cxgbe(4): Do not panic when tx is called with invalid checksum requests.
There is no need to panic in if_transmit if the checksums requested are
inconsistent with the frame being transmitted. This typically indicates
that the kernel and driver were built with different INET/INET6 options,
or there is some other kernel bug. The driver should just throw away
the requests that it doesn't understand and move on.
Kevin Bowling [Wed, 21 Apr 2021 02:35:14 +0000 (19:35 -0700)]
ixgbe: Improve device name strings
This is just clerical work to ease bug triage and may be used to set
expectations around the ability for anyone in the community to perform
testing and development on older parts.
Mark Johnston [Wed, 28 Apr 2021 14:42:59 +0000 (10:42 -0400)]
pipe: Avoid calling selrecord() on a closing pipe
pipe_poll() may add the calling thread to the selinfo lists of both ends
of a pipe. It is ok to do this for the local end, since we know we hold
a reference on the file and so the local end is not closed. It is not
ok to do this for the remote end, which may already be closed and have
called seldrain(). In this scenario, when the polling thread wakes up,
it may end up referencing a freed selinfo.
Add rib_walk_from() wrapper for selective rib tree traversal.
Provide wrapper for the rnh_walktree_from() rib callback.
As currently `struct rib_head` is considered internal to the
routing subsystem, this wrapper is necessary to maintain isolation
from the external code.
cxgbe(4): Add support for NIC suspend/resume and live reset.
Add suspend/resume callbacks to the driver and a live reset built around
them. This commit covers the basic NIC and future commits will expand
this functionality to other stateful parts of the chip. Suspend and
resume operate on the chip (the t?nex nexus device) and affect all its
ports. It is not possible to suspend/resume or reset individual ports.
All these operations can be performed on a running NIC. A reset will
look like a link bounce to the networking stack.
Here are some ways to exercise this functionality:
/* Manual reset with driver sysctl. */
# sysctl dev.t6nex.0.reset=1
/* Automatic adapter reset on any fatal error. */
# hw.cxgbe.reset_on_fatal_err=1
Suspend disables the adapter (DMA, interrupts, and the port PHYs) and
marks the hardware as unavailable to the driver. All ifnets associated
with the adapter are still visible to the kernel but operations that
require hardware interaction will fail with ENXIO. All ifnets report
link-down while the adapter is suspended.
Resume will reattach to the card, reconfigure it as before, and recreate
the queues servicing the existing ifnets. The ifnets are able to send
and receive traffic as soon as the link comes back up.
Reset is roughly the same as a suspend and a resume with at least one of
these events in between: D0->D3Hot->D0, FLR, PCIe link retrain.
Rick Macklem [Wed, 28 Apr 2021 00:30:16 +0000 (17:30 -0700)]
nfscl: add check for NULL clp and forced dismounts to nfscl_delegreturnvp()
Commit aad780464fad added a function called nfscl_delegreturnvp()
to return delegations during the NFS VOP_RECLAIM().
The function erroneously assumed that nm_clp would
be non-NULL. It will be NULL for NFSV4.0 mounts until
a regular file is opened. It will also be NULL during
vflush() in nfs_unmount() for a forced dismount.
This patch adds a check for clp == NULL to fix this.
Also, since it makes no sense to call nfscl_delegreturnvp()
during a forced dismount, the patch adds a check for that
case and does not do the call during forced dismounts.
Rick Macklem [Tue, 27 Apr 2021 22:32:35 +0000 (15:32 -0700)]
nfsd: fix a NFSv4.1 Linux client mount stuck in CLOSE_WAIT
It was reported that a NFSv4.1 Linux client mount against
a FreeBSD12 server was hung, with the TCP connection in
CLOSE_WAIT state on the server.
When a NFSv4.1/4.2 mount is done and the back channel is
bound to the TCP connection, the soclose() is delayed until
a new TCP connection is bound to the back channel, due to
a reference count being held on the SVCXPRT structure in
the krpc for the socket. Without the soclose() call, the socket
will remain in CLOSE_WAIT and this somehow caused the Linux
client to hang.
