Division by zero triggers an arithmetic exception and should not be very
common. Predict this.
No functional change intended.
MFC after: 1 week
Sponsored by: NVIDIA Networking
TCP has an idle-reduce feature that allows a connection to reduce its
cwnd after it has been idle more than an RTT. This feature only works
for a sending side connection. It does this by at output checking the
idle time (t_rcvtime vs ticks) to see if its more than the RTO timeout.
The problem comes if you are a web server. You get a request and
then send out all the data.. then go idle. The next time you would
send is in response to a request from the peer asking for more data.
But the thing is you updated t_rcvtime when the request came in so
you never reduce.
The fix is to do the idle reduce check also on inbound.
Reviewed by: tuexen, rscheff
Sponsored by: Netflix Inc
Differential Revision: https://reviews.freebsd.org/D36721
Two functions to call tcp_drop() and tcp_close() from a callout context.
Garbage collect tcp_inpinfo_lock_del(), it has a single use now.
Differential revision: https://reviews.freebsd.org/D36397
In the original design of the network stack from the protocol control
input method pr_ctlinput was used notify the protocols about two very
different kinds of events: internal system events and receival of an
ICMP messages from outside. These events were coded with PRC_ codes.
Today these methods are removed from the protosw(9) and are isolated
to IPv4 and IPv6 stacks and are called only from icmp*_input(). The
PRC_ codes now just create a shim layer between ICMP codes and errors
or actions taken by protocols.
- Change ipproto_ctlinput_t to pass just pointer to ICMP header. This
allows protocols to not deduct it from the internal IP header.
- Change ip6proto_ctlinput_t to pass just struct ip6ctlparam pointer.
It has all the information needed to the protocols. In the structure,
change ip6c_finaldst fields to sockaddr_in6. The reason is that
icmp6_input() already has this address wrapped in sockaddr, and the
protocols want this address as sockaddr.
- For UDP tunneling control input, as well as for IPSEC control input,
change the prototypes to accept a transparent union of either ICMP
header pointer or struct ip6ctlparam pointer.
- In icmp_input() and icmp6_input() do only validation of ICMP header and
count bad packets. The translation of ICMP codes to errors/actions is
done by protocols.
- Provide icmp_errmap() and icmp6_errmap() as substitute to inetctlerrmap,
inet6ctlerrmap arrays.
- In protocol ctlinput methods either trust what icmp_errmap() recommend,
or do our own logic based on the ICMP header.
Differential revision: https://reviews.freebsd.org/D36731
where struct ipsec_methods is defined. Not a functional change.
Allows further modification of method prototypes without breaking
compilation of other ipsec compilation units.
Differential revision: https://reviews.freebsd.org/D36730
Now these functions are called only from icmp*_input(). The pointer
to the ICMP data is never NULL and cmd has a limited set of values.
In the past the functions were demultiplexing control messages from
ICMP layer, as well as internally generated events. In the latter
case the the pointer to IP would be NULL.
Differential revision: https://reviews.freebsd.org/D36729
and mark those PRC_* codes, that are used. The rest are dead code.
This is not a functional change, but illustrative to make easier
review of following changes.
After decoupling of protosw(9) and IP wire protocols in 78b1fc05b2 for
IPv4 we got vector ip_ctlprotox[] that is executed only and only from
icmp_input() and respectively for IPv6 we got ip6_ctlprotox[] executed
only and only from icmp6_input(). This allows to use protocol specific
argument types in these methods instead of struct sockaddr and void.
Reviewed by: melifaro
Differential revision: https://reviews.freebsd.org/D36727
The netinet/ipprotosw.h and netinet6/ip6protosw.h were KAME relics, with
the former removed in f0ffb944d2 in 2001 and the latter survived until
today. It has been reduced down to only one useful declaration that
moves to ip6_var.h
Reviewed by: melifaro
Differential revision: https://reviews.freebsd.org/D36726
With this change one can make a forward declaration of a function
that is of UDP tunneling type.
Reviewed by: melifaro
Differential revision: https://reviews.freebsd.org/D36724
The netmap-specific data stored at be->opaque is freed by the caller
on error as part of freeing be.
