Andrew Turner [Mon, 28 Mar 2022 11:37:09 +0000 (12:37 +0100)]
Handle non-page aligned/sized memory in physmem
In some configurations the firmware may pass memory regions that are
not page sized or aligned, e.g. when using 16k pages on arm64. If this
is the case we will calculate many small regions because the alignment
is applied before being inserted. As we round the start up and end down
this will leave a 1 page hole between what should have been a single
region.
Fix by keeping the original alignment until we are just about to insert
the region into the avail array.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D34694
Andrew Turner [Mon, 4 Apr 2022 09:28:59 +0000 (10:28 +0100)]
Fix a coherent bus check in the arm64 busdma
In the arm64 busdma we have an internal flag to signal when a tag is
for a cache-coherent device. In this case we don't need to adjust the
size and alignment of allocated buffers to be within a cache line.
The cache line adjustment was incorrectly using the coherent flag
passed in to bus_dma_tag_create and not the internal flag. Fix it to
use the latter to reduce the memory usage slightly.
Reviewed by: bz
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D34763
Andrew Turner [Tue, 15 Mar 2022 17:05:48 +0000 (17:05 +0000)]
Move the arm64 DMAP creation to C
To simplify the creation of the direct map (DMAP) region on arm64 move
it from the pre-C code into pmap. This simplifies the DMAP creation
as we can use the notmal index macros, and should reduce the number
of pages needed to hold the level 1 tables to just those needed.
Reviewed by: alc, dch
Tested by: dch, kevans
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D34568
Warner Losh [Wed, 6 Apr 2022 03:35:27 +0000 (21:35 -0600)]
fix integer overflow bugs in *stosbt
68f57679d660 Fixed another class of integer overflows, but introduced a
boundary condition for 2-4s in ns conversion, 2-~4000s in us conversions
and 2-~4,000,000s in ms conversions. This was because we bogusly used
SBT_1S for the notion of 1 second, instead of the appropriate power of
10. To fix, just use the appropriate power of 10, which avoids these
overflows.
This caused some sleeps in ZFS to be on the order of an hour.
The correct logic is a lot simpler than the previous iteration. We
record the base fts_name to avoid having to worry about whether we
needed the root symlink name or not (as applicable), then we can simply
shift all of that logic to after path translation to make it less
fragile.
If we're copying to DNE, then we'll have swapped out the NULL root_stat
pointer and then attempted to recurse on it. The previously nonexistent
directory shouldn't exist at all in the new structure, so just back out
from that tree entirely and move on.
The tests have been amended to indicate our expectations better with
subdirectory recursion. If we copy A to A/B, then we expect to copy
everything from A/B/* into A/B/A/B, with exception to the A that we
create in A/B.
Reviewed by: bapt
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D34655
Warner Losh [Tue, 5 Apr 2022 19:40:01 +0000 (13:40 -0600)]
mii: Add opt_platform.h to all miibus drivers
miivar.h includes opt_platform.h. Make sure all the drivers that use the
miibus_if.h interface file have opt_platform.h as well. While some of
these may not, strictly speaking, need it, it's easier to include it
universally for miibus.
Warner Losh [Tue, 5 Apr 2022 19:09:27 +0000 (13:09 -0600)]
bwi: Mark write-only variable as __unused
We read TX_STATUS1 to acknowledge a TX interrupt. We don't use this
value for anything, so mark it as unused. We may be able to eliminate
this read, but since this hardware is poorly documented and difficult to
test, I'm leaving the read in place.
Warner Losh [Tue, 5 Apr 2022 19:07:20 +0000 (13:07 -0600)]
bwi: Remove write-only variable
bbp_atten is write-only: we don't use it later in
bwi_rf_calc_nrssi_slope_11b. However, we may also be able to eliminate
the read of this value from the device. However, The documationation for
this hardware is thing and I can't test this card so I've left the read
of this register in case it's required.
Mitchell Horne [Thu, 6 Jan 2022 19:40:16 +0000 (15:40 -0400)]
livedump: add event handler hooks
Add three hooks to the livedump process: before, after, and for each
block of dumped data. This allows, for example, quiescing the system
before the dump begins or protecting data of interest to ensure its
consistency in the final output.
Mitchell Horne [Tue, 23 Mar 2021 20:47:14 +0000 (17:47 -0300)]
Add new vnode dumper to support live minidumps
This dumper can instantiate and write the dump's contents to a
file-backed vnode.
Unlike existing disk or network dumpers, the vnode dumper should not be
invoked during a system panic, and therefore is not added to the global
dumper_configs list. Instead, the vnode dumper is constructed ad-hoc
when a live dump is requested using the new ioctl on /dev/mem. This is
similar in spirit to a kgdb session against the live system via
/dev/mem.
