nasty bug. The comparison to tset for an instruciton with the $0xf prefix
should have jumped down to the next non-prefix instruction test. Instead,
it jumped down to the next instruction test, which happened to be prefixed
instruction test. This test assumed that the earlier test had succeeded,
thus in some rare cases, this test would actually succeed, and we would
actually attempt to emulate a RDMSR instruction instead of the instruction
we were supposed to be emulating. Since %ecx usually did not contain a
valid MSR index at the time of the trap, this usually resulted in a #GP
due to an invalid MSR address and a lovely BTX fault when one tried to boot
the machine.
Noticed by: unfurl and others
The BM algorithm works by scanning the pattern from right to left,
and jumping as many characters as viable based on the text's mismatched
character and the pattern's already matched suffix.
This typically enable us to test only a fraction of the text's characters,
but has a worse performance than the straight-forward method for small
patterns. Because of this, the BM algorithm will only be used if the
pattern size is at least 4 characters.
Notice that this pre-matching is done on the largest substring of the
regular expression that _must_ be present on the text for a succesful
match to be possible at all.
For instance, "(xyzzy|grues)" will yield a null "must" substring, and,
therefore, not benefit from the BM algorithm at all. Because of the
lack of intelligence of the algorithm that finds the "must" string,
things like "charjump|matchjump" will also yield a null string. To
optimize that, "(char|match)jump" should be used.
The setup time (at regcomp()) for the BM algorithm will most likely
outweight any benefits for one-time matches. Given the slow regex(3)
we have, this is unlikely to be even perceptible, though.
The size of a regex_t structure is increased by 2*sizeof(char*) +
256*sizeof(int) + strlen(must)*sizeof(int). This is all inside the
regex_t's "guts", which is allocated dynamically by regcomp(). If
allocation of either of the two tables fail, the other one is freed.
In this case, the straight-forward algorithm is used for pre-matching.
Tests exercising the code path affected have shown a speed increase of
50% for "must" strings of length four or five.
API and ABI remain unchanged by this commit.
The patch submitted on the PR was not used, as it was non-functional.
PR: 14342
On unload, remove references from freelist to memory type defined by module.
Print a warning if module defines and allocate its own memory type, but
didn't free it all on unload.
Reviewed by: peter
system. Well, not really required if you know what you are doing, but
there's enough people that don't fit into this class that are getting
burned now that we need to say it is required. The actual message
says that one should treat it as if it was required to try to be
weasilly for the nitpickers amoung us :-)Killed by signal 2.
instead of bumping the recvAck counter by one, pretend that
all outstanding xmit packets are acknowleged, and restart
transmitting anew, with an empty (but halved) transmit window.
Put a lower bound on the adaptive timeout value.
Don't fake any file types, just set vap->va_type to IFTOVT(stb.st_mode).
If something does not report its mode, vap->va_type is set to VNON
accordingly.
use the current setting for tagged queueing when deciding whether or not to
print "Tagged Queueing Enabled" instead of using the device's actual
capabilities.
This is more consistent with the rest of the transfer rate display, which
relies on current settings, and is more consistent with the way we display
things on boot.
Reported by: Gustavo Vieira Goncalves Coelho Rios <kernel@tdnet.com.br>
Reviewed by: mjacob
