In run.c::assign(), assigning to $0 from $F, a field,
where F >= 2, produces an incoherent cell.
The assignment occurs in two steps, first the string value
and then the float. When the string value is assigned to $0,
setsval invalidates the fields. If FS hasn't changed, after
getfval rebuilds the fields, NF = 1 and F >= 2, therefore $F
is definitely uninitialized. The result is a float val of
0.0, producing a boolean false in the pattern expression.
Coercing a string comparison gives the expected result
because the incoherent cell has the correct string
value, which is not empty and evaluates to true.
when using records in BEGIN, we would read
from the field table before we read into it;
this ensures that the fields are an empty
string before we start touching their contents.
when reading a long line such as a dkim key in a
txt record, ndb calls Brdstr, which is limited
to the size of the buffer. This means we would
fail to parse the line from NDB, and bail out
early.
Increasing the buffer size allows us to read and
parse longer lines.
mdio interrupt command completion handling was broken,
as the interrupt handler would clear the mii status
register before mdiodone() sees it.
handle errors in miistatus() and miiane(), to not get
confused.
for unknown reasons, multiwrite is busted in usdhc:
sdhc: write error intr 10 stat ff88858e
usdhccmd: need to reset Datinhibit intr 10 stat ff88858e
usdhc: cmd 193a0027 arg 1e5b6b error intr 18010 stat ff88858f
i'm disabling it for now, adding a flag to the SDio struct.
The original intention was to put devskel in to the
kernel to detach what it provides from devsrv.
That is not a good reason, just move it to userspace.
auth/box has been changed to exec skelfs instead
of relying on '#z'.
do all the magic dwc3 specific initalization as well
as reset the hub and power on the phys.
with this, "usb start" command is not needed anymore
from u-boot.
Git9 was sloppy about telling git what commits we have.
We would list the commits at the tip of the branch, but not
walk down it, which means we would request too much data if
our local branches were ahead of the remote.
This patch changes that, sending the tips *and* the first
256 commits after them, so that git can produce a better
pack for us, with fewer redundant commits.
cwfs had an issue with iounit negotiation as a result
of the conversion to 9p2000 -- with the move to variable
size messages, the fixed message overhead decreased, but
the advertised message size was still adding the old
fixed overhead.
This meant that if the kernel negotiated the maximum io
size, cwfs would negotiate something larger than it
supported, and would hang up when an io of that size
was made.
In addition, the size of messages was stored in a short,
which means that negotiating an iounit larger than 16384
bytes would overflow the message count, and cause things
to fall over.
Finally, whle we're here, we clean up some duplicated
and unused constants.
Changing the user to none doesn't do much for us
here. For kernel drivers that check the user of the
current proc we'll be none, but anything from devmnt
will still be accessed using creds from the original
attachment. Instead, running with none can be done
by chaining with auth/none:
auth/none auth/box ...
