GitLab

Define that bit 12 of the RISCOS_AddRAM flags indicates that the
supplied start, end, and sigbits values are in 4KB units instead of byte
units. This allows a 44 bit address space to be used, higher than the 40
bit LPAE limit.

The page list that RISCOS_AddRAM constructs will now store everything in
4KB page units, however any RAM above 4GB will currently be thrown away
when the list is later transferred to the PhysRamTable which the OS uses
at runtime.

Version 6.54. Tagged as 'Kernel-6_54'

Previously the CAM sat inside a fixed 16MB window, restricting it to
storing the details of 1 million pages, i.e. 4GB of RAM. Shuffle things
around a bit to allow this restriction to be removed: the CAM is now
located just above the IO region, and the CAM start address /
IO top will calculated appropriately during kernel init. This change
paves the way for us to support machines with over 4GB of RAM.

FixedAreasTable has also been removed, since it's no longer really
necessary (DAs can only be created between the top of application space
and the bottom of the used IO space, and it's been a long time since
we've had any fixed bits in the middle of there)

This adds initial support for the "long descriptor" MMU page table
format, which allows the CPU to (flexibly) use a 40-bit physical address
space.

There are still some features that need fixing (e.g. RISCOS_MapInIO
flags), and the OS doesn't yet support RAM above the 32bit limit, but
this set of changes is enough to allow for working ROMs to be produced.

Also, move MMUControlSoftCopy initialisation out of ClearWkspRAM, since
it's unrelated to whether the HAL has cleared the RAM or not.

Remove the lazy initialisation of PhysIllegalMask and instead manually
initialise it during MMU init. This fixes some situations where the lazy
initialisation doesn't work (PhysIllegalMask isn't in a zero-initialised
area of workspace, so if the HAL isn't doing a RAM clear then it could
be garbage)

Currently RISCOS_AccessPhysicalAddress allows the caller to specify the
permissions/properties of the mapped memory by directly specifying some
of the L1 page table entry flags. This will complicate things when
adding support for more page table formats, so change it so that
standard RISC OS page flags are used instead (like the alternate entry
point, RISCOS_AccessPhysicalAddressUnchecked, already uses).

Also, drop the "RISCOS_" prefix from RISCOS_AccessPhysicalAddress and
RISCOS_ReleasePhysicalAddress, and remove the references to these
routines from the HAL docs. These routines have never been exposed to
the HAL, so renaming them and removing them from the docs should make
their status clearer.

Version 6.52. Tagged as 'Kernel-6_52'

RISCOS_LogToPhys and OS_Memory 20 (compatibility page) changed to use
suitable subroutines for reading the page tables instead of accessing
them directly.

Change the DecodeL1/L2Entry routines so that instead of accepting a page
table entry as input, they accept a (suitable-aligned) logical address
and fetch the page table entry themselves. This helps insulate the
calling code from the finer details of the page table format.

ppn_to_physical, logical_to_physical, physical_to_ppn & ppn_to_physical
have now all been changed to accept/receive 64bit physical addresses in
R8,R9 instead of a 32bit address in R5. However, where a phys addr is
being provided as an input, they may currently only pay attention to the
bottom 32 bits of the address.

This ensures that the debug macros may be freeley placed without
concern for if LR and CPSR contain important values, and avoid the
of code only working because the debug code changes the flags with
the effect of a bug appearing when debugging is enabled.

Version 6.49. Tagged as 'Kernel-6_49'

Changes in Kernel-6_44 left an undefined symbol in s.HAL in DebugTerminal_Rdch. These mods resolve that, and also add a note in s.Kernel to reflect this usage.

Version 6.45. Tagged as 'Kernel-6_45'

If HAL+OS are contiguous, the combined image is initially mapped as
OSAP_None in order to allow for decompression. Fix the discontiguous
case to use the same permissions

Detail:

- s/HAL - Fix ADD v. SUB muddle that could prevent addresses from being rounded down correctly. Fix incorrect logical address being returned to caller on pre-ARMv6 machines due to PageTableSync corrupting a1.
- s/NewReset - Initialising the CMOS RAM cache while in the middle of setting up the processor vectors feels a bit silly. Move the code to just afterwards so that it feels a bit safer, and so that early crashes are easier to debug (processor vectors in stable state)

Admin:

Tested on Iyonix.
Fixes ROM softload failure reported on forums:
https://www.riscosopen.org/forum/forums/11/topics/14749

Version 6.23. Tagged as 'Kernel-6_23'