This patch adds calls to soshutdown(.., SHUT_WR) that
are performed when the server side krpc sees that the
socket is no longer usable. Since this can be done
before the back channel is bound to a new TCP connection,
it allows the TCP connection to proceed to CLOSED state.
Kevin Bowling [Sun, 25 Apr 2021 08:22:23 +0000 (01:22 -0700)]
e1000: Rework em_msi_link interrupt filter
* Fix 82574 Link Status Changes, carrying the OTHER mask bit around as
needed.
* Move igb-class LSC re-arming out of FAST back into the handler.
* Clarify spurious/other interrupt re-arms in FAST.
In MSI-X mode, 82574 and igb-class devices use an interrupt filter to
handle Link Status Changes. We want to do LSC re-arms in the handler
to take advantage of autoclear (EIAC) single shot behavior.
82574 uses 'Other' in ICR and IMS for LSC interrupt types when in MSI-X
mode, so we need to set and re-arm the 'Other' bit during attach and
after ICR reads in the FAST handler if not an LSC or after handling on
LSC due to autoclearing.
This work was primarily done to address the referenced PR, but inspired
some clarification and improvement for igb-class devices once the
intentions of previous bug fix attempts became clearer.
[fib algo] Delay algo init at fib growth to to allow to reliably use rib KPI.
Currently, most of the rib(9) KPI does not use rnh pointers, using
fibnum and family parameters to determine the rib pointer instead.
This works well except for the case when we initialize new rib pointers
during fib growth.
In that case, there is no mapping between fib/family and the new rib,
as an entirely new rib pointer array is populated.
Address this by delaying fib algo initialization till after switching
to the new pointer array and updating the number of fibs.
Set datapath pointer to the dummy function, so the potential callers
won't crash the kernel in the brief moment when the rib exists, but
no fib algo is attached.
This change allows to avoid creating duplicates of existing rib functions,
with altered signature.
Merge commit f511dc75e4c1 from llvm git (by Mark Johnston):
[asan] Add an offset for the kernel address sanitizer on FreeBSD
This is based on a port of the sanitizer runtime to the FreeBSD kernel
that has been commited as https://cgit.freebsd.org/src/commit/?id=38da497a4dfcf1979c8c2b0e9f3fa0564035c147
and the following commits.
Reviewed By: emaste, dim
Differential Revision: https://reviews.llvm.org/D98285
mmccam: Add mmc_sim, a generic sim for mmc driver to use
This adds a generic sim that abstract a lot of what needs to be implemented
in a driver for mmccam support.
A new interface with three methods is added :
- mmc_sim_get_tran_settings: Use to get what the controller supports in term
of capabilities, freq etc ...
- mmc_sim_set_tran_settings: Use to change the speed/freq/etc of the
sdcard host controller
- mmc_sim_cam_request: Used for MMCIO requests
Modular fib lookup framework features logic that allows
route update batching for the algorithms that cannot easily
apply the routing change without rebuilding. As a result,
dataplane lookups may return old data until the the sync
takes place. With the default sync timeout of 50ms, it is
possible that new binary like ping(8) executed exactly after
route(8) will still use the old fib data.
To address some aspects of the problem, framework executes
all rtable changes without RTF_GATEWAY synchronously.
To fix the aforementioned problem, this diff extends sync
execution for all RTF_STATIC routes (e.g. ones maintained by
route(8).
This fixes a bunch of tests in the networking space.
33cb3cb2e321 introduced an `rib_head` structure field under the
FIB_ALGO define. This may be problematic for the CTF, as some
of the files including `route_var.h` do not have `fib_algo`
defined.
Make dtrace happy by making the field unconditional.
