1.44:
Fix typo which could lead into a double free
1.43:
Commands that create objects return a 24 bit object ID, so mask off the
high 8 bits of the value we extract, in case the firmware leaves junk there.
Hrvoje Popovski has seen this with newer firmware on a ConnectX 5 card,
which now works properly.
1.42:
Increase the completion queue size to prevent overflow. Under reasonably
unlikely circumstances - lots of single-fragment packets being sent, a
significant number of packets being received, while the interrupt handler
was unable to process the completion queue - the completion queue could
overflow, which would result in the interface locking up.
1.41:
Check if we've reached the end of the current mailbox before writing past
the end of it, rather than after. Now we can actually allocate queues
big enough to need multiple mailboxes.
1.40:
Don't call mcx_intr() from mcx_cmdq_poll(); this was a leftover from early
development that I forgot about, but turns out to be a potential race with
the actual interrupt handler.
1.39:
fix previous: use the correct offset for sq/rq creation, and don't
reset the mbox counter to 0 after calculating it.
1.38:
Add a helper function for writing physical addresses for queues into
command queue mailboxes, and use this for all queue setup commands.
Previously we just assumed the addresses would fit in the first mailbox,
which is currently true but may not be for much longer.
1.37:
(skipped)
1.36:
The event queue consumer counter also needs to be unsigned like the others.
1.35:
try to make if_baudrate look plausible.
this updates the eth proto capability map so it records the baudrate
against the different link types and their media, and then reads
it when the link state changes.
1.34:
(skipped)
- restrictions that existed before merging isaki-audio2 branch.
- better support for 6 channels hardware.
- audio layer's requirement.
This may help PR kern/54474.
bus_dmamap_destroy() can not be executed in soft interrupt context,
and msk_stop() can be called in soft interrupt context.
As such, move creation and destruction of tx dmamaps to attach() and
detach() functions.
http://mail-index.netbsd.org/source-changes/2020/04/16/msg116278.html
The reasoning turned out to be wrong; __KERNEL_RCSID() in header files
does *not* overwrite RCSID in main source files. The real problem is that
it inserts its RCSID into *every* object files. However, it can be still
useful even if heavily duplicated.
overwriting RCSID in main source files.
XXX
The first argument of __KERNEL_RCSID() is neglected for ELF. If we wish
to have RCSID of header files in kernel binary, we need something like
__FBSDID() macro in FreeBSD.
driver ata_bio hooks read parts of the xfer after ata_exec_xfer()
call in order to determine return value, change so that the hook
doesn't return any value - callers do not care already,
as all I/O requests are asynchronous
this problem was uncovered by recent change for wd(4) to not hold
wd mutex during ata_bio call, the interrupt for the xfer might
thus actually fire immediately
adjust also ata_exec_command driver hooks similarily - remove all
completion and waiting logic from drivers, upper layer ata code
using AT_WAIT/AT_POLL changed to call ata_wait_cmd() itself
PR kern/55169 by Nick Hudson
and pool_prime() (and their pool_cache_* counterparts):
- the pool_set*wat() APIs are supposed to specify thresholds for the count of
free items in the pool before pool pages are automatically allocated or freed
during pool_get() / pool_put(), whereas pool_sethardlimit() and pool_prime()
are supposed to specify minimum and maximum numbers of total items
in the pool (both free and allocated). these were somewhat conflated
in the existing code, so separate them as they were intended.
- change pool_prime() to take an absolute number of items to preallocate
rather than an increment over whatever was done before, and wait for
any memory allocations to succeed. since pool_prime() can no longer fail
after this, change its return value to void and adjust all callers.
- pool_setlowat() is documented as not immediately attempting to allocate
any memory, but it was changed some time ago to immediately try to allocate
up to the lowat level, so just fix the manpage to describe the current
behaviour.
- add a pool_cache_prime() to complete the API set.
We are returning an ACPI_INTEGER (= uint64_t), so it doesn't make
sense to handle more than 64 bits.
Apparently there are some ACPIs out there that ask for unreasonably
large widths here. Just reject those requests, rather than writing
past the caller's stack buffer.
Previously we attempted to fix this by copying byte by byte as large
as the caller asked, in order to avoid the undefined behaviour of
shifting past the size of ACPI_INTEGER, but that just turned a shift
(which might have been harmless on real machines) into a stack buffer
overflow (!).
ok msaitoh