As described briefly in the mem(4) man page, live dumps are not
guaranteed to result in a usuable output file, but offer some debugging
value where forcefully panicing a system to dump its memory is not
desirable/feasible.
A future change to savecore(8) will add an option to save a live dump.
Reviewed by: markj, Pau Amma <pauamma@gundo.com> (manpages)
Discussed with: kib
MFC after: 3 weeks
Sponsored by: Juniper Networks, Inc.
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D33813
Mitchell Horne [Mon, 9 Aug 2021 17:21:07 +0000 (14:21 -0300)]
Split out dumper allocation from list insertion
Add a new function, dumper_create(), to allocate a dumper.
dumper_insert() will call this function and retains the existing
behaviour.
This is desirable for performing live dumps of the system. Here, there
is a need to allocate and configure a dumper structure that is invoked
outside of the typical debugger context. Therefore, it should be
excluded from the list of panic-time dumpers.
free_single_dumper() is made public and renamed to dumper_destroy().
Reviewed by: kib, markj
MFC after: 1 week
Sponsored by: Juniper Networks, Inc.
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D34068
Mitchell Horne [Thu, 24 Feb 2022 16:15:17 +0000 (12:15 -0400)]
mem(4): Improve ioctl section formatting
Add the missing .El which fixes the indentation of the memory range
definitions and operation. Add subsection headings to further clarify
this section. Do the same for the RETURN VALUES section, and mention
explicitly that MEM_EXTRACT_PADDR always returns zero.
Reviewed by: markj, 0mp, Pau Amma <pauamma@gundo.com>
Sponsored by: Juniper Networks, Inc.
Sponsored by: Klara, Inc.
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D34574
Rick Macklem [Tue, 5 Apr 2022 15:11:37 +0000 (08:11 -0700)]
nfscl: Add support for a NFSv4 AppendWrite RPC
For IO_APPEND VOP_WRITE()s, the code first does a
Getattr RPC to acquire the file's size, before it
can do the Write RPC.
Although NFS does not have an append write operation,
an NFSv4 compound can use a Verify operation to check
that the client's notion of the file's size is
correct before doing the Write operation.
This patch prepares the NFSv4 client for such an
RPC, which will be added in a future commit.
mrsas: if controller reset is in progress, refrain from firing DCMDs to
firmware in shutdown
If controller reset is in progress, at same time if system shutdown is
issued then corresponding shutdown function in driver will be invoked
where driver is waiting 15 seconds to complete the controller reset.
If the reset is not complteted within that time frame driver will go
ahead and fire cache flush and shutdown DCMDs which will end up
accessing the the queues which are not initialized due to undergoing
reset leads to FMU error in firmware.
Fix:
In shutdown function, if controller reset is not finished within 15
seconds than driver will return to the OS without firing any DCMDs.
When the unicode locales files where split off the non unicode locales
be_BY.ISO8859-5 and ca_IT.ISO8859-15 where non installed one due to
a typo, the other one was just missing its entry in the Makefile.
It was only notice a year later, by some postgres developpers as it was
breaking some of their tests.
Warner Losh [Tue, 5 Apr 2022 05:06:29 +0000 (23:06 -0600)]
linuxkpi: Move lkpi_pcim_iomap_devres_find to .c file
lkpi_pcim_iomap_devres_find encodes the size of struct pcim_iomap_devres
in the code, so move from .h to .c to move from client driver to
linuxkpi module.
Warner Losh [Tue, 5 Apr 2022 05:06:21 +0000 (23:06 -0600)]
linuxkpi: Move pci_alloc_irq_vectors to .c file
pci_alloc_irq_vectors encodes the size of struct msix_entry
into its code. Move from .h to .c to move this knowledge from
client modules to linuxkpi module.
Warner Losh [Tue, 5 Apr 2022 05:06:14 +0000 (23:06 -0600)]
linuxkpi: Move pci_request_region and _lkpi_pci_iomap into .c
Both pci_request_region and _lkpi_pci_iomap encode the size of struct
pci_mmio_region into their code. Move from .h to .c files to move that
knowledge from the client drivers into the linuxkpi module.
Warner Losh [Tue, 5 Apr 2022 05:05:43 +0000 (23:05 -0600)]
linuxkpi: Move class_create to .c file
class_create encodes the size of struct class into the generated
code. Move from .h file to .c file to move this knowledge from the
client modules that call this into the linuxkpi module.
Warner Losh [Tue, 5 Apr 2022 05:05:36 +0000 (23:05 -0600)]
linuxkpi: Move device_create_groups_vargs to linux_compat.c
device_create_groups_vargs encodes the size of struct device. Move
definition from .h to .c to move this size into the linuxkpi module
rather than encoding it in all client driver modules.
projects / FreeBSD / FreeBSD.git / log