For extended attribute name max length, kernel filesystem-independant
code use either EXTATTR_MAXNAMELEN (BSD API) or XATTR_NAME_MAX (Linux API),
which are both defined as KERNEL_NAME_MAX and fits Linux limit of 255
without training \0.
UFS1 code had a lower limit that broke Linux compatibility. We can bump
the limit without sacrifying backward compatibility, because:
1) There is no API exposing this limit outside the kernel. Upper kernel
layers have a larger limit handle the increase without a hitch
2) Each attribute has its own backing store in the fileystem, the name
of the backing store matching the attribute name. A newer kernel can
create/read/write backing store for longer attribute names and will
have no problem with existing shorter names.
Softint-based interrupt handling is considered as a future direction
of the (network) device driver architecture in NetBSD. pq3etsec of
ppc is already implemented based on the architecture (unlike pq3etsec,
this change doesn't include softint-based if_start). In this
architecture, a hardware interrupt handler just schedules a softint
and the softint performs actual interrupt processing. It reduces
processing in hardware interrupt context and allows Layer 2 network
stack (e.g., bridge, vlan and even bpf) run in softint context,
which makes it easy to implement fine-grain locking in the layer.
This is an experimental implementation of the architecture in if_viof.
virtio introduces a new flag VIRTIO_F_PCI_INTR_SOFTINT. If a driver
of virtio sets it to sc_flags of virtio_softc, virtio calls
softint_schedule in virtio_intr instead of directly calling the
interrupt handler of the driver.
When VIOIF_SOFTINT_INTR is on, vioif doesn't use the existing softint
(vioif_rx_softint) that is called from vioif_rx_vq_done. Because
vioif_rx_softint already runs in softint context and another softint
isn't needed. This change actually improves performance in some cases.
The feature is disabled by default and enabled when SOFTINT_INTR is
set somewhere (normally in a kernel configuration).
- Fix a conditional
- Don't include bsd.lib.mk if we are not building libraries
(otherwise we end up with an empty library name that installed
bogus library files lib.a etc.)
More or less an Alliance Semiconductors ProMotion AT24 with some PCI-SBus
glue and 4MB framebuffer. No acceleration yet, just enough wsdisplay/vcons
setup and DAC programming to be functional.
A concrete result is enabling unpatched libpthread to run on the
rumprun stacks (e.g. Xen and bare metal) with a non-NetBSD scheduler.
Those schedulers hook into the existing _lwp_frobnitz() NetBSD syscall
interfaces (well, "syscall" interfaces in that scenario ;)
More specifically about the change itself:
1) instead of calling _lwp_makecontext() followed by _lwp_create()
and passing the entry point in ucontext_t (MD) through the calls, roll
the calls into pthread__makelwp() and allow alternate implementations
for that MI interface.
2) allow compile-time overriding of __lwp_gettcb_fast() or
__lwp_getprivate_fast, which are inline and leak MD scheduler/thread
details into libpthread
Additionally, two small nits:
I) define LIB=pthread before including mk.conf so that it's possible
to test for LIB==pthread in mk.conf
II) make it possible to leave out pthread_cancelstub.c. This is required
by the current implementation of rumprun-posix (i.e. rumprun on
POSIX hosts) due to symbol collisions. It needs to be fixed properly
some day, but for now allows an almost-correct libpthread to run.
I am sure @justin will be happy to explain the details ;)
no change to NetBSD
tested: anita+atf
0001. Do not recognize paths, mail folders, and pipes in mail addresses
by default. That avoids a direct command injection with syntactically
valid email addresses starting with |.
Such addresses can be specified both on the command line, the mail
headers (with -t) or in address lines copied over from previous
while replying.
This was assigned CVE-2014-7844 for some versions of BSD mailx. It is
documented behavior for Heirloom mailx, and was mentioned in an old
technical report about BSD mailx (which does not usually make its way
into operating system installations). The patch switches off this
processing and updates the documentation.
Added expandaddr option to explicitly enable this behavior.
0002. When invoking sendmail, prevent option processing for email
address arguments. This prevents changing e.g. the Postfix
configuration file in unexpected ways. This behavior was documented for
BSD mailx (sort of), but not for Heirloom mailx. We did not assign a
CVE to this because it is more of a missing feature, and code invoking
mailx needs adjustment in the caller as well.
Fixed.
0003. Make wordexp support mandatory. (No functional change.)
Fixed (replaced explicit shell pipe implementation).
0004. Prevent command execution in the expand function, which is IMHO
unexpected. (Not really required with patch 1, and there is still
information disclosure/DoS potential if this expansion occurs.) This is
a historic vulnerability already fixed in the Debian package,
retroactively assigned CVE-2004-2771:
Fixed (as part of the pipe replacement with wordexp).
to unwind the stack. Add a temporary workaround where we simply don't
allow the thread to exit (a kernel thread exit is a relatively uncommon
event in a rump kernel anyway).
