This isn't supposed to be here... looks like Mr. Jacob is experiencing
CVS lossage that keeps this directory from being properly pruned.

missing 'objs' target.

A stub makefile has to hang around (as far as *I* can tell) for makes to
continue to work.

Remove alpha 'mmclock' memory-mapped clock interfaces. In retrospect,
while it's sort of nice, it's sufficiently hard to do this properly
for the general case that, in the long run, it probably won't make sense
to support in the way i was planning to. (There are serious problems
with doing this, on multiprocessors.)

rearrange build structure so different architecture-specific libraries
can have separate shared library versions, by building in the arch
subdirectory.

Oops, forgot this file.

rearrange build structure so different architecture-specific libraries
can have separate shared library versions, by building in the arch
subdirectory.

makefile for alpha libarch

branches: 1.1.4;
libarch: Add expected symbols list.

PR lib/58838: shared libraries in base should all have expsym lists

Make 'i' unsigned to avoid signed/unsigned comparison warnings from lint.

branches: 1.3.4;
Remove clause 3 and 4 from TNF licenses

branches: 1.2.32;
De-lint, cleanup.

Userland interface to the bus window translation support now present
in the alpha sysarch(2). In a perfect world, these would go away in
favor of a generic bus space mapping framework, but XFree86 is a long
ways away from being able to use one.

Remove alpha 'mmclock' memory-mapped clock interfaces. In retrospect,
while it's sort of nice, it's sufficiently hard to do this properly
for the general case that, in the long run, it probably won't make sense
to support in the way i was planning to. (There are serious problems
with doing this, on multiprocessors.)

internal libalpha header, and functions to init and read from the
memory mapped clock.

Remove alpha 'mmclock' memory-mapped clock interfaces. In retrospect,
while it's sort of nice, it's sufficiently hard to do this properly
for the general case that, in the long run, it probably won't make sense
to support in the way i was planning to. (There are serious problems
with doing this, on multiprocessors.)

internal libalpha header, and functions to init and read from the
memory mapped clock.

Remove clause 3 and 4 from TNF licenses

branches: 1.2.32;
De-lint, cleanup.

Userland interface to the bus window translation support now present
in the alpha sysarch(2). In a perfect world, these would go away in
favor of a generic bus space mapping framework, but XFree86 is a long
ways away from being able to use one.

Fix -O0 build; <machine/bwx.h> should be included after ".arch ev56" pseudo-op.

branches: 1.7.12; 1.7.16;
Include <machine/alpha_cpu.h> for alpha_mb() definition.

branches: 1.6.2;
Add a cast of the shift count to int32_t, so that we don't try
to do int32_t << long, since ANSI C doesn't perform "balancing"
before the shift operation according to lint. Should not make a
difference, offset is limited to 0..3 anyway.

branches: 1.5.4;
Remove clause 3 and 4 from TNF licenses

branches: 1.4.10;
fix more caddr_t -> void * fallout.

Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.

De-lint, cleanup.

Userland interface to the bus window translation support now present
in the alpha sysarch(2). In a perfect world, these would go away in
favor of a generic bus space mapping framework, but XFree86 is a long
ways away from being able to use one.

Remove clause 3 and 4 from TNF licenses

branches: 1.4.30;
Fix multiple bugs:
* Select dense or sparse space based on the prefetchable flag, like the kernel
does. In particular, non-prefetchable mappings are forced to sparse space.
* If no window is matched, return MAP_FAILED, rather than falling through and
blowing chunks.
* Ignore addr_shift (set it to 0) for dense mappings.

De-lint, cleanup.

Pass the abst to alpha_pci_mem_unmap() so that the real size of the
mapped area can be unmapped.

branches: 1.1.4;
Userland interface to the bus window translation support now present
in the alpha sysarch(2). In a perfect world, these would go away in
favor of a generic bus space mapping framework, but XFree86 is a long
ways away from being able to use one.

Remove alpha 'mmclock' memory-mapped clock interfaces. In retrospect,
while it's sort of nice, it's sufficiently hard to do this properly
for the general case that, in the long run, it probably won't make sense
to support in the way i was planning to. (There are serious problems
with doing this, on multiprocessors.)

internal libalpha header, and functions to init and read from the
memory mapped clock.

Remove alpha 'mmclock' memory-mapped clock interfaces. In retrospect,
while it's sort of nice, it's sufficiently hard to do this properly
for the general case that, in the long run, it probably won't make sense
to support in the way i was planning to. (There are serious problems
with doing this, on multiprocessors.)

internal libalpha header, and functions to init and read from the
memory mapped clock.

Add "Remember to update distrib/sets..." lines (and NetBSD RCS IDs in
some cases).

Userland interface to the bus window translation support now present
in the alpha sysarch(2). In a perfect world, these would go away in
favor of a generic bus space mapping framework, but XFree86 is a long
ways away from being able to use one.