Alexander Motin [Wed, 21 Aug 2019 02:17:39 +0000 (02:17 +0000)]
Formalize NVMe controller consumer life cycle.
This fixes possible double call of fail_fn, for example on hot removal.
It also allows ctrlr_fn to safely return NULL cookie in case of failure
and not get useless ns_fn or fail_fn call with NULL cookie later.
Toomas Soome [Tue, 20 Aug 2019 19:31:11 +0000 (19:31 +0000)]
loader.efi: efipart should be more careful about constructing block device lists
The cd handles should be collected as list of partitions. Some systems also
provide base name for block device (like PciRoot(0x0)/Pci(0x5,0x0)), we need
to be careful about those.
To make out life a bit easier, we prepare not just an array of handles, but we
allocate pdinfo struct for each handle and devicepath, then we can simplify
our work to sort the devices.
Ed Maste [Tue, 20 Aug 2019 18:20:45 +0000 (18:20 +0000)]
makefs: add msdosfs (FAT) support
Add FAT support to makefs by copying some files from sys/fs/msdosfs/ and
updating others with changes from NetBSD.
The six files copied from sys/fs/msdosfs at r348251 and modified are:
denode.h direntry.h fat.h msdosfs_fat.c msdosfs_lookup.c msdosfsmount.h
I would prefer to avoid the duplication, but reluctance to doing so was
expressed in a previous review (D11197); for now copy the files and
revisit in the future.
Mark Johnston [Tue, 20 Aug 2019 17:52:12 +0000 (17:52 +0000)]
Use a sleepable lock for midistat functions.
Otherwise the mutex needs to be dropped when copying out the midistat
sbuf, leading to a race which allows one to read kernel memory beyond
the end of the sbuf buffer.
Reported and tested by: pho
Security: CVE-2019-5612
Dimitry Andric [Tue, 20 Aug 2019 17:39:32 +0000 (17:39 +0000)]
Pull in r368867 from upstream libc++ trunk (by Marshall Clow):
Rework recursive_timed_mutex so that it uses __thread_id instead of
using the lower-level __libcpp_thread_id. This is prep for fixing
PR42918. Reviewed as https://reviews.llvm.org/D65895
Pull in r368916 from upstream libc++ trunk (by Marshall Clow):
Fix thread comparison by making sure we never pass our special 'not a
thread' value to the underlying implementation. Fixes PR#42918.
This should fix std::thread::id::operator==() attempting to call
pthread_equal(3) with zero values.
Reported by: andrew@tao11.riddles.org.uk
PR: 239038, 239550
MFC after: 3 days
Brooks Davis [Tue, 20 Aug 2019 16:07:17 +0000 (16:07 +0000)]
Remove some compatability with Seventh Edition UNIX realloc().
In Seventh Edition UNIX, the last pointer passed to free() was
guaranteed to not actually have been freed allowing memory to be
"compacted" via the following pattern:
free(foo);
foo = realloc(foo, newsize);
Further, Andrew Koenig reports in "C Traps and Pitfalls" that the
original realloc() implementation required this pattern.
The C standard is clear that this is Undefined Behavior. Modern
allocators don't support it and no portable code could rely on it so
remove this support.
Note: the removed implementation contains an off-by-one error and if
an item isn't found on the freelist, then twice as much memory as the
largest possible allocation will be copied.
Kristof Provost [Tue, 20 Aug 2019 14:31:22 +0000 (14:31 +0000)]
netpfil tests: Add the set ToS test for ipfw and pf
This test tests the following:
- The firewall is able to set the tos bits
- The firewall is able to set the DSCP bits when EN bits is already set and
the EN bits remains unchanged.
- The firewall is able to drop the packets based on ToS value
usb_ethernet.h includes a number of mii headers, but only does so in
order to have struct mii_data available. However, it only really needs
a forward declaration of struct mii_data for use in pointer form for
the return type of a function prototype.
Custom kernel configuration that have usb and fdt enabled, but no miibus,
end up with compilation failures because miibus_if.h will not get
generated.
Due to the above, the following changes have been made to usb_ethernet.h:
* remove the inclusion of mii headers
* forward-declare struct mii_data
* include net/ifq.h to satify the need for complete struct ifqueue
Reviewed by: ian
Obtained from: Juniper Networks, Inc.
Differential Revision: https://reviews.freebsd.org/D21293
Conrad Meyer [Mon, 19 Aug 2019 22:57:03 +0000 (22:57 +0000)]
gdb(4): Pack 'info threads' responses into fewer packets
We suffer at least one round trip ACK latency every command / packet that
GDB has to send and receive, and the response format for 'info threads'
supports packing many threads IDs into a single packet, so do so.
Adds and uses a new API, gdb_txbuf_has_capacity(), which checks for a
certain number of bytes available in the outgoing txbuf.
On an example amd64 VM, the number of RTTs to transmit this list is reduced
by a factor of 110x. This is especially beneficial with recent GDB, which
seems to request the list at least twice during attach.
Jeff Roberson [Mon, 19 Aug 2019 22:25:28 +0000 (22:25 +0000)]
Permit vm_pager_has_page() to run with a shared lock. Introduce
VM_OBJECT_DROP/VM_OBJECT_PICKUP to handle functions that are called with
uncertain lock state.
Kyle Evans [Mon, 19 Aug 2019 19:01:59 +0000 (19:01 +0000)]
tuntap: belatedly add MODULE_VERSION for if_tun and if_tap
When tun/tap were merged, appropriate MODULE_VERSION should have been added
for things like modfind(2) to continue to do the right thing with the old
names.
Alan Somers [Mon, 19 Aug 2019 17:28:12 +0000 (17:28 +0000)]
ping: fix -Wformat-truncating warning with GCC
Increase buffer size for the string representation of n_time
ICMP timestamp is a 32-bit number. In pr_ntime(), number of minutes
and seconds is always 2 characters wide. Max. number of hours is 4
characters wide. The buffer size should be at least:
4 + 2 + 2 + 1 (':') + 1 (':') + 1 ('\0') = 11
Submitted by: Ján Sučan <sucanjan@gmail.com>
MFC after: 2 weeks
Sponsored by: Google, inc. (Google Summer of Code 2019)
Differential Revision: https://reviews.freebsd.org/D21325
Mark Johnston [Mon, 19 Aug 2019 16:29:51 +0000 (16:29 +0000)]
Fix netdump buffering after r348473.
nd_buf is used to buffer headers (for both the kernel dump itself and
for EKCD) before the final call to netdump_dumper(), which flushes
residual data in nd_buf. As a result, a small portion of the residual
data would be corrupted. This manifests when kernel dump compression
is enabled since both zstd and zlib detect the corruption during
decompression.
Reviewed by: cem
Differential Revision: https://reviews.freebsd.org/D21294
Emmanuel Vadot [Mon, 19 Aug 2019 14:33:22 +0000 (14:33 +0000)]
ti: sdhci: Correct voltage caps
ti,dual-volt property say that the eMMC support 1.8V and 3.3V not 3.0V
Use the correct caps for the mmc stack.
Note that the MMCHS_SD_CAPA register can only be written once after bootup
so if one is using a u-boot compiled with eMMC support (this is the default)
this code is a no-op but just in case someone have u-boot compiled without
eMMC support this make eMMC works when the kernel is booted.
Emmanuel Vadot [Mon, 19 Aug 2019 14:28:54 +0000 (14:28 +0000)]
arm64: a37x0_gpio: Use syscon instead of MMIO region
The fdt node for this driver is a simple-mfd and syscon compatible one
meaning that simplemfd will be the driver attached for it. The gpio driver
is attached to the 'gpio' subnode so use syscon_get_handle_default to
obtain the handle of the syscon from the parent device and use this
to read/write to the memory region.
Alan Somers [Mon, 19 Aug 2019 14:20:26 +0000 (14:20 +0000)]
ping6: revert r350857
Some socket options require root privileges to set. The old code did indeed
drop privileges at the earliest opportunity.