Moritz Schmitt [Tue, 27 Apr 2021 01:59:12 +0000 (03:59 +0200)]
Make pkg(7) use environment variables specified in pkg.conf
Modify /usr/sbin/pkg to use environment variables specified in pkg.conf.
This allows control over underlying libraries like fetch(3), which can
be configured by setting HTTP_PROXY.
When fsck_ffs is creating a lost+found directory it must allocate
an inode and a filesystem block. If it encounters a cylinder group
with a bad check hash, it complains twice: once for the inode and
again for the filesystem block.
Rick Macklem [Tue, 27 Apr 2021 00:48:21 +0000 (17:48 -0700)]
nfscl: fix the handling of NFSERR_DELAY for Open/LayoutGet RPCs
For a pNFS mount, the NFSv4.1/4.2 client uses compound RPCs that
have both Open and LayoutGet operations in them.
If the pNFS server were tp reply NFSERR_DELAY for one of these
compounds, the retry after a delay cannot be handled by
newnfs_request(), since there is a reference held on the open
state for the Open operation in them.
Fix this by adding these RPCs to the "don't do delay here"
list in newnfs_request().
This patch is only needed if the mount is using pNFS (the "pnfs"
mount option) and probably only matters if the MDS server
is issuing delegations as well as pNFS layouts.
Make fsck_ffs more persistent in creating a lost+found directory.
When fsck_ffs is running in interactive mode and finds unlinked files,
it offers to either unlink them or place them in a lost+found directory.
If the lost+found directory option is requested and no lost+found
directory exists, fsck_ffs offers to create one. When creating one,
it must allocate an inode and a filesystem block. It attempts to
allocate them from the first cylinder group. If the first cylinder
group has a bad check hash, it gives up.
This change expands the search into later cylinder groups when the
first one fails with a bad check hash.
Rick Macklem [Mon, 26 Apr 2021 23:24:10 +0000 (16:24 -0700)]
nfsd: fix the slot sequence# when a callback fails
Commit 4281bfec3628 patched the server so that the
callback session slot would be free'd for reuse when
a callback attempt fails.
However, this can often result in the sequence# for
the session slot to be advanced such that the client
end will reply NFSERR_SEQMISORDERED.
To avoid the NFSERR_SEQMISORDERED client reply,
this patch negates the sequence# advance for the
case where the callback has failed.
The common case is a failed back channel, where
the callback cannot be sent to the client, and
not advancing the sequence# is correct for this
case. For the uncommon case where the client's
reply to the callback is lost, not advancing the
sequence# will indicate to the client that the
next callback is a retry and not a new callback.
But, since the FreeBSD server always sets "csa_cachethis"
false in the callback sequence operation, a retry
and a new callback should be handled the same way
by the client, so this should not matter.
Until you have this patch in your NFSv4.1/4.2 server,
you should consider avoiding the use of delegations.
Even with this patch, interoperation with the
Linux NFSv4.1/4.2 client in kernel versions prior
to 5.3 can result in frequent 15second delays if
delegations are enabled. This occurs because, for
kernels prior to 5.3, the Linux client does a TCP
reconnect every time it sees multiple concurrent
callbacks and then it takes 15seconds to recover
the back channel after doing so.
cxgbe(4): Separate the sw- and hw-specific parts of resource allocations
The driver uses both software resources (locks, callouts, memory for
descriptors and for bookkeeping, sysctls, etc.) and hardware resources
(VIs, DMA queues, TCAM entries, etc.) to operate the NIC. This commit
splits the single *_ALLOCATED flag used to track all these resources
into separate *_SW_ALLOCATED and *_HW_ALLOCATED flags.