To achieve this:
* DecodeL1Entry and DecodeL2Entry return 64-bit physical addresses in
  r0 and r1, with additional return values shuffled up to r2 and r3
* DecodeL1Entry now returns the section size, so callers can distinguish
  section- from supersection-mapped memory
* PhysAddrToPageNo now accepts a 64-bit address (though since the physical
  RAM table is currently still all 32-bit, it will report any top-word-set
  addresses as being not in RAM)

Version 6.22. Tagged as 'Kernel-6_22'

Add a new reason code, OS_Memory 22, equivalent to OS_Memory 14, but
accepting a 64-bit physical address in r1/r2. Current ARM architectures can
only express 40-bit or 32-bit physical addresses in their page tables
(depending on whether they feature the LPAE extension or not) so unlike
OS_Memory 14, OS_Memory 22 can return an error if an invalid physical
address has been supplied. OS_Memory 15 should still be used to release a
temporary mapping, whether you claimed it using OS_Memory 14 or OS_Memory 22.

The logical memory map has had to change to accommodate supersection mapping
of the physical access window, which needs to be 16MB wide and aligned to a
16MB boundary. This results in there being 16MB less logical address space
available for dynamic areas on all platforms (sorry) and there is now a 1MB
hole spare in the system address range (above IO).

The internal function RISCOS_AccessPhysicalAddress has been changed to
accept a 64-bit physical address. This function has been a candidate for
adding to the kernel entry points from the HAL for a long time - enough that
it features in the original HAL documentation - but has not been so added
(at least not yet) so there are no API compatibility issues there.

Requires RiscOS/Sources/Programmer/HdrSrc!2

This is facilitated by two extended calls. From the HAL:
* RISCOS_MapInIO64 allows the physical address to be specified as 64-bit

From the OS:
* OS_Memory 21 acts like OS_Memory 13, but takes a 64-bit physical address

There is no need to extend RISCOS_LogToPhys, instead we change its return
type to uint64_t. Any existing HALs will only read the a1 register, thereby
narrowing the result to 32 bits, which is fine because all existing HALs
only expected a 32-bit physical address space anyway.

Internally, RISCOS_MapInIO has been rewritten to detect and use supersections
for IO regions that end above 4GB. Areas that straddle the 4GB boundary should
also work, although if you then search for a sub-area that doesn't, it won't
find a match and will instead map it in again using vanilla sections - this is
enough of an edge case that I don't think we need to worry about it too much.

The rewrite also conveniently fixes a bug in the old code: if the area being
mapped in went all the way up to physical address 0xFFFFFFFF (inclusive) then
only the first megabyte of the area was actually mapped in due to a loop
termination issue.

Requires RiscOS/Sources/Programmer/HdrSrc!2

Without this, some CPUs (including Cortex-A72) take the undefined instruction
trap when encountering any CP15-encoded barrier instructions.

Requires RiscOS/Sources/Programmer/HdrSrc!1

Version 6.20. Tagged as 'Kernel-6_20'

Detail:
  s/HAL - OS_Hardware 3 (remove HAL device) will now re-scan the device list for the device following the Service_Hardware call, so that the device list won't become corrupt if the service call triggers addition/removal of devices.
Admin:
  Tested on iMX6
  *HDMIOff now correctly removes the HDMI audio device and SoundDMA's software mixer device (SoundDMA removes mixer in response to the HDMI audio device vanishing, but re-entrancy bug meant that the HDMI device was left on the list)
  Note that this only covers re-entrancy via Service_Hardware. OS_Hardware 2/3/4/5 are not re-entrant from other locations (e.g. IRQ handlers or memory allocation service calls).


Version 6.12. Tagged as 'Kernel-6_12'

Detail:
  To make it easier to support arbitrary complexity keyboard controllers (eg. USB via DWCDriver on the Pi) have the kernel do the early keyboard recovery key press detection instead of the HAL.
  During the first pass those modules used for reading the keyboard are started, ignoring the CMOS frugal bits.
  The keyboard is then scanned for 3s, during which time the RAM is cleared (unless the HAL indicated it has already been done).
  During the second pass the remaining modules are started respecting the CMOS frugal bits. Any which were already started in the first pass are inserted into the new chain, so the keyboard is reset once and only once.