Submitted by: Ján Sučan <sucanjan@gmail.com>
MFC after: Never
Sponsored by: Google, Inc. (Google Summer of Code 2019)
Differential Revision: https://reviews.freebsd.org/D21319
Mateusz Guzik [Mon, 19 Aug 2019 14:11:54 +0000 (14:11 +0000)]
vfs: fix up r351193 ("stop always overwriting ->mnt_stat in VFS_STATFS")
fs-specific part of vfs_statfs routines only fill in small portion of the
structure. Previous code was always copying everything at a higher layer to
acoomodate it and this patch does the same.
'df' (no arguments) worked fine because the caller uses mnt_stat itself as the
target buffer, making all the copying a no-op for its own case.
'df /' and similar use a different consumer which passes its own buffer and
this is where you can run into trouble.
Reported by: cy Fixes: r351193
Sponsored by: The FreeBSD Foundation
Kristof Provost [Mon, 19 Aug 2019 10:48:27 +0000 (10:48 +0000)]
netpfil tests: Move pft_ping.py and sniffer.py to the common test directory
The pft_ping.py and sniffer.py tool is moved from tests/sys/netpfil/pf to
tests/sys/netpfil/common directory because these tools are to be used in
common for all the firewalls.
Inform D that C executed procctl(PROC_PDEATHSIG_CTL). Otherwise D
might allow B to exit before C is set up to receive a signal on the
parent exit. In this case, C waits forever for the signal and test
hangs.
PR: 237657
Reported and tested by: lwhsu
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
ptrace(PT_DETACH) requires stopped debuggee, otherwise it fails. When
the call fails, the C process is left as debuggee of the process D,
and might be killed too early if process D exits occurs fast enough.
Since pipes are not closed in the forked children, this resulted in
the test hanging, since no write occured from C to wake A.
PR: 237657
Reported and tested by: lwhsu
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Rick Macklem [Mon, 19 Aug 2019 00:29:05 +0000 (00:29 +0000)]
Add a vop_stdioctl() that performs a trivial FIOSEEKDATA/FIOSEEKHOLE.
Without this patch, when an application performed lseek(SEEK_DATA/SEEK_HOLE)
on a file in a file system that does not have its own VOP_IOCTL(), the
lseek(2) fails with errno ENOTTY. This didn't seem appropriate, since
ENOTTY is not listed as an error return by either the lseek(2) man page
nor the POSIX draft for lseek(2).
A discussion on freebsd-current@ seemed to indicate that implementing
a trivial algorithm that returns the offset argument for FIOSEEKDATA and
returns the file's size for FIOSEEKHOLE was the preferred fix.
http://docs.FreeBSD.org/cgi/mid.cgi?CAOtMX2iiQdv1+15e1N_r7V6aCx_VqAJCTP1AW+qs3Yg7sPg9wA
The Linux kernel appears to implement this trivial algorithm as well.
This patch adds a vop_stdioctl() that implements this trivial algorithm.
It returns errors consistent with vn_bmap_seekhole() and, as such, will
still return ENOTTY for non-regular files.
I have proposed a separate patch that maps errors not described by the
lseek(2) man page nor POSIX draft to EINVAL. This patch is under separate
review.
ng_ubt(4): do not attach Intel Wireless 8260/8265 in bootloader mode.
Add helper function for synchronous execution of HCI commands at probe
stage and use this function to check firmware state of Intel Wireless
8260/8265 bluetooth devices found in many post 2016 year laptops.
Attempt to initialize FreeBSD bluetooth stack while such a device is in
bootloader mode locks the adapter hardly so it requires power on/off
cycle to restore.
This change blocks ng_ubt attachment unless operational firmware is
loaded thus preventing the lock up.
Suppose that a binary was executed from tmpfs mount, and the text
vnode was reclaimed while the binary was still running. It is
possible during even the normal operations since tmpfs vnode'
vm_object has swap type, and no references on the vnode is held. Also
assume that the text vnode was revived for some reason. Then, on the
process exit or exec, unmapping of the text mapping tries to remove
the text reference from the vnode, but since it went from
recycle/instantiation cycle, there is no reference kept, and assertion
in VOP_UNSET_TEXT_CHECKED() triggers.
Fix this by keeping a use reference on the tmpfs vnode for each exec
reference. This prevents the vnode reclamation while executable map
entry is active.
Do it by adding per-mount flag MNTK_TEXT_REFS that directs
vop_stdset_text() to add use ref on first vnode text use, and
per-vnode VI_TEXT_REF flag, to record the need on unref in
vop_stdunset_text() on last vnode text use going away. Set
MNTK_TEXT_REFS for tmpfs mounts.
Reported by: bdrewery
Tested by: sbruno, pho (previous version)
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Alan Somers [Sun, 18 Aug 2019 17:12:06 +0000 (17:12 +0000)]
periodic: fix anticongestion for scripts run after security
Revision 316342, which introduced the anticongestion feature, failed to
consider that the periodic scripts are executed by a recursive invocation of
periodic. The recursive invocation wrongly cleaned up a temporary file that
should've been cleaned up only by the original invocation. The result is
that if the first script that requests an anticongestion sleep runs after
the security scripts, the sleep won't happen.
Fix this bug by delaying cleanup until the end of the original invocation.
Conrad Meyer [Sun, 18 Aug 2019 16:04:01 +0000 (16:04 +0000)]
random(4): Reorder configuration of random source modules
Move fast entropy source registration to the earlier
SI_SUB_RANDOM:SI_ORDER_FOURTH and move random_harvestq_prime after that.
Relocate the registration routines out of the much later randomdev module
and into random_harvestq.
This is necessary for the fast random sources to actually register before we
perform random_harvestq_prime() early in the kernel boot.
Michal Meloun [Sun, 18 Aug 2019 09:19:33 +0000 (09:19 +0000)]
Improve rk_pinctrl driver:
- add support for 'output-low', 'output-high', 'output-low' and
'output-enable' properties. These are use in RK3288 DT files
- add support for RK3288
- to reduce overall file size, use local macros for initialization
of pinctrl description structures.
Michal Meloun [Sun, 18 Aug 2019 09:11:43 +0000 (09:11 +0000)]
Improve rk_i2c driver:
- Properly handle IIC_M_NOSTOP and IIC_M_NOSTART flags.
- add polling mode, so driver can be used even if interrupts are not
enabled (this is necessary for proper support of PMICs).
- add support for RK3288
Michal Meloun [Sun, 18 Aug 2019 08:08:56 +0000 (08:08 +0000)]
Add method for getting of syscon handle from parent device.
If simple multifuction device also provides syscon interface, its
childern should be able to consume it. Due to this:
- declare coresponding method in syscon interface
- implement it in simple multifunction device driver
Jeff Roberson [Sun, 18 Aug 2019 07:06:31 +0000 (07:06 +0000)]
Encapsulate phys_avail manipulation in a set of simple routines. Add a
NUMA aware boot time memory allocator that will be used to allocate early
domain correct structures. Code partially submitted by gallatin.
Conrad Meyer [Sun, 18 Aug 2019 04:19:41 +0000 (04:19 +0000)]
OFED: Fix accidental double-copy of rdma_sdp.h in r351176
The mistake came about like this: the first attempt to commit was blocked by
a pre-commit hook due to missing SVN tags. svn revert doesn't delete new
files, I guess. While reapplying the fixed diff, the non-empty target file
was just concatenated with the new contents? Ugh. :-(
Mateusz Guzik [Sat, 17 Aug 2019 18:19:49 +0000 (18:19 +0000)]
fork: rework locking around do_fork
- move allproc lock into the func, it is of no use prior to it
- the code would lock p1 and p2 while holding allproc to partially
construct it after it gets added to the list. instead we can do the
work prior to adding anything.
- protect lastpid with procid_lock
As a side effect we do less work with allproc held.
Mateusz Guzik [Sat, 17 Aug 2019 17:42:01 +0000 (17:42 +0000)]
fork: stop skipping < 100 ids on wrap around
Code doing this is commented with a claim that these IDs are occupied by
daemons, but that's demonstrably false. To an extent the range is used by init
and kernel processes (and on sufficiently big machines it indeed is fully
populated).