This is the simplified pseudocode that now applies to most queues (foo
can be ctrlq/txq/rxq/ofld_txq/ofld_rxq):
/* Idempotent */
alloc_foo
{
if (!SW_ALLOCATED)
init_iq/init_eq/init_fl no-fail sw init
alloc_iq_fl/alloc_eq/alloc_wrq may-fail sw alloc
add_foo_sysctls, etc. no-fail post-alloc items
if (!HW_ALLOCATED)
alloc_iq_fl_hwq/alloc_eq_hwq hw resource allocation
}
/* Idempotent */
free_foo
{
if (!HW_ALLOCATED)
free_iq_fl_hwq/free_eq_hwq release hw resources
if (!SW_ALLOCATED)
free_iq_fl/free_eq/free_wrq release sw resources
}
The routines that take the driver to FULL_INIT_DONE and VI_INIT_DONE and
back are now all idempotent. The quiesce routines pay attention to the
HW_ALLOCATED flag and will not wait on the hardware for pidx/cidx
updates and other completions if this flag is not set.
Michael Tuexen [Mon, 26 Apr 2021 08:38:05 +0000 (10:38 +0200)]
sctp: improve handling of illegal packets containing INIT chunks
Stop further processing of a packet when detecting that it
contains an INIT chunk, which is too small or is not the only
chunk in the packet. Still allow to finish the processing
of chunks before the INIT chunk.
Thanks to Antoly Korniltsev and Taylor Brandstetter for reporting
an issue with the userland stack, which made me aware of this
issue.
Mark Johnston [Mon, 26 Apr 2021 19:15:31 +0000 (15:15 -0400)]
pkgbase: Remove package name mangling in generate-ucl.sh
The mangling was present in the initial revision of the script, but its
purpose is not clear. It may have been to avoid defining make(1)
variables with a dash in the name, but this is permitted. Furthermore,
it results in invalid dependency information if a dependency's name
contains an underscore, causing e.g., libcompiler_rt-dev to depend on
libcompiler-rt, and resulting in warnings when installing base system
packages. Remove the mangling.
Reviewed by: manu
MFC after: 2 months
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D29883
Mark Johnston [Mon, 26 Apr 2021 18:53:16 +0000 (14:53 -0400)]
imgact_elf: Ensure that the return value in parse_notes is initialized
parse_notes relies on the caller-supplied callback to initialize "res".
Two callbacks are used in practice, brandnote_cb and note_fctl_cb, and
the latter fails to initialize res. Fix it.
In the worst case, the bug would cause the inner loop of check_note to
examine more program headers than necessary, and the note header usually
comes last anyway.
Reviewed by: kib
Reported by: KMSAN
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D29986
Warner Losh [Mon, 26 Apr 2021 17:04:15 +0000 (11:04 -0600)]
newbus: remove support for SINGLETON
Revert rest of de8dd262c43b since it's now unused.
jhibbits@ introduced this to give powerpc MMU functions IFUNC like
performance while retaining the kobj interface, speeding up operations
10-20%. Since there was only ever one instance of the mmu interface
active at any given time, we could cache the looked up results more
agressively.
powerpc migrated to using IFUNCs to get an even larger performance boost
in 45b69dd63e, deleting the two files it was added to in de8dd262c43b.
However, there's few, if any, other potential applications of this to
the tree today. It's now unused and undocumented. Retire it to eliminate
this wart and to preclude the need to document it. Should a simmilar
case arise in the future, the code is in git...
Add a test case where the pfctl optimizer will generate a table
automatically. These tables have long names, which we accidentally broke
in the nvlist ADDRULE ioctl.
When parsing the nvlist for a struct pf_addr_wrap we unconditionally
tried to parse "ifname". This broke for PF_ADDR_TABLE when the table
name was longer than IFNAMSIZ. PF_TABLE_NAME_SIZE is longer than
IFNAMSIZ, so this is a valid configuration.
Only parse (or return) ifname or tblname for the corresponding
pf_addr_wrap type.
This manifested as a failure to set rules such as these, where the pfctl
optimiser generated an automatic table:
pass in proto tcp to 192.168.0.1 port ssh
pass in proto tcp to 192.168.0.2 port ssh
pass in proto tcp to 192.168.0.3 port ssh
pass in proto tcp to 192.168.0.4 port ssh
pass in proto tcp to 192.168.0.5 port ssh
pass in proto tcp to 192.168.0.6 port ssh
pass in proto tcp to 192.168.0.7 port ssh
Allow up to 5 labels to be set on each rule.
This offers more flexibility in using labels. For example, it replaces
the customer 'schedule' keyword used by pfSense to terminate states
according to a schedule.
Kevin Bowling [Wed, 21 Apr 2021 05:27:48 +0000 (22:27 -0700)]
e1000: Improve device name strings
This is just clerical work to ease bug triage and may be used to set
expectations around the ability for anyone in the community to perform
testing and development on older parts (this driver covers over 20 years
of silicon)
Reviewed by: erj
Approved by: markj
Sponsored by: Pink Floyd - Any Colour You Like (in kind)
Differential Revision: https://reviews.freebsd.org/D29872
Patrick Kelsey [Mon, 26 Apr 2021 04:25:59 +0000 (00:25 -0400)]
iflib: Improve mapping of TX/RX queues to CPUs
iflib now supports mapping each (TX,RX) queue pair to the same CPU
(default), to separate CPUs, or to a pair of physical and logical CPUs
that share the same L2 cache. The mapping mechanism supports unequal
numbers of TX and RX queues, with the excess queues always being
mapped to consecutive physical CPUs. When the platform cannot
distinguish between physical and logical CPUs, all are treated as
physical CPUs. See the comment on get_cpuid_for_queue() for the
entire matrix.
The following device-specific tunables influence the mapping process:
dev.<device>.<unit>.iflib.core_offset (existing)
dev.<device>.<unit>.iflib.separate_txrx (existing)
dev.<device>.<unit>.iflib.use_logical_cores (new)
The following new, read-only sysctls provide visibility of the mapping
results:
dev.<device>.<unit>.iflib.{t,r}xq<n>.cpu
When an iflib driver allocates TX softirqs without providing reference
RX IRQs, iflib now binds those TX softirqs to CPUs using the above
mapping mechanism (that is, treats them as if they were TX IRQs).
Previously, such bindings were left up to the grouptaskqueue code and
thus fell outside of the iflib CPU mapping strategy.
Rick Macklem [Sun, 25 Apr 2021 19:52:48 +0000 (12:52 -0700)]
nfscl: fix delegation recall when the file is not open
Without this patch, if a NFSv4 server recalled a
delegation when the file is not open, the renew
thread would block in the NFS VOP_INACTIVE()
trying to acquire the client state lock that it
already holds.
This patch fixes the problem by delaying the
vrele() call until after the client state
lock is released.
This bug has been in the NFSv4 client for
a long time, but since it only affects
delegation when recalled due to another
client opening the file, it got missed
during previous testing.
Until you have this patch in your client,
you should avoid the use of delegations.
Previously it depended on sysctl, which itself has no dependencies,
so rcorder(8) had a bit too much flexibility when choosing when to run
it. Make sure it runs just between 'fsck' and 'root'.
Option `FIB_ALGO` gates new modular fib lookup functionality,
enabling more performant routing table lookups and improving
control plane convergence under the load.
Detailed feature description is available in D27401.
Traditionally we had 2 sources of information whether the
added/delete route request targets network or a host route:
netmask (RTA_NETMASK) and RTF_HOST flag.
The former one is tricky: netmask can be empty or can explicitly
specify the host netmask. Parsing netmask sockaddr requires per-family
parsing and that's what rtsock code traditionally avoided. As a result,
consistency was not enforced and it was possible to specify network with
the RTF_HOST flag and vice versa.
Continue normalization efforts from D29826 and D29826 and ensure that
RTF_HOST flag always reflects host/network data from netmask field.
Differential Revision: https://reviews.freebsd.org/D29958
MFC after: 2 days
Improve debugging output on routing tests failure.
Most of the routing tests create per-test VNET, making
it harder to repeat the failure with CLI tools.
Provide an additional route/nexthop data on failure.
Robert Watson [Sat, 2 Jan 2021 16:42:28 +0000 (16:42 +0000)]
Support run-time configuration of the PIPE_MINDIRECT threshold.
PIPE_MINDIRECT determines at what (blocking) write size one-copy
optimizations are applied in pipe(2) I/O. That threshold hasn't
been tuned since the 1990s when this code was originally
committed, and allowing run-time reconfiguration will make it
easier to assess whether contemporary microarchitectures would
prefer a different threshold.
(On our local RPi4 baords, the 8k default would ideally be at least
32k, but it's not clear how generalizable that observation is.)
Allow the tcp_lro_flush_all() function to be called when the control
structure is zeroed, by setting the VNET after checking the mbuf count
for zero. It appears there are some cases with early interrupts on some
network devices which still trigger page-faults on accessing a NULL "ifp"
pointer before the TCP LRO control structure has been initialized.
This basically preserves the old behaviour, prior to 9ca874cf740ee68c5742df8b5f9e20910085c011 .
Alexander Motin [Sat, 24 Apr 2021 03:18:01 +0000 (23:18 -0400)]
mpr/mps(4): Make device mapping some more robust.
Allow new enclosure to replace previously existing one if there is
no completely unused table entry, same as it is done for devices.
If we can not process DPM due to corruption -- wipe it and restart
from scratch. Otherwise I don't see a way to recover persistence if
something go wrong and there is no BIOS to recover it for us.
Together this solves a problem that appeared when 9300-8i firmware
update to 16.00.10.00 somehow switched its mapping mode from Device
Persistence to Enclosure/Slot without wiping the DPM table. It made
HBA completely unusable, since overflowed and conflicting mapping
table was unable to map any of enclosures and so devices.
Also while there make some enclosure mapping errors more informative.
3e7bae08210e0 turns the BUS_READ_IVAR() failure from a warning into a
KASSERT. For certain PCI audio devices such like snd_csa(4) and
snd_emu10kx(4), the ac97_create() keeps the device handler generated
by device_add_child(pci_dev, "pcm"), which is not really a PCI device
handler. This in turn causes the subsequent pci_get_subdevice()
inside ac97_initmixer() triggering a panic.
This patch tries to put a bandaid for the aforementioned pcm device
children such that they can use the correct PCI handler(from parent)
to avoid a KASSERT panic in the INVARIANTS kernel.
Rick Macklem [Fri, 23 Apr 2021 22:24:47 +0000 (15:24 -0700)]
nfsd: fix session slot handling for failed callbacks
When the NFSv4.1/4.2 server does a callback to a client
on the back channel, it will use a session slot in the
back channel session. If the back channel has failed,
the callback will fail and, without this patch, the
session slot will not be released.
As more callbacks are attempted, all session slots
can become busy and then the nfsd thread gets stuck
waiting for a back channel session slot.
This patch frees the session slot upon callback
failure to avoid this problem.
Without this patch, the problem can be avoided by leaving
delegations disabled in the NFS server.
John Baldwin [Fri, 23 Apr 2021 16:40:04 +0000 (09:40 -0700)]
etcupdate: Remove the old pre-world tree when rotating for -p.
This fixes a bug in an earlier change to move tree rotation to
the end of the update where the step to make room for the new
preworld tree was deleting the old "current" tree instead of
the old "preworld" tree.
Mark Johnston [Fri, 23 Apr 2021 13:14:42 +0000 (09:14 -0400)]
Re-enable network ioctls in capability mode
This reverts a portion of 274579831b61 ("capsicum: Limit socket
operations in capability mode") as at least rtsol and dhcpcd rely on
being able to configure network interfaces while in capability mode.
Reported by: bapt, Greg V
Sponsored by: The FreeBSD Foundation
projects / FreeBSD / FreeBSD.git / log