  Boot times, with a 300cs key scan time in NewReset.
  Risc PC with 160MB RAM (128+32+0).
  Times from turning on power to initial "beep", using a stopwatch.
                RISC OS 3.70 RISC OS 5.22 This OS
  ARM610        12.5         10.4         10.3
  ARM710        11.8         10.2         9.7
  StrongARM 233 11.1         9.5          8.4

  In NewReset.s:
  Remove old KbdScan code (leave Reset_IRQ_Handler for IIC only)
  If HAL_KbdScanDependencies returns a null string then present KbdDone flag and skip to full init.
  A few vestiges of soft resets removed.
  Do RAM clear when waiting for INKEY (being careful not to trash the running modules...).
  Clearing just the freepool on a 2GB Titanium cleared 7EFD6 pages (99.2%).

  In ModHand.s:
  2nd pass need to sneaky renumber the nodes (so *ROMModules is in the right order, frugal bits line up) without resetting the chain

  In HAL.s:
  Change ClearPhysRAM to ClearWkspRAM, such that it only clears the kernel workspace rather than all RAM. The bulk of the RAM is cleared during the keyboard scan by new function ClearFreePoolSection.
  Add a variant of Init_MapInRAM which clears the mapped in RAM too (as these very early claims will not be in the free pool when the RAM is cleared later).
  Remove HAL keyboard scan setup & IRQ handler.
  Fix bug in HALDebugHexTX2, the input value needs pre-shifting by 16b before continuing.

  In GetAll.s, PMF/osbyte.s:
  Use Hdr:Countries and Hdr:OsBytes for constants.

  In PMF/key.s, PMF/osinit.s:
  Relocate the key post init from PostInit to KeyPostInit.
  Changed PostInit to not tail call KeyPostInit so they can be called independently.

  In hdr/KernelWs:
  Improve comments, add InitWsStart label to refer to.

  In hdr/HALEntries:
  Add HAL_KbdScanDependencies.
  Delete KbdFlag exports.
  Took the opportunity to reorder some of the higher numbered HAL entries and re-grouping, specifically (112,120) (84,106,108,117).
Admin:
  Tested on an ARM6/ARM7/SA Risc PC, BeagleBoard xM, Iyonix, Pandaboard ES, Wandboard Quad, IPEGv5, Titanium, Pi 2 and 3.
  Requires corresponding HAL change.
  Submission for USB bounty.

Version 5.89. Tagged as 'Kernel-5_89'

Detail:
  s/ARMops, s/HAL - Add workarounds for some of the scary errata that were previously weren't dealing with (720013, 716151, 714068)
Admin:
  Tested on Raspberry Pi 1


Version 5.72. Tagged as 'Kernel-5_72'

Detail:
  Modern ARMs (ARMv6+) introduce the possibility for the page table walk hardware to make use of the data cache(s) when performing memory accesses. This can significantly reduce the cost of a TLB miss on the system, and since the accesses are cache-coherent with the CPU it allows us to make the page tables cacheable for CPU (program) accesses also, improving the performance of page table manipulation by the OS.
  Even on ARMs where the page table walk can't use the data cache, it's been measured that page table manipulation operations can still benefit from placing the page tables in write-through or bufferable memory.
  So with that in mind, this set of changes updates the OS to allow cacheable/bufferable page tables to be used by the OS + MMU, using a system-appropriate cache policy.
  File changes:
  - hdr/KernelWS - Allocate workspace for storing the page flags that are to be used by the page tables
  - hdr/OSMem - Re-specify CP_CB_AlternativeDCache as having a different behaviour on ARMv6+ (inner write-through, outer write-back)
  - hdr/Options - Add CacheablePageTables option to allow switching back to non-cacheable page tables if necessary. Add SyncPageTables var which will be set {TRUE} if either the OS or the architecture requires a DSB after writing to a faulting page table entry.
  - s/ARM600, s/VMSAv6 - Add new SetTTBR & GetPageFlagsForCacheablePageTables functions. Update VMSAv6 for wider XCBTable (now 2 bytes per element)
  - s/ARMops - Update pre-ARMv7 MMU_Changing ARMops to drain the write buffer on entry if cacheable pagetables are in use (ARMv7+ already has this behaviour due to architectural requirements). For VMSAv6 Normal memory, change the way that the OS encodes the cache policy in the page table entries so that it's more compatible with the encoding used in the TTBR.
  - s/ChangeDyn - Update page table page flag handling to use PageTable_PageFlags. Make use of new PageTableSync macro.
  - s/Exceptions, s/AMBControl/memmap - Make use of new PageTableSync macro.
  - s/HAL - Update MMU initialisation sequence to make use of PageTable_PageFlags + SetTTBR
  - s/Kernel - Add PageTableSync macro, to be used after any write to a faulting page table entry
  - s/MemInfo - Update OS_Memory 0 page flag conversion. Update OS_Memory 24 to use new symbol for page table access permissions.
  - s/MemMap2 - Use PageTableSync. Add routines to enable/disable cacheable pagetables
  - s/NewReset - Enable cacheable pagetables once we're fully clear of the MMU initialision sequence (doing earlier would be trickier due to potential double-mapping)
Admin:
  Tested on pretty much everything currently supported
  Delivers moderate performance benefits to page table ops on old systems (e.g. 10% faster), astronomical benefits on some new systems (up to 8x faster)
  Stats: https://www.riscosopen.org/forum/forums/3/topics/2728?page=2#posts-58015