On a sample box 40-way box the highest id in the range is 63. On a different one
it is 23. Just use the range.
Alan Somers [Sat, 17 Aug 2019 15:25:01 +0000 (15:25 +0000)]
ping: Move in_cksum() to a separate source file
This is a preparation step for adding ATF tests of in_cksum(), which has been
modified to operate on unaligned data. ping.o cannot be linked to the test
executable because both of them contain 'main' symbol.
Submitted by: Ján Sučan <sucanjan@gmail.com>
MFC after: 2 weeks
Sponsored by: Google, inc. (Google Summer of Code 2019)
Differential Revision: https://reviews.freebsd.org/D21288
Alan Somers [Sat, 17 Aug 2019 14:28:32 +0000 (14:28 +0000)]
ping6: Fix dnsdecode() bug introduced by r350859
Revision 350859 removed level of indirection that was needed for setting the
caller's `cp' pointer. dnsdecode() uses return value to indicate error or
success. It returns pointer to a buffer holding a decompressed DNS name or
NULL. The caller uses that value only to find out the result, not for accessing
the buffer.
We use the return value to propagate the new value of `cp' pointer to
the caller instead of using an output argument.
Tom Jones [Sat, 17 Aug 2019 06:47:43 +0000 (06:47 +0000)]
Reduce size of EFI_STAGING_SIZE to 32 on arm
Reduce the size of the EFI_STAGING area we allocate on arm to 32. On arm SBC
such as the NanoPi-NEOLTS the staging area allocation will fail on the 256MB
model with a staging size of 64.
Conrad Meyer [Sat, 17 Aug 2019 03:03:26 +0000 (03:03 +0000)]
SDP: SYSCTL_DECL SDP-wide sysctl node in header
This allows use of the shared _net_inet_sdp in more than one compilation
unit. (Nothing in-tree uses this today, but some of Isilon's out-of-tree
SDP enhancements add sysctls below the node.)
Warner Losh [Sat, 17 Aug 2019 02:36:42 +0000 (02:36 +0000)]
Fix small bug in wrapping introduced in r325955.
When local support was fixed, it introduced a minor bug in formatting. We don't
increment the lpos by enouogh, so lines are a little too long. Adjust to be
correct now with variable length srcprefix.
Warner Losh [Sat, 17 Aug 2019 02:35:53 +0000 (02:35 +0000)]
Move initializations of config earlier.
Inizialize global variables earlier in the process. It doesn't matter today, but
may in the future if we want to access these lists earlier in config's run.
Eric Joyner [Fri, 16 Aug 2019 23:33:44 +0000 (23:33 +0000)]
iflib: add iflib_deregister to help cleanup on exit
Commit message by Jake:
The iflib_register function exists to allocate and setup some common
structures used by both iflib_device_register and iflib_pseudo_register.
There is no associated cleanup function used to undo the steps taken in
this function.
Both iflib_device_deregister and iflib_pseudo_deregister have some of
the necessary steps scattered in their flow. However, most of the
necessary cleanup is not done during the error path of
iflib_device_register and iflib_pseudo_register.
Some examples of missed cleanup include:
the ifp pointer is not free'd during error cleanup
the STATE and CTX locks are not destroyed during error cleanup
the vlan event handlers are not removed during error cleanup
media added to the ifmedia structure is not removed
the kobject reference is never deleted
Additionally, when initializing the kobject class reference counter is
increased even though kobj_init already increases it. This results in
the class never being free'd again because the reference count would
never hit zero even after all driver instances are unloaded.
To aid in proper cleanup, implement an iflib_deregister function that
goes through the reverse steps taken by iflib_register.
Call this function during the error cleanup for iflib_device_register
and iflib_pseudo_register. Additionally call the function in the
iflib_device_deregister and iflib_pseudo_deregister functions near the
end of their flow. This helps reduce code duplication and ensures that
proper steps are taken to cleanup allocations and references in both the
regular and error cleanup flows.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Submitted by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed by: shurd@, erj@
MFC after: 3 days
Sponsored by: Intel Corporation
Differential Revision: https://reviews.freebsd.org/D21005