(Only for architectures that support it at all; on others,
__HAVE_MM_MD_DIRECT_MAPPED_PHYS/IO are still undefined and the
functions unimplemented.)
This gives modules like zfs an opportunity to use it.
While here, fix the one caller of mm_md_direct_mapped_phys that
ignored the return value (and make sure to call pmap_kremove/update
before uvm_km_free).
It turns out this code always sees __HAVE_MM_MD_DIRECT_MAPPED_PHYS as
disabled on amd64, for reasons that I shall address forthwith, but it
is enabled on aarch64 and so the MKZFS=yes build breaks.
As it happens, on x86 both _HARDKERNEL and _KERNEL get defined; see
the conditional in sys/rump/Makefile.rump that _refrains_ from
defining _RUMPKERNEL on x86.
So the only version of this code that has been tested is the one with
all of it included. But on, e.g., aarch64, we do not get _HARDKERNEL
here, and the code fails to build because some things use the field
struct spa::spa_deadman_cycid under _KERNEL when it is declared only
under _HARDKERNEL.
If there's a reason _not_ to use this in rump -- and it's not obvious
to me why -- then all access to the field needs to agree to use
_HARDKERNEL.
The PN533 is known to mishandle the USB toggle bit, causing replies to
be filtered out by the host controller. As a result, the kernel sees
a timed out operation.
Vendor errata suggests that userland applications should detect the
situation on read timeout, and write a dumy frame to resync the toggle bit.
NFC Tools's libnfc does just that, but in order to succeed, the dummy
frame write must not be reported as timed out.
We therefore introduce a new USB quirk for devices known to miss output
acks. When that occur, we pretend that the operation succeeded, leaving
userland the duty to check that everything went okay.
This workaround lets libnfc recover from interrupted communications
without the need te reboot the system.
- For UAREA, arrange layout same as FreeBSD for now to hopefully
ease porting woes. add some related macros
locore.S is incorrectly assuming same layout and seems painful
to change bspstore in startup
- use ia64_init_return same as FreeBSD
- change some "printf" to "panic" for incompleted items
context switching is still broken but maybe less so
Rather than write out a table for each architecture, rely on the C
compiler to define _LP64 for 64-bit ones, on the assumption that
anything not LP64 is ILP32, and on CTASSERTs to verify this
assumption so that if it's wrong it'll fail safely with a noisy build
failure.
Gives zfs half a chance of building on, e.g., powerpc.
Don't call desidhash_l on pcxl2 as it doesn't support it. QEMU emulates
this cpu and would trap on illegal instruction for the diag in desidhash_l
Allow a FPU to be missing... more fixes are likely here.
QEMU doesn't set C bit properly yet
wm_gmii_{hv,i82544}_{read,write}reg(). *_locked() functions are not
mii(4) API functions, so it's not required to keep the mii API. Change
the PHY register type from int to uint16_t. It also change the usage of
return value. It returns zero on success and non-zero on error.
- Check the return value of *_locked() function and treat it.
- Use *writereg_locked() function to reduce race condition in
wm_init_lcd_from_nvm().
- Add comment.