Version 5.71. Tagged as 'Kernel-5_71'

Detail:
  For a while we've known that the correct way of doing cache maintenance on ARMv6+ (e.g. when converting a page from cacheable to non-cacheable) is as follows:
  1. Write new page table entry
  2. Flush old entry from TLB
  3. Clean cache + drain write buffer
  The MMU_Changing ARMops (e.g. MMU_ChangingEntry) implement the last two items, but in the wrong order. This has caused the operations to fall out of favour and cease to be used, even in pre-ARMv6 code paths where the effects of improper cache/TLB management perhaps weren't as readily visible.
  This change re-specifies the relevant ARMops so that they perform their sub-operations in the correct order to make them useful on modern ARMs, updates the implementations, and updates the kernel to make use of the ops whereever relevant.
  File changes:
  - Docs/HAL/ARMop_API - Re-specify all the MMU_Changing ARMops to state that they are for use just after a page table entry has been changed (as opposed to before - e.g. 5.00 kernel behaviour). Re-specify the cacheable ones to state that the TLB invalidatation comes first.
  - s/ARM600, s/ChangeDyn, s/HAL, s/MemInfo, s/VMSAv6, s/AMBControl/memmap - Replace MMU_ChangingUncached + Cache_CleanInvalidate pairs with equivalent MMU_Changing op
  - s/ARMops - Update ARMop implementations to do everything in the correct order
  - s/MemMap2 - Update ARMop usage, and get rid of some lingering sledgehammer logic from ShuffleDoublyMappedRegionForGrow
Admin:
  Tested on pretty much everything currently supported


Version 5.70. Tagged as 'Kernel-5_70'

Add new ARMops. Add macros which map the ARMv7/v8 cache/TLB maintenance mnemonics (as featured in recent ARM ARMs) to MCR ops.

Detail:
  - Docs/HAL/ARMop_API - Document the new ARMops. These ops are intended to help with future work (DMA without OS_Memory 0 "make temp uncacheable", and minimising cache maintenance when unmapping pages) and aren't in use just yet.
  - hdr/Copro15ops - Add new macros for ARMv7+ which map the mnemonics seen in recent ARM ARMs to the corresponding MCR ops. This should make things easier when cross-referencing docs and reduce the risk of typos.
  - hdr/KernelWS - Shuffle kernel workspace a bit to make room for the new ARMops
  - hdr/OSMisc - Expose new ARMops via OS_MMUControl 2
  - s/ARMops - Implement the new ARMops. Change the ARMv7+ ARMops to use the new mnemonic macros. Also get rid of myDSB / myISB usage from ARMv7+ code paths; use DSB/ISB/etc. directly to ensure correct behaviour
  - s/HAL - Mnemonic + ISB/DSB updates. Change software RAM clear to do 16 bytes at a time for kernel workspace instead of 32 to allow the kernel workspace tweaks to work.
Admin:
  Binary diff shows that mnemonics map to the original MCR ops correctly
  Note: Raspberry Pi builds will now emit lots of warnings due to increased DSB/ISB instruction use. However it should be safe to ignore these as they should only be present in v7+ code paths.
  Note: New ARMops haven't been tested yet, will be disabled (or at least hidden from user code) in a future checkin


Version 5.68. Tagged as 'Kernel-5_68'

