016llx, Status 0x%016llx\n", + sbuf_printf(&sb, "MCA: Global Cap 0x%016llx, Status 0x%016llx\n", (long long)rec->mr_mcg_cap, (long long)rec->mr_mcg_status); - printf("MCA: Vendor \"%s\", ID 0x%x, APIC ID %d\n", cpu_vendor, - rec->mr_cpu_id, rec->mr_apic_id); - printf("MCA: CPU %d ", rec->mr_cpu); + sbuf_printf(&sb, "MCA: Vendor \"%s\", ID 0x%x, APIC ID %d\n", + cpu_vendor, rec->mr_cpu_id, rec->mr_apic_id); + sbuf_printf(&sb, "MCA: CPU %d ", rec->mr_cpu); if (rec->mr_status & MC_STATUS_UC) - printf("UNCOR "); + sbuf_printf(&sb, "UNCOR "); else { - printf("COR "); + sbuf_printf(&sb, "COR "); if (cmci_supported(rec->mr_mcg_cap)) - printf("(%lld) ", ((long long)rec->mr_status & + sbuf_printf(&sb, "(%lld) ", ((long long)rec->mr_status & MC_STATUS_COR_COUNT) >> 38); if (tes_supported(rec->mr_mcg_cap)) { switch ((rec->mr_status & MC_STATUS_TES_STATUS) >> 53) { case 0x1: - printf("(Green) "); + sbuf_printf(&sb, "(Green) "); break; case 0x2: - printf("(Yellow) "); + sbuf_printf(&sb, "(Yellow) "); break; } } } if (rec->mr_status & MC_STATUS_EN) - printf("EN "); + sbuf_printf(&sb, "EN "); if (rec->mr_status & MC_STATUS_PCC) - printf("PCC "); + sbuf_printf(&sb, "PCC "); if (ser_supported(rec->mr_mcg_cap)) { if (rec->mr_status & MC_STATUS_S) - printf("S "); + sbuf_printf(&sb, "S "); if (rec->mr_status & MC_STATUS_AR) - printf("AR "); + sbuf_printf(&sb, "AR "); } if (rec->mr_status & MC_STATUS_OVER) - printf("OVER "); + sbuf_printf(&sb, "OVER "); mca_error = rec->mr_status & MC_STATUS_MCA_ERROR; event_type = MCA_T_COUNT; switch (mca_error) { /* Simple error codes. */ case 0x0000: - printf("no error"); + sbuf_printf(&sb, "no error"); event_type = MCA_T_NONE; break; case 0x0001: - printf("unclassified error"); + sbuf_printf(&sb, "unclassified error"); event_type = MCA_T_UNCLASSIFIED; break; case 0x0002: - printf("ucode ROM parity error"); + sbuf_printf(&sb, "ucode ROM parity error"); event_type = MCA_T_UCODE_ROM_PARITY; break; case 0x0003: - printf("external error"); + sbuf_printf(&sb, "external error"); event_type = MCA_T_EXTERNAL; break; case 0x0004: - printf("FRC error"); + sbuf_printf(&sb, "FRC error"); event_type = MCA_T_FRC; break; case 0x0005: - printf("internal parity error"); + sbuf_printf(&sb, "internal parity error"); event_type = MCA_T_INTERNAL_PARITY; break; case 0x0006: - printf("SMM handler code access violation"); + sbuf_printf(&sb, "SMM handler code access violation"); event_type = MCA_T_SMM_HANDLER; break; case 0x0400: - printf("internal timer error"); + sbuf_printf(&sb, "internal timer error"); event_type = MCA_T_INTERNAL_TIMER; break; case 0x0e0b: - printf("generic I/O error"); + sbuf_printf(&sb, "generic I/O error"); event_type = MCA_T_GENERIC_IO; if (rec->mr_cpu_vendor_id == CPU_VENDOR_INTEL && (rec->mr_status & MC_STATUS_MISCV)) { - printf(" (pci%d:%d:%d:%d)", + sbuf_printf(&sb, " (pci%d:%d:%d:%d)", (int)((rec->mr_misc & MC_MISC_PCIE_SEG) >> 32), (int)((rec->mr_misc & MC_MISC_PCIE_BUS) >> 24), (int)((rec->mr_misc & MC_MISC_PCIE_SLOT) >> 19), @@ -548,7 +596,8 @@ mca_log(const struct mca_record *rec) break; default: if ((mca_error & 0xfc00) == 0x0400) { - printf("internal error %x", mca_error & 0x03ff); + sbuf_printf(&sb, "internal error %x", + mca_error & 0x03ff); event_type = MCA_T_INTERNAL; break; } @@ -557,14 +606,16 @@ mca_log(const struct mca_record *rec) /* Memory hierarchy error. */ if ((mca_error & 0xeffc) == 0x000c) { - printf("%s memory error", mca_error_level(mca_error)); + sbuf_printf(&sb, "%s memory error", + mca_error_level(mca_error)); event_type = MCA_T_MEMORY; break; } /* TLB error. */ if ((mca_error & 0xeff0) == 0x0010) { - printf("%sTLB %s error", mca_error_ttype(mca_error), + sbuf_printf(&sb, "%sTLB %s error", + mca_error_ttype(mca_error), mca_error_level(mca_error)); event_type = MCA_T_TLB; break; @@ -572,19 +623,19 @@ mca_log(const struct mca_record *rec) /* Memory controller error. */ if ((mca_error & 0xef80) == 0x0080) { - printf("%s channel ", mca_error_mmtype(mca_error, - &event_type)); + sbuf_printf(&sb, "%s channel ", + mca_error_mmtype(mca_error, &event_type)); if ((mca_error & 0x000f) != 0x000f) - printf("%d", mca_error & 0x000f); + sbuf_printf(&sb, "%d", mca_error & 0x000f); else - printf("??"); - printf(" memory error"); + sbuf_printf(&sb, "??"); + sbuf_printf(&sb, " memory error"); break; } /* Cache error. */ if ((mca_error & 0xef00) == 0x0100) { - printf("%sCACHE %s %s error", + sbuf_printf(&sb, "%sCACHE %s %s error", mca_error_ttype(mca_error), mca_error_level(mca_error), mca_error_request(mca_error)); @@ -594,77 +645,129 @@ mca_log(const struct mca_record *rec) /* Extended memory error. */ if ((mca_error & 0xef80) == 0x0280) { - printf("%s channel ", mca_error_mmtype(mca_error, - &event_type)); + sbuf_printf(&sb, "%s channel ", + mca_error_mmtype(mca_error, &event_type)); if ((mca_error & 0x000f) != 0x000f) - printf("%d", mca_error & 0x000f); + sbuf_printf(&sb, "%d", mca_error & 0x000f); else - printf("??"); - printf(" extended memory error"); + sbuf_printf(&sb, "??"); + sbuf_printf(&sb, " extended memory error"); break; } /* Bus and/or Interconnect error. */ if ((mca_error & 0xe800) == 0x0800) { - printf("BUS%s ", mca_error_level(mca_error)); + sbuf_printf(&sb, "BUS%s ", mca_error_level(mca_error)); event_type = MCA_T_BUS; switch ((mca_error & 0x0600) >> 9) { case 0: - printf("Source"); + sbuf_printf(&sb, "Source"); break; case 1: - printf("Responder"); + sbuf_printf(&sb, "Responder"); break; case 2: - printf("Observer"); + sbuf_printf(&sb, "Observer"); break; default: - printf("???"); + sbuf_printf(&sb, "???"); break; } - printf(" %s ", mca_error_request(mca_error)); + sbuf_printf(&sb, " %s ", mca_error_request(mca_error)); switch ((mca_error & 0x000c) >> 2) { case 0: - printf("Memory"); + sbuf_printf(&sb, "Memory"); break; case 2: - printf("I/O"); + sbuf_printf(&sb, "I/O"); break; case 3: - printf("Other"); + sbuf_printf(&sb, "Other"); break; default: - printf("???"); + sbuf_printf(&sb, "???"); break; } if (mca_error & 0x0100) - printf(" timed out"); + sbuf_printf(&sb, " timed out"); break; } - printf("unknown error %x", mca_error); + sbuf_printf(&sb, "unknown error %x", mca_error); event_type = MCA_T_UNKNOWN; break; } - printf("\n"); + sbuf_printf(&sb, "\n"); if (rec->mr_status & MC_STATUS_ADDRV) { - printf("MCA: Address 0x%llx", (long long)rec->mr_addr); + sbuf_printf(&sb, "MCA: Address 0x%llx", + (long long)rec->mr_addr); if (ser_supported(rec->mr_mcg_cap) && (rec->mr_status & MC_STATUS_MISCV)) { - printf(" (Mode: %s, LSB: %d)", + sbuf_printf(&sb, " (Mode: %s, LSB: %d)", mca_addres_mode(rec->mr_misc), (int)(rec->mr_misc & MC_MISC_RA_LSB)); } - printf("\n"); + sbuf_printf(&sb, "\n"); } if (rec->mr_status & MC_STATUS_MISCV) - printf("MCA: Misc 0x%llx\n", (long long)rec->mr_misc); + sbuf_printf(&sb, "MCA: Misc 0x%llx\n", (long long)rec->mr_misc); + if (event_type < 0 || event_type >= MCA_T_COUNT) { KASSERT(0, ("%s: invalid event type (%d)", __func__, event_type)); event_type = MCA_T_UNKNOWN; } - atomic_add_64(&mca_stats[event_type], 1); + numskipped = 0; + if (!fatal && !uncor) { + /* + * Update statistics and check the rate limit for + * correctable errors. The rate limit is only applied + * after the system records a reasonable number of errors + * of the same type. The goal is to reduce the impact of + * the system seeing and attempting to log a burst of + * similar errors, which (especially when printed to the + * console) can be expensive. + */ + mtx_lock_spin(&mca_lock); + mca_stats[event_type]++; + if (mca_log_interval.tv_sec > 0 && mca_stats[event_type] > 50 && + ratecheck(&mca_last_log_time, &mca_log_interval) == 0) { + mca_log_skipped++; + mtx_unlock_spin(&mca_lock); + goto done; + } + numskipped = mca_log_skipped; + mca_log_skipped = 0; + mtx_unlock_spin(&mca_lock); + } + + error = sbuf_finish(&sb); + if (fatal || !mca_uselog) { + if (numskipped > 0) + printf("MCA: %d events skipped due to rate limit\n", + numskipped); + if (error) + printf("MCA: error logging message (sbuf error %d)\n", + error); + else + sbuf_putbuf(&sb); + } else { + if (numskipped > 0) + log(LOG_ERR, + "MCA: %d events skipped due to rate limit\n", + numskipped); + if (error) + log(LOG_ERR, + "MCA: error logging message (sbuf error %d)\n", + error); + else + log(uncor ? LOG_CRIT : LOG_ERR, "%s", sbuf_data(&sb)); + } + +done: + sbuf_delete(&sb); + if (using_shared_buf) + mtx_unlock_spin(&mca_msg_buf_lock); } static bool @@ -825,7 +928,7 @@ mca_record_entry(enum scan_mode mode, const struct mca_record *record) if (rec == NULL) { mtx_unlock_spin(&mca_lock); printf("MCA: Unable to allocate space for an event.\n"); - mca_log(record); + mca_log(mode, record, false); return; } STAILQ_REMOVE_HEAD(&mca_freelist, link); @@ -982,7 +1085,7 @@ mca_scan(enum scan_mode mode, bool *recoverablep) if (*recoverablep) mca_record_entry(mode, &rec); else - mca_log(&rec); + mca_log(mode, &rec, true); } #ifdef DEV_APIC @@ -1066,7 +1169,7 @@ mca_process_records(enum scan_mode mode) mtx_unlock_spin(&mca_lock); STAILQ_FOREACH(mca, &tmplist, link) - mca_log(&mca->rec); + mca_log(mode, &mca->rec, false); mtx_lock_spin(&mca_lock); while ((mca = STAILQ_FIRST(&tmplist)) != NULL) { @@ -1231,6 +1334,7 @@ mca_setup(uint64_t mcg_cap) mca_banks = mcg_cap & MCG_CAP_COUNT; mtx_init(&mca_lock, "mca", NULL, MTX_SPIN); + mtx_init(&mca_msg_buf_lock, "mca_msg_buf", NULL, MTX_SPIN); STAILQ_INIT(&mca_records); STAILQ_INIT(&mca_pending); mca_tq = taskqueue_create_fast("mca", M_WAITOK, |