Reviewed by: markj
Reported by: GCC -Wfree-nonheap-object
Differential Revision: https://reviews.freebsd.org/D36828
wdc_get_dui_log_size allocates a buffer and then advances the
returned pointer. Passing this advanced pointer to free() is UB,
so save the original pointer to pass to free() instead.
Reviewed by: imp
Reported by: GCC 12 -Wfree-nonheap-object
Differential Revision: https://reviews.freebsd.org/D36827
DevPathToTextUsbWWID allocates a separate copy of the SerialNumber
string to append a null terminator if the original string is not
null terminated. However, by using AllocateCopyPool, it tries to
copy 'Length + 1' words from the existing string containing 'Length'
characters into the target string. Split the copy out to only
copy 'Length' characters instead.
Reviewed by: imp, emaste
Reported by: GCC 12 -Wstringop-overread
Differential Revision: https://reviews.freebsd.org/D36826
The length passed to strncpy is the length of the source string, not
the destination buffer. This triggers a non-fatal warning in GCC 12.
Hoewver, the code is also odd. It is really just a memcpy of the
string without its nul terminator. For that use case, memcpy is
clearer.
Reviewed by: imp, emaste
Differential Revision: https://reviews.freebsd.org/D36824
When -B or -I are used, change() evaluates the lines in a hunk to
determine if it is a hunk that should be ignored. It does this by
reading each candidate line into a mallocated buffer via preadline()
and then calling ignoreline(). Previously the buffer was freed as a
side effect of ignoreline_pattern() called from ignoreline().
However, if only -B was specified, then ignoreline_pattern() was not
called and the lines were leaked. If both options were specified,
then ignoreline_pattern() was called before checking for a blank line
so that the second check was a use after free.
To fix, pull the free() out of ignoreline_pattern() and instead do it
up in change() so that is paired with preadline().
While here, simplify ignoreline() by checking for the -B and -I cases
individually without a separate clause for when both are set. Also,
do the cheaper check (-B) first, and remove a false comment (this
function is only called if at least one of -I or -B are specified).
Reviewed by: emaste
Reported by: GCC 12 -Wuse-after-free
Differential Revision: https://reviews.freebsd.org/D36822
Use TAILQ_FOREACH_SAFE to walk to list of children mnemonics to free
them instead of TAILQ_FOREACH.
Reviewed by: emaste
Reported by: GCC 12 -Wuse-after-free
Differential Revision: https://reviews.freebsd.org/D36821
GCC 12 thinks ct_visual_string can reuse a pointer after it has been
reallocated, but in this case the warning appears false.
Reviewed by: emaste
Differential Revision: https://reviews.freebsd.org/D36820
The _xrealloc() function prints pointer values for internal assertion
failures and in one case does so after it has freed the pointer.
Reviewed by: emaste
Differential Revision: https://reviews.freebsd.org/D36819
The debug traces for reference counting in ZFS use the pointer of the
owning object as a "tag" for references to check that when an object
drops a reference it had actually held one. In a couple of places ZFS
drops references after freeing the owning object. In userland GCC
realizes this is a use after free. However, since only the value of
the pointer is used, and it isn't indirected, the use is harmless.
Reviewed by: emaste
Differential Revision: https://reviews.freebsd.org/D36818
NLM 4 status values are a superset of the older NLM protocol so these
casts are safe while pacifying -Wenum-conversion warnings from GCC.
Reviewed by: rmacklem
Differential Revision: https://reviews.freebsd.org/D36816
The debug printf is intended to execute after the loop has ended to
log the selected file.
Reviewed by: imp, emaste
Reported by: GCC
Differential Revision: https://reviews.freebsd.org/D36815
libzfs depends on symbols defined in libzutil. For static linking
ld.bfd is less forgiving than lld, so rescue was failing to link when
using ld.bfd due to unresolved symbols from libzutil used by libzfs.
Reviewed by: imp, kib, emaste
Differential Revision: https://reviews.freebsd.org/D36811
This could use -fmax-errors on GCC, but tweaking the error limit is
unusual in the tree anyway. Just remove it.
Reviewed by: erj, imp, emaste
Differential Revision: https://reviews.freebsd.org/D36808
The Adler32 digest consists of two 16-bit words whose values are
calculated modulo 65521 (largest prime < 2^16). To avoid two division
instructions per byte, this version copies an optimization found in
zlib which defers the modulus until close to the point that the
intermediate sums can overflow 2^32. (zlib uses NMAX == 5552 for
this, this version uses 5000)
The bug is that in the deferred modulus case, the modulus was
only applied to the high word (and twice at that) but not to
the low word. The fix is to apply it to both words.
Reviewed by: jhibbits
Reported by: Miod Vallat <miod@openbsd.org>
Differential Revision: https://reviews.freebsd.org/D36798
GCC warns about the mismatched sizes on i386 where vm_paddr_t is 64
bits.
Reviewed by: imp, markj
Differential Revision: https://reviews.freebsd.org/D36750
r325506 (3cf8254f1e) extended struct pkthdr to add packet timestamp in
mbuf(9) chain. For example, cxgbe(4) and mlx5en(4) support this feature.
Use the timestamp for bpf(4) if it is available.
Reviewed by: hselasky, kib, np
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D36868
Ensure that an RFC3168 ECN reaction only occurs on non-SYN
segments.
Reviewed By: tuexen, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D36867
Flags should be M_WAITOK | M_ZERO instead of just M_ZERO
Reviewed by: markj
MFC after: 2 days
Sponsored by: Instituto de Pesquisas Eldorado (eldorado.org.br)
Differential Revision: https://reviews.freebsd.org/D36865
GCC 12.1.1_p20220625's -fanalyzer found and reported this.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Damian Szuberski <szuberskidamian@gmail.com>
Signed-off-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Closes#13975
Simple change to allow for the individual toggling of
RFC7323 window scaling and timestamp option.
Reviewed By: rrs, tuexen, glebius, guest-ccui, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D36863
This presents the existing information in a slightly more readable way.
Reviewed by: jrtc27, markj, jhb
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D36565
Only i386 and amd64 print the decoded syscall name in the backtrace.
This de-duplication facilitates further changes and adoption by other
platforms.
Reviewed by: jrtc27, markj, jhb
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D36565
Whilst febca0e643 added -N for both build and extract, it only fully
worked for extract; build would perform the actual tree build with
-DNO_ROOT and construct the intended METALOG, but the subsequent tarball
creation did not take this into account and just tarred up the the
directory as-is rather than using the METALOG. This resulted in the
permissions and ownership not being correct, as well as there being a
stray METALOG file in the tarball's root.
Reported by: avg
Reviewed by: avg, jhb, imp
Tested by: avg
Fixes: febca0e643 ("etcupdate: Add a -N flag to perform a NO_ROOT build")
Differential Revision: https://reviews.freebsd.org/D36845
Currently the installer is only started on the primary ("high level")
console. For systems where this is the video console and serial consoles
aren't of interest, and headless systems with just a serial console,
this works just fine, but for systems where both video and serial
consoles are present and meaningful this requires the user to select the
right primary console in loader, with the poor user experience of the
system appearing to hang if they leave the wrong one selected. This
notably differs from our multi-user behaviour of spawning getty on every
console, where the only issue with selecting the wrong primary console
is a quieter boot process until the login prompt appears (or the system
crashes).
Instead, use the newly-added runconsoles helper to run the installer on
every console (except for ttyv*, where only ttyv0 will be used). For
interactive installations, any of the consoles can be used, though only
one should be used at a time as no effort is made to avoid multiple
installations running at the same time clobbering each other. If the
Live CD option is selected, the other installers (which should, if the
user is well-behaved, be sitting at the welcome screen) will be killed.
If an automated install is in use, the primary console will be used to
display its output, and the others will direct the user to the primary
console.
Reviewed by: brooks, gjb
Differential Revision: https://reviews.freebsd.org/D36805
This separates out the install media-specific environment (creating
bsdinstall_etc) from actually running the installer on a given console.
This will be used by a future change to start the installer on multiple
consoles.
Reviewed by: brooks, gjb
Differential Revision: https://reviews.freebsd.org/D36803
This helper binary will run a given command on every on console, as
defined by /etc/ttys (except for ttyv*, where only ttyv0 will be used).
If one of the command processes exits, the rest will be killed. This
will be used by a future change to start the installer on multiple
consoles.
Reviewed by: brooks, imp, gjb
Differential Revision: https://reviews.freebsd.org/D36804
