changelog

linux-2.6.15.6

    ChangeLog

    Prehistory Martin Tischh鋟ser <[e-mail]>
	       Initial register-setting code (from proform-1.0).
    19971216   Richard Gooch <[e-mail]>
               Original version for /proc/mtrr interface, SMP-safe.
  v1.0
    19971217   Richard Gooch <[e-mail]>
               Bug fix for ioctls()'s.
	       Added sample code in Documentation/mtrr.txt
  v1.1
    19971218   Richard Gooch <[e-mail]>
               Disallow overlapping regions.
    19971219   Jens Maurer <[e-mail]>
               Register-setting fixups.
  v1.2
    19971222   Richard Gooch <[e-mail]>
               Fixups for kernel 2.1.75.
  v1.3
    19971229   David Wragg <[e-mail]>
               Register-setting fixups and conformity with Intel conventions.
    19971229   Richard Gooch <[e-mail]>
               Cosmetic changes and wrote this ChangeLog ;-)
    19980106   Richard Gooch <[e-mail]>
               Fixups for kernel 2.1.78.
  v1.4
    19980119   David Wragg <[e-mail]>
               Included passive-release enable code (elsewhere in PCI setup).
  v1.5
    19980131   Richard Gooch <[e-mail]>
               Replaced global kernel lock with private spinlock.
  v1.6
    19980201   Richard Gooch <[e-mail]>
               Added wait for other CPUs to complete changes.
  v1.7
    19980202   Richard Gooch <[e-mail]>
               Bug fix in definition of <set_mtrr> for UP.
  v1.8
    19980319   Richard Gooch <[e-mail]>
               Fixups for kernel 2.1.90.
    19980323   Richard Gooch <[e-mail]>
               Move SMP BIOS fixup before secondary CPUs call <calibrate_delay>
  v1.9
    19980325   Richard Gooch <[e-mail]>
               Fixed test for overlapping regions: confused by adjacent regions
    19980326   Richard Gooch <[e-mail]>
               Added wbinvd in <set_mtrr_prepare>.
    19980401   Richard Gooch <[e-mail]>
               Bug fix for non-SMP compilation.
    19980418   David Wragg <[e-mail]>
               Fixed-MTRR synchronisation for SMP and use atomic operations
	       instead of spinlocks.
    19980418   Richard Gooch <[e-mail]>
	       Differentiate different MTRR register classes for BIOS fixup.
  v1.10
    19980419   David Wragg <[e-mail]>
	       Bug fix in variable MTRR synchronisation.
  v1.11
    19980419   Richard Gooch <[e-mail]>
	       Fixups for kernel 2.1.97.
  v1.12
    19980421   Richard Gooch <[e-mail]>
	       Safer synchronisation across CPUs when changing MTRRs.
  v1.13
    19980423   Richard Gooch <[e-mail]>
	       Bugfix for SMP systems without MTRR support.
  v1.14
    19980427   Richard Gooch <[e-mail]>
	       Trap calls to <mtrr_add> and <mtrr_del> on non-MTRR machines.
  v1.15
    19980427   Richard Gooch <[e-mail]>
	       Use atomic bitops for setting SMP change mask.
  v1.16
    19980428   Richard Gooch <[e-mail]>
	       Removed spurious diagnostic message.
  v1.17
    19980429   Richard Gooch <[e-mail]>
	       Moved register-setting macros into this file.
	       Moved setup code from init/main.c to i386-specific areas.
  v1.18
    19980502   Richard Gooch <[e-mail]>
	       Moved MTRR detection outside conditionals in <mtrr_init>.
  v1.19
    19980502   Richard Gooch <[e-mail]>
	       Documentation improvement: mention Pentium II and AGP.
  v1.20
    19980521   Richard Gooch <[e-mail]>
	       Only manipulate interrupt enable flag on local CPU.
	       Allow enclosed uncachable regions.
  v1.21
    19980611   Richard Gooch <[e-mail]>
	       Always define <main_lock>.
  v1.22
    19980901   Richard Gooch <[e-mail]>
	       Removed module support in order to tidy up code.
	       Added sanity check for <mtrr_add>/<mtrr_del> before <mtrr_init>.
	       Created addition queue for prior to SMP commence.
  v1.23
    19980902   Richard Gooch <[e-mail]>
	       Ported patch to kernel 2.1.120-pre3.
  v1.24
    19980910   Richard Gooch <[e-mail]>
	       Removed sanity checks and addition queue: Linus prefers an OOPS.
  v1.25
    19981001   Richard Gooch <[e-mail]>
	       Fixed harmless compiler warning in include/asm-i386/mtrr.h
	       Fixed version numbering and history for v1.23 -> v1.24.
  v1.26
    19990118   Richard Gooch <[e-mail]>
	       Added devfs support.
  v1.27
    19990123   Richard Gooch <[e-mail]>
	       Changed locking to spin with reschedule.
	       Made use of new <smp_call_function>.
  v1.28
    19990201   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
	       Extended the driver to be able to use Cyrix style ARRs.
    19990204   Richard Gooch <[e-mail]>
	       Restructured Cyrix support.
  v1.29
    19990204   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
	       Refined ARR support: enable MAPEN in set_mtrr_prepare()
	       and disable MAPEN in set_mtrr_done().
    19990205   Richard Gooch <[e-mail]>
	       Minor cleanups.
  v1.30
    19990208   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
               Protect plain 6x86s (and other processors without the
               Page Global Enable feature) against accessing CR4 in
               set_mtrr_prepare() and set_mtrr_done().
    19990210   Richard Gooch <[e-mail]>
	       Turned <set_mtrr_up> and <get_mtrr> into function pointers.
  v1.31
    19990212   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
               Major rewrite of cyrix_arr_init(): do not touch ARRs,
               leave them as the BIOS have set them up.
               Enable usage of all 8 ARRs.
               Avoid multiplications by 3 everywhere and other
               code clean ups/speed ups.
    19990213   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
               Set up other Cyrix processors identical to the boot cpu.
               Since Cyrix don't support Intel APIC, this is l'art pour l'art.
               Weigh ARRs by size:
               If size <= 32M is given, set up ARR# we were given.
               If size >  32M is given, set up ARR7 only if it is free,
               fail otherwise.
    19990214   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
               Also check for size >= 256K if we are to set up ARR7,
               mtrr_add() returns the value it gets from set_mtrr()
    19990218   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
               Remove Cyrix "coma bug" workaround from here.
               Moved to linux/arch/i386/kernel/setup.c and
               linux/include/asm-i386/bugs.h
    19990228   Richard Gooch <[e-mail]>
	       Added MTRRIOC_KILL_ENTRY ioctl(2)
	       Trap for counter underflow in <mtrr_file_del>.
	       Trap for 4 MiB aligned regions for PPro, stepping <= 7.
    19990301   Richard Gooch <[e-mail]>
	       Created <get_free_region> hook.
    19990305   Richard Gooch <[e-mail]>
	       Temporarily disable AMD support now MTRR capability flag is set.
  v1.32
    19990308   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
	       Adjust my changes (19990212-19990218) to Richard Gooch's
	       latest changes. (19990228-19990305)
  v1.33
    19990309   Richard Gooch <[e-mail]>
	       Fixed typo in <printk> message.
    19990310   Richard Gooch <[e-mail]>
	       Support K6-II/III based on Alan Cox's <[e-mail]> patches.
  v1.34
    19990511   Bart Hartgers <[e-mail]>
	       Support Centaur C6 MCR's.
    19990512   Richard Gooch <[e-mail]>
	       Minor cleanups.
  v1.35
    19990707   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
               Check whether ARR3 is protected in cyrix_get_free_region()
               and mtrr_del(). The code won't attempt to delete or change it
               from now on if the BIOS protected ARR3. It silently skips ARR3
               in cyrix_get_free_region() or returns with an error code from
               mtrr_del().
    19990711   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
               Reset some bits in the CCRs in cyrix_arr_init() to disable SMM
               if ARR3 isn't protected. This is needed because if SMM is active
               and ARR3 isn't protected then deleting and setting ARR3 again
               may lock up the processor. With SMM entirely disabled, it does
               not happen.
    19990812   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
               Rearrange switch() statements so the driver accomodates to
               the fact that the AMD Athlon handles its MTRRs the same way
               as Intel does.
    19990814   Zolt醤 B鰏z鰎m閚yi <[e-mail]>
	       Double check for Intel in mtrr_add()'s big switch() because
	       that revision check is only valid for Intel CPUs.
    19990819   Alan Cox <[e-mail]>
               Tested Zoltan's changes on a pre production Athlon - 100%
               success.
    19991008   Manfred Spraul <[e-mail]>
    	       replaced spin_lock_reschedule() with a normal semaphore.
  v1.36
    20000221   Richard Gooch <[e-mail]>
               Compile fix if procfs and devfs not enabled.
	       Formatting changes.
  v1.37
    20001109   H. Peter Anvin <[e-mail]>
	       Use the new centralized CPU feature detects.

  v1.38
    20010309   Dave Jones <[e-mail]>
	       Add support for Cyrix III.

  v1.39
    20010312   Dave Jones <[e-mail]>
               Ugh, I broke AMD support.
	       Reworked fix by Troels Walsted Hansen <[e-mail]>

  v1.40
    20010327   Dave Jones <[e-mail]>
	       Adapted Cyrix III support to include VIA C3.

  v2.0
    20020306   Patrick Mochel <[e-mail]>
               Split mtrr.c -> mtrr/*.c
               Converted to Linux Kernel Coding Style
               Fixed several minor nits in form
               Moved some SMP-only functions out, so they can be used
                for power management in the future.
               TODO: Fix user interface cruft.