There is no reason to do this. Instead just calculate it later.
Reviewed by: jhb
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D36302
tf_trapno is checked on return from interrupt/exception to determine if
special handling is needed for switching address space. This is due to
the possibility of NMI/MCHK/DBG to occur at arbitrary place in kernel,
where both address space and stack used could be transient. Kernel
saves current %cr3 in tf_err for such events, to restore on return.
If user is able to set tf_trapno, it can trigger that special handling,
and since tf_err is also user-controlled by sigreturn(2), the result is
undefined.
PR: 265889
Reported by: lwhsu
Reviewed by: jhb
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D36302
Which means that we must not copy top 8 bytes from the trampoline stack
for the exception frame to the regular thread kstack. As consequence,
this stops corruption of the pcb. The visible effect was often a broken
fork(2) on the CPU where corruption occured.
Account for the detail by substracting 8 from the copy byte count when
moving exception frames from trampoline to the regular stack.
[irettraps handles segmentation/stack/protection faults which could
occur on the doreti path, where we might already switched stack and
address space]
Reported and tested by: pho
Reviewed by: jhb
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D36302
Do not blindly account a page fault occuring on the trampoline area,
as the userspace access fault. Check that it occured exactly in the
instruction that does that.
This avoids unneeded switches of address space on faults not needing the
switch, effectively converting machine resets due to tripple faults,
into regular panics.
Reviewed by: jhb
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D36302
When a freebsd32 caller uses all or most allowed space for control
messages (MCLBYTES == 2K) then the message may no longer fit when
the messages are padded for 64-bit alignment. Historically we've just
shrugged and said there is no ABI guarantee. We ran into this on
CheriBSD where a capsicumized 64-bit nm would fail when called with more
than 64 files.
Fix this by not gratutiously capping size of mbuf data we'll allocate
to MCLBYTES and let m_get2 allocate up to MJUMPAGESIZE (4K or larger).
Instead of hard-coding a length check, let m_get2 do it and check for a
NULL return.
Reviewed by: markj, jhb, emaste
Sponsored by: DARPA, AFRL
Differential Revision: https://reviews.freebsd.org/D36322
This could only happen on systems with PAGE_SIZE < 4K and FreeBSD
doesn't support such systems.
Reviewed by: np, imp, jhb
Sponsored by: DARPA, AFRL
Differential Revision: https://reviews.freebsd.org/D36321
The Vax supported such things, but FreeBSD does not. This further
implies that MJUMPAGESIZE > MCLBYTES so assert this and remove code
handling them being equal.
Reviewed by: kp, imp, jhb
Sponsored by: DARPA, AFRL
Differential Revision: https://reviews.freebsd.org/D36320
Add an internal debug level global:
- Level 1 (-d) currently does nothing.
- Level 2 (-d -d) enables libfetch debugging (quite verbose) so it's
possible to see what pkg is attempting to download without having
to sniff traffic.
Reviewed by: debdrup, bapt
Differential Revision: https://reviews.freebsd.org/D35756
Summary:
Currently netinet6/ code allocates IPv6 prefixes (nd_prefix) for
both manually-assigned addresses and advertised prefixes. As a result,
prefixes from manually-assigned prefixes can be seen in `ndp -p` list
and be cleared via `ndp -P`. The latter relies on the SIOCSPFXFLUSH_IN6
ioctl to clear to prefix list.
The original intent of the SIOCSPFXFLUSH_IN6 was to clear prefixes
originated from the advertising routers:
```
1998-09-02 JINMEI, Tatuya <jinmei@isl.rdc.toshiba.co.jp>
* nd6.c (nd6_ioctl): added 2 new ioctls; SIOCSRTRFLUSH_IN6 and
SIOCSPFXFLUSH_IN6. The former is to flush all default routers
in the default router list, and the latter is to flush all the
prefixes and the addresses derived from them in the prefix list.
```
Restore the intent by marking prefixes derived from the RA messages
with newly-added ndpr_flags.ra_derived flag and skip prefixes not marked
with such flag during deletion and listing.
Differential Revision: https://reviews.freebsd.org/D36312
MFC after: 2 weeks
This is a joint work with manu.
- fixed conditions in do_i2c_transfer and i2c_transfer as linux_i2cbb
does not set adapter->algo->master_xfer but does set
adapter->algo_data;
- fixed parent bus specification for linux_i2cbb driver module;
- actually implemented iicbb_transfer method;
- added iicbb_pre_xfer and iicbb_post_xfer methods;
- removed unnecessary and harmful delays (and other extra logic) from
iicbb methods as iicbb driver already has them;
- added setting of iicbb speed based on algo_data->udelay, so that iicbb
uses correct delays;
PR: 265920
Fixes: 1961a14a47 linuxkpi: Add i2c support
MFC after: 2 weeks
Sponsored by: Beckhoff Automation GmbH & Co. KG (manu's work)
We don't use EFI_MEMORY_DESCRIPTOR that's typedef'd here. We use the one
from sys/efi.h instead. Remove the clutter here as these two are subtly
different (though wind up with the same layout due to alignment rules).
Sponsored by: Netflix
The fsck_ffs(8) utility has two subsystems for reading and writing
inodes. The getnextinode() interface is used in Pass 1 (and Pass
1b if needed) to sequentially walk through all the inodes in the
filesystem. The ginode() interface is used to read and write
individual inodes. Pass 1 uses a mix of both interfaces. This
change ensures that ginode() returns a pointer to the inode in the
cache maintained by getnextinode() when that interface holds the
requested inode so that all modifications to the inode are made in
a single place and are all written to the disk together.
Reported by: Peter Holm
Tested by: Peter Holm
Sponsored by: The FreeBSD Foundation
Historically fsck_ffs(8) would only use alternate superblocks when
running in manual mode. When the standard superblock fails, it now
tries to find and use a backup superblocks even when running in `preen'
mode. If an alternate superblock is found and the filesystem is
successfully cleaned up using it, write the alternate superblock
back to the standard superblock so that the filesystem can be
subsequently mounted and used.
Reported by: Peter Holm
Tested by: Peter Holm
Sponsored by: The FreeBSD Foundation
This applies one of the changes from
5567d6b441 to other architectures
besides arm64.
Reviewed by: kib
Sponsored by: DARPA
Differential Revision: https://reviews.freebsd.org/D36263
Current `.arcconfig` specifies explicit mapping to the FreeBSD/SVN
project in Phabricator, which is inactive. This mapping only gets
updated to the current "production" (FreeBSD/git) project when the
underlying change is committed.
Update arcanist configuration to create all new diffs in the
FreeBSD/git project.
Reviewed By: dim, emaste
Differential Revision: https://reviews.freebsd.org/D36204
In that case, there is only one page size.
Reviewed by: kib
MFC after: 1 week
Sponsored by: Dell EMC Isilon
Differential Revision: https://reviews.freebsd.org/D36265
Currently when the peer collapses its rwnd, we mark packets to be retransmitted
and use the must_retran flags like we do when a PMTU collapses to retransmit the
collapsed packets. However this causes a problem with some middle boxes that
play with the rwnd to control flow. As soon as the rwnd increases we start resending
which may be not even a rtt.. and in fact the peer may have gotten the packets. Which
means we gratuitously retransmit packets we should not.
The fix here is to make sure that a rack time has passed before retransmitting the packets.
This makes sure that the rwnd collapse was real and the packets do need retransmission.
Reviewed by: tuexen
Sponsored by: Netflix Inc
Differential Revision: https://reviews.freebsd.org/D35166
A while back Hans optimized the LRO code. This is great but one
optimization he did degrades the timestamp precision so that
all flushed LRO entries end up with the same LRO timestamp
if there is not a hardware timestamp. The intent of the LRO timestamp
is to get as close to the time that the packet arrived as possible. Without
the LRO queuing this works out fine since a binuptime is taken and then
the rx_common code is called. But when you go through the queue path
you end up *not* updating the M_LRO_TSTMP fields.
Another issue in the LRO code is several places that cause packet reordering. In
general TCP can handle reordering but it can cause extra un-needed retransmission
as well as other oddities. We will fix all of the reordering problems.
Lets fix this so that we restore the precision to the timestamp.
Reviewed by: tuexen, gallatin
Sponsored by: Netflix Inc
Differential Revision: https://reviews.freebsd.org/D36043
When scrubbing an encrypted filesystem with unloaded key still report an
error in zpool status.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Alek Pinchuk <apinchuk@axcient.com>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Closes#13675Closes#13717
Get rid of struct pfsync_pkt. It was used to store data on the stack to
pass to all the submessage handlers, but only the flags part of it was
ever used. Just pass the flags directly instead.
Reviewed by: kp
Obtained from: OpenBSD
Sponsored by: InnoGames GmbH
Differential Revision: https://reviews.freebsd.org/D36294
Currently make.py has a hack to add the cross-build headers to the
include search path when bootstrapping bmake on Linux (but not macOS).
This is a bit of an abuse of these headers, and e9ba1fd5ed was not
prepared for this, since sys/bitcount.h won't exist in that instance (it
gets copied into WORLDTMP during the legacy build). Work around this
until we can wean the bmake bootstrap off using these headers by not
including sys/bitcount.h when it doesn't exist.
Fixes: e9ba1fd5ed ("tools/build: Provide FreeBSD's bitstring API when cross-building")
During a discussion with someone working on NFS-over-TLS
for a non-FreeBSD platform, we agreed that a single server
daemon for TLS handshakes could become a bottleneck when
an NFS server first boots, if many concurrent NFS-over-TLS
connections are attempted.
This patch modifies the kernel RPC code so that it can
handle multiple rpc.tlsservd daemons. A separate commit
currently under review as D35886 for the rpc.tlsservd
daemon.
There are a couple changes included here. The first is to introduce
a cap on the size the ZED will grow the zevent list to. One million
entries is more than enough for most use cases, and if you are
overflowing that value, the problem needs to be addressed another
way. The value is also tunable, for those who want the limit to be
higher or lower.
The other change is to add a kernel module parameter that allows
snapshot creation/deletion to be exempted from the history logging;
for most workloads, having these things logged is valuable, but for
some workloads it produces large quantities of log spam and isn't
especially helpful.
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Paul Dagnelie <pcd@delphix.com>
Issue #13374Closes#13753
Fix the off-by-one in the KASSERT in arm_gic_reserve_msi_range()
allowing the MSIs being allocated at the end of the IRQ range.
Reported by: bz
Submitted by: andrew
Reviewed by: andrew
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D36291
Classify VLAN devices as ethernet in rdma_copy_addr().
This fixes the following error message:
rdma_bind_addr: No such file or directory
Submitted by: bartosz.sobczak@intel.com (Bartosz Sobczak)
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D36120
Sponsored by: NVIDIA Networking
When specifying a custom band-pass filter for the sound(4) equalizer,
use the same format as used by the resulting C-code define to avoid
confusion.
PR: 195760
Sponsored by: NVIDIA Networking
NXP defines unique name for LS1046A UART - "NXP0018".
It is ns8250 compatible, adding a new uart compat data entry is enough
to make it work.
Tested on LS1046ARDB.
Approved by: mw(mentor)
Reviewed by: mw, kd
Obtained from: Semihalf
Differential Revision: https://reviews.freebsd.org/D36260
For now it is read-only, but eventually the cycle that goes over
all fragments should be refactored and this MIB should also become
read/write.
This MIB will allow SNMP daemons to implement MIB-II ipReasmTimeout MIB
straightfoward. Right now net-snmp compilation is broken by 1922eb3e9c.
The base system bsnmpd is not broken just because it ignored PR_SLOWTIMO,
and thus always returned incorrectly doubled value for ipReasmTimeout.
We cannot see a thread with the flag set in unsuspend, after we stopped
doing SINGLE_ALLPROC from user processes.
Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D36207
It does not serve any purpose after we stopped doing
thread_single(SINGLE_ALLPROC) from stoppable user processes.
Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D36207
There is a problem still left after the fixes to REAP_KILL_PROC. The
handling of the stopping signals by sig_suspend_threads() can occur
outside the stopping process context by tdsendsignal(), and it uses
mostly the same mechanism of aborting sleeps as suspension. In other
words, it badly interacts with thread_single(SINGLE_ALLPROC).
But unlike single threading from the process context, we cannot wait by
sleep for other single threading requests to pass, because we own
spinlock(s).
Fix this by moving both the thread_single(p2, SINGLE_ALLPROC), and the
signalling, to the threaded taskqueue which cannot be single-threaded
itself.
Reported and tested by: pho
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D36207