Detail:
  This set of changes:
  * Refactors page table entry encoding/decoding so that it's (mostly) performed via functions in the MMU files (s.ARM600, s.VMSAv6) rather than on an ad-hoc basis as was the case previously
  * Page table entry encoding/decoding performed during ROM init is also handled via the MMU functions, which resolves some cases where the wrong cache policy was in use on ARMv6+
  * Adds basic support for shareable pages - on non-uniprocessor systems all pages will be marked as shareable (however, we are currently lacking ARMops which broadcast cache maintenance operations to other cores, so safe sharing of cacheable regions isn't possible yet)
  * Adds support for the VMSA XN flag and the "privileged ROM" access permission. These are exposed via RISC OS access privileges 4 and above, taking advantage of the fact that 4 bits have always been reserved for AP values but only 4 values were defined
  * Adds OS_Memory 17 and 18 to convert RWX-style access flags to and from RISC OS access privelege numbers; this allows us to make arbitrary changes to the mappings of AP values 4+ between different OS/hardware versions, and allows software to more easily cope with cases where the most precise AP isn't available (e.g. no XN on <=ARMv5)
  * Extends OS_Memory 24 (CheckMemoryAccess) to return executability information
  * Adds exported OSMem header containing definitions for OS_Memory and OS_DynamicArea
  File changes:
  - Makefile - export C and assembler versions of hdr/OSMem
  - Resources/UK/Messages - Add more text for OS_Memory errors
  - hdr/KernelWS - Correct comment regarding DCacheCleanAddress. Allocate workspace for MMU_PPLTrans and MMU_PPLAccess.
  - hdr/OSMem - New file containing exported OS_Memory and OS_DynamicArea constants, and public page flags
  - hdr/Options - Reduce scope of ARM6support to only cover builds which require ARMv3 support
  - s/AMBControl/Workspace - Clarify AMBNode_PPL usage
  - s/AMBControl/growp, mapslot, mapsome, memmap - Use AreaFlags_ instead of AP_
  - s/AMBControl/main, memmap - Use GetPTE instead of generating page table entry manually
  - s/ARM600 - Remove old coments relating to lack of stack. Update BangCam to use GetPTE. Update PPL tables, removing PPLTransL1 (L1 entries are now derived from L2 table on demand) and adding a separate table for ARM6. Implement the ARM600 versions of the Get*PTE ('get page table entry') and Decode*Entry functions
  - s/ARMops - Add Init_PCBTrans function to allow relevant MMU_PPLTrans/MMU_PCBTrans pointers to be set up during the pre-MMU stage of ROM init. Update ARM_Analyse to set up the pointers that are used post MMU init.
  - s/ChangeDyn - Move a bunch of flags to hdr/OSMem. Rename the AP_ dynamic area flags to AreaFlags_ to avoid name clashes and confusion with the page table AP_ values exported by Hdr:MEMM.ARM600/Hdr:MEMM.VMSAv6. Also generate the relevant flags for OS_Memory 24 so that it can refer to the fixed areas by their name instead of hardcoding the permissions.
  - s/GetAll - GET Hdr:OSMem
  - s/HAL - Change initial page table setup to use DA/page flags and GetPTE instead of building page table entries manually. Simplify AllocateL2PT by removing the requirement for the user to supply the access perimssions that will be used for the area; instead for ARM6 we just assume that cacheable memory is the norm and set L1_U for any L1 entry we create here.
  - s/Kernel - Add GetPTE macro (for easier integration of Get*PTE functions) and GenPPLAccess macro (for easy generation of OS_Memory 24 flags)
  - s/MemInfo - Fixup OS_Memory 0 to not fail on seeing non-executable pages. Implement OS_Memory 17 & 18. Tidy up some error generation. Make OS_Memory 13 use GetPTE. Extend OS_Memory 24 to return (non-) executability information, to use the named CMA_ constants generated by s/ChangeDyn, and to use the Decode*Entry functions when it's necessary to decode page table entries.
  - s/NewReset - Use AreaFlags_ instead of AP_
  - s/VMSAv6 - Remove old comments relating to lack of stack. Update BangCam to use GetPTE. Update PPL tables, removing PPLTransL1 (L1 entries are now derived from L2 table on demand) and adding a separate table for shareable pages. Implement the VMSAv6 versions of the Get*PTE and Decode*Entry functions.
Admin:
  Tested on Raspberry Pi 1, Raspberry Pi 3, Iyonix, RPCEmu (ARM6 & ARM7), comparing before and after CAM and page table dumps to check for any unexpected differences


Version 5.55. Tagged as 'Kernel-5_55'

Detail:
  This change gets rid of the following switches from the source (picking appropriate code paths for a 32bit HAL build):
  * FixCallBacks
  * UseProcessTransfer
  * CanLiveOnROMCard
  * BleedinDaveBell
  * NewStyleEcfs
  * DoVdu23_0_12
  * LCDPowerCtrl
  * HostVdu
  * Print
  * EmulatorSupport
  * TubeInfo
  * AddTubeBashers
  * TubeChar, TubeString, TubeDumpNoStack, