Andrew Rybchenko [Sun, 29 Nov 2015 05:42:49 +0000 (05:42 +0000)]
sfxge: add prefast annotation to common code return types
Using a typedef for common code return types (rather than "int")
allows the Prefast static analyser to understand when a function
has been successful (and thus when its postconditions must hold).
This greatly reduces then number of false positives reported by
prefast for error paths in common code functions.
Submitted by: Andy Moreton <amoreton at solarflare.com>
Sponsored by: Solarflare Communications, Inc.
MFC after: 2 days
Andrew Rybchenko [Sun, 29 Nov 2015 05:08:23 +0000 (05:08 +0000)]
sfxge: modify nvram update functions for uio platform to support RFID-selectable presets
Dynamic config partitions on boards that support RFID are divided into
a number of segments, each formatted like a partition, with header,
trailer and end tags. The first segment is the current active
configuration.
The segments are initialised by manftest and each contain a different
configuration e.g. firmware variant. The firmware can be instructed
via RFID to copy a segment over the first segment, hence changing the
active configuration. This allows ops to change the configuration of
a board prior to shipment using RFID.
Changes to the dynamic config may need to be written to all segments (in
particular firmware versions written by manftest) or just the first
segment (changes to the active configuration). See SF-111324-SW.
If only the first segment is written the code still needs to be aware of
the possible presence of subsequent segments as writing to a segment may
cause its size to increase, which would overwrite the subsequent
segments and invalidate them.
Boards that do not support RFID will only have one segment in their
dynamic config partition.
Submitted by: Paul Fox <pfox at solarflare.com>
Sponsored by: Solarflare Communications, Inc.
MFC after: 2 days
Differential Revision: https://reviews.freebsd.org/D4302
Michal Meloun [Sat, 28 Nov 2015 12:09:36 +0000 (12:09 +0000)]
ARM: Cumulative fixes for GIC
- fix detection of interrupt root controller
- allow (but warn) unsupported configuration bits
- dont send EOI for spurious interrupts
- print more informations for spurious interrupts
- use device_printf() where appropriate
Reviewed by: ian (earlier version)
Approved by: kib (mentor)
Remove sv_prepsyscall, sv_sigsize and sv_sigtbl members of the struct
sysent.
sv_prepsyscall is unused.
sv_sigsize and sv_sigtbl translate signal number from the FreeBSD
namespace into the ABI domain. It is only utilized on i386 for iBCS2
binaries. The issue with this approach is that signals for iBCS2 were
delivered with the FreeBSD signal frame layout, which does not follow
iBCS2. The same note is true for any other potential user if
sv_sigtbl. In other words, if ABI needs signal number translation, it
really needs custom sv_sendsig method instead.
In vm_pageout_grow_cache(), do not re-try the inactive queue when
active queue scan initiated write.
Re-trying from the inactive queue when doing active scan makes the
loop never end if number of domains is greater than 1 and inactive or
active scan cannot reach the target.
Reported and tested by: Andrew Gallatin <gallatin@netflix.com>
Reviewed by: alc
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Zbigniew Bodek [Fri, 27 Nov 2015 19:03:59 +0000 (19:03 +0000)]
Add helper to catch single step debug event and distinguish it from bkpt
Some architectures (including ARMv6/v7) do not have separate single step
events and cannot see difference between breakpoint and single step.
Add db_pc_is_singlestep() to avoid skipping instruction we stepped on
to trigger debug event.
This commit does not change the existing functionality but adds possibility
to implement custom db_pc_is_singlestep().
Reviewed by: imp
Submitted by: Zbigniew Bodek <zbb@semihalf.com>
Obtained from: Semihalf
Sponsored by: Juniper Networks Inc.
Differential Revision: https://reviews.freebsd.org/D4036
John Baldwin [Fri, 27 Nov 2015 18:58:26 +0000 (18:58 +0000)]
Add support to libkvm for reading vmcores from other architectures.
- Add a kvaddr_type to represent kernel virtual addresses instead of
unsigned long.
- Add a struct kvm_nlist which is a stripped down version of struct nlist
that uses kvaddr_t for n_value.
- Add a kvm_native() routine that returns true if an open kvm descriptor
is for a native kernel and memory image.
- Add a kvm_open2() function similar to kvm_openfiles(). It drops the
unused 'swapfile' argument and adds a new function pointer argument for
a symbol resolving function. Native kernels still use _fdnlist() from
libc to resolve symbols if a resolver function is not supplied, but cross
kernels require a resolver.
- Add a kvm_nlist2() function similar to kvm_nlist() except that it uses
struct kvm_nlist instead of struct nlist.
- Add a kvm_read2() function similar to kvm_read() except that it uses
kvaddr_t instead of unsigned long for the kernel virtual address.
- Add a new kvm_arch switch of routines needed by a vmcore backend.
Each backend is responsible for implementing kvm_read2() for a given
vmcore format.
- Use libelf to read headers from ELF kernels and cores (except for
powerpc cores).
- Add internal helper routines for the common page offset hash table used
by the minidump backends.
- Port all of the existing kvm backends to implement a kvm_arch switch and
to be cross-friendly by using private constants instead of ones that
vary by platform (e.g. PAGE_SIZE). Static assertions are present when
a given backend is compiled natively to ensure the private constants
match the real ones.
- Enable all of the existing vmcore backends on all platforms. This means
that libkvm on any platform should be able to perform KVA translation
and read data from a vmcore of any platform.
Tested on: amd64, i386, sparc64 (marius)
Differential Revision: https://reviews.freebsd.org/D3341
Zbigniew Bodek [Fri, 27 Nov 2015 18:17:53 +0000 (18:17 +0000)]
Implement simple ops for umass_disk
The initial IOCTL implementation supports reading disk physical
geometry.
Two additional functions were added. They allow reading/writing raw
data to the disk (default partition).
Submitted by: Wojciech Macek <wma@semihalf.com>
Obtained from: Semihalf
Sponsored by: Juniper Networks Inc.
Differential Revision: https://reviews.freebsd.org/D4143
Zbigniew Bodek [Fri, 27 Nov 2015 18:16:10 +0000 (18:16 +0000)]
Do not zero memory in umass_detach
The detach function is called very often, for example from
get_capacity function. We don't want to loose any pointers
here, so disable detaching for umass driver.
Submitted by: Wojciech Macek <wma@semihalf.com>
Obtained from: Semihalf
Sponsored by: Juniper Networks Inc.
Differential Revision: https://reviews.freebsd.org/D4141
Andrew Rybchenko [Fri, 27 Nov 2015 16:23:27 +0000 (16:23 +0000)]
sfxge: cleanup: report error on failure path in efx_vpd_hunk_verify
If the VPD is corrupt and contains an 'RV' keyword before the
END tag, then this function could return without setting the
return code to report the error.
Found by prefast.
Submitted by: Andy Moreton <amoreton at solarflare.com>
Sponsored by: Solarflare Communications, Inc.
MFC after: 2 days
Andrew Rybchenko [Fri, 27 Nov 2015 16:18:59 +0000 (16:18 +0000)]
sfxge: fix prefast warning in falconsiena_tx_qcreate
Keep prefast happy by returning the initial queue index
from falconsiena_tx_qcreate(). No change in behaviour, as
etxo_qcreate already zeros *addedp before the call.
Submitted by: Andy Moreton <amoreton at solarflare.com>
Sponsored by: Solarflare Communications, Inc.
MFC after: 2 days
Andrew Rybchenko [Fri, 27 Nov 2015 16:03:51 +0000 (16:03 +0000)]
sfxge: infer external port numbering for Pavia
Adjust external port mapping table to distinguish Pavia from Monza.
Now the presence of any 40G mode implies at least 2 outputs per
external port. So Pavia 4x10G ports are now mapped to 1,2,3,4;
Monza 4x10G ports map to 1,1,2,2 as before.
Submitted by: Richard Houldsworth <rhouldsworth at solarflare.com>
Sponsored by: Solarflare Communications, Inc.
MFC after: 2 days
Andrew Rybchenko [Fri, 27 Nov 2015 15:54:36 +0000 (15:54 +0000)]
sfxge: do not use unnamed union in siena_mc_combo_rom_hdr_t
GCC 4.2.1 used on FreeBSD 8 and 9 branches does not like unnamed
union member in the structure. It is not strictly required in head,
but nice to have to minimize difference with out-of-tree driver.
Sponsored by: Solarflare Communications, Inc.
MFC after: 2 days
Move the comment about resident pages preventing vnode from leaving
active list, into the header comment for vdrop(), which is the
function that decides whether to leave the vnode on the list. Note
that dirty page write-out in vinactive() is asynchronous.
Discussed with: alc
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Ed Maste [Fri, 27 Nov 2015 00:04:39 +0000 (00:04 +0000)]
vidfont: with vt(4) omit size from vidcontrol -f
When using syscons, vidfont extracts the font size from the filename
passes it to vidcontrol -f. In vt(4) mode the size argument is not
required, and some of the fonts in /usr/share/vt/fonts do not have the
size in the filename, which caused vidfont to fail. Thus, just omit the
size argument in vt(4) mode.
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Alan Cox [Thu, 26 Nov 2015 19:12:18 +0000 (19:12 +0000)]
Correct an error in vm_reserv_reclaim_contig(). In the highly unusual
case that the reservation contained "low", the starting position in the
popmap for the free page search was incorrectly calculated. The most
likely (and visible) symptom of this error was the assertion failure,
"vm_reserv_reclaim_contig: pa is too low".
Ian Lepore [Thu, 26 Nov 2015 17:26:52 +0000 (17:26 +0000)]
Rename sysctl node hw.imx6 to hw.imx. Move its definition to imx_machdep.c
so that code shared between imx5 and imx6 can work with OIDs under that node.
Add last_reset_status (integer) and last_reset_reason (string) OIDs that
provide info about the last chip reset (power-on, software reset, watchdog
timeout).
Alexander Motin [Thu, 26 Nov 2015 13:04:58 +0000 (13:04 +0000)]
One more round of port scanner rewrite.
- Make scan aborted by event restart immediately and infinitely.
- Improve handling of some loop events from firmware.
- Remove loop down timer, adding its functionality to scanner thread.
- Some more unification and simplification.
Michael Tuexen [Thu, 26 Nov 2015 09:25:20 +0000 (09:25 +0000)]
When receiving an SCTP/UDP packet and the interface performed
the UDP checksum computation and signals that it was OK,
clear this bit when passing the packet to SCTP. Since the
bits indicating a valid UDP checksum and a valid SCTP
checksum are the same, the SCTP stack would assume
that also an SCTP checksum check has been performed.
Adrian Chadd [Thu, 26 Nov 2015 01:58:50 +0000 (01:58 +0000)]
[ifconfig] handle IBSS mediatype correctly.
Right now net80211 is configured as type IBSS but then treats it
as mediatype ADHOC. This doesn't change that; it just correctly handles
being given a mediatype of IBSS.
Adrian Chadd [Thu, 26 Nov 2015 01:54:58 +0000 (01:54 +0000)]
[net80211] decode WPA cipher config '0' as "clear cipher config".
wpa_supplicant actually calls the wpa ioctl with cipher 0 as part
of the teardown process and this returns an ioctl error.
It's required as part of the (hopefully!) upcoming encrypted
IBSS support which does indeed do the above as part of interface
setup and then exits ungracefully when it gets an ioctl error.
(I'll fix wpa_supplicant in a later commit as part of other work.)
Bryan Drewery [Thu, 26 Nov 2015 01:54:09 +0000 (01:54 +0000)]
Always build libevent, which fixes the build for usr.sbin/ypldap after r291230.
The dependency for ypldap is only with WITH_NIS, but libevent is small
enough to not warrant the extra logic and maintenance cost here so just
always build it.
Bryan Drewery [Wed, 25 Nov 2015 20:38:17 +0000 (20:38 +0000)]
Replace DPSRCS that work fine in SRCS.
This is so that 'make depend' is not a required build step in these
files.
DPSRCS is overall unneeded. DPSRCS already contains SRCS, so anything
which can safely be in SRCS should be. DPSRCS is mostly just a way to
generate files that should not be linked into the final PROG/LIB. For
headers and grammars it is safe for them to be in SRCS since they will
be excluded during linking and installation.
The only remaining uses of DPSRCS are for generating .c or .o files that
must be built before 'make depend' can run 'mkdep' on the SRCS c files
list. A semi-proper example is in tests/sys/kern/acct/Makefile where a
checked-in .c file has an #include on a generated .c file. The
generated .c file should not be linked into the final PROG though since
it is #include'd. The more proper way here is just to build/link it in
though without DPSRCS. Another example is in sys/modules/linux/Makefile
where a shell script runs to parse a DPSRCS .o file that should not be
linked into the module. Beyond those, the need for DPSRCS is largely
unneeded, redundant, and forces 'make depend' to be ran. Generally,
these Makefiles should avoid the need for DPSRCS and define proper
dependencies for their files as well.
An example of an improper usage and why this matters is in usr.bin/netstat.
nl_defs.h was only in DPSRCS and so was not generated during 'make all',
but only during 'make depend'. The files including it lacked proper
depenencies on it, which forced running 'make depend' to workaround that
bug. The 'make depend' target should mostly be used for incremental build
help, not to produce a working build. This specific example was broken in
the meta build until r287905 since it does not run 'make depend'.
The gnu/lib/libreadline/readline case is fine since bsd.lib.mk has 'OBJS:
SRCS:M*.h' when there is no .depend file.
Bryan Drewery [Wed, 25 Nov 2015 19:49:55 +0000 (19:49 +0000)]
Define a LIB<NAME>DIR for every library that LIBADD provides.
This is going to be used to allow DIRDEPS to be bootstrapped off of
LIBADD/DPADD. It currently works for internal libraries which have a
DIR defined for them but also use the .a library from a src-mapped obj
directory. It can also be useful for using -L without a --sysroot per
LIBADD to use the OBJDIR version of the libraries.
I didn't review every LIBADD, so it is possible this is missing some.
Bryan Drewery [Wed, 25 Nov 2015 19:45:04 +0000 (19:45 +0000)]
META MODE: Avoid dirdep dependency on lib/libmd.
This avoids using the staged headers for sys/crypto/sha2/*.h, such as sha256.h,
which added an unneeded pre-build dependency on libmd to libcrypt. This
header is an INCS in lib/libmd, but found via .PATH in sys/crypto/sha2.
Since the libcrypt build was already using the in-src libmd headers
directly, just teach it how to find the sha256.h header as well.
Bryan Drewery [Wed, 25 Nov 2015 19:45:01 +0000 (19:45 +0000)]
META MODE: Fix MACHINE=host builds.
We need to not use -nostdinc since it breaks building of clang itself. Use
-isystem rather than -I/usr/include and -nostdinc which gets us using
the stage include directory before searching the real host headers.
This allows removing more of the -I hacks to get host headers since the
headers are no longer excluded. The -B seemed unneeded.
This fixes building of secure/lib/libcrypto which was looking at the
/usr/include/openssl/asn1.h header rather than the staged one.
This fixes building of clang which wants to find its own internal
headers in the STAGEDIR/usr/lib/clang/* path.
Bryan Drewery [Wed, 25 Nov 2015 19:44:51 +0000 (19:44 +0000)]
META MODE: Fix 'make bootstrap-tools'.
The main problem was bitrot after elftoolchain being swapped in for the
GNU toolchain.
This also reworks how the list of 'host allowed' libraries is determined
to only allow INTERNALLIBs, which is needed for libelftc to come in.
For usr.bin/readelf use the same hack, as libelf and libdward, to bring in
the needed sys/ headers for host builds. This has not yet been a problem due
to readelf not being built as a host tool in buildworld. This is possible
in the meta build though when building the toolchain.
Bryan Drewery [Wed, 25 Nov 2015 19:44:43 +0000 (19:44 +0000)]
META MODE: Don't create .meta files when symlinking sources into the obj directory.
Tracking these leads to situations where meta mode will consider the
file to be out of date if /bin/sh or /bin/ln are newer than the source
file. There's no reason for meta mode to do this as make is already
handling the rebuild dependency fine.
Bryan Drewery [Wed, 25 Nov 2015 19:44:28 +0000 (19:44 +0000)]
META MODE: Need to include the LEGACY_TOOLS directories into PATH.
This is just as Makefile.inc1 does it for these phases. Otherwise some
of the build tools are not found and used properly, such as 'make-roken'
in the kerberos5 build on older releases.
Bryan Drewery [Wed, 25 Nov 2015 19:13:45 +0000 (19:13 +0000)]
META MODE: Don't add staged DPADDs to DIRDEPS when bootstrapping gendirdeps.
This was finding libraries that were installed into DESTDIR/usr/lib,
where DESTDIR is the stage directory, and then adding in usr/lib to
DIRDEPS. Just exclude the STAGE_ROOT if defined.
Bryan Drewery [Wed, 25 Nov 2015 19:13:35 +0000 (19:13 +0000)]
META MODE: Rework circular dependency guard for librtld_db/libproc.
librtld_db only needs libutil.h to build, not the libproc library. So
it can safely use its header and allow libproc to depend on librtld_rb
to be built first to link. This is required after fixing ld --sysroot
in r291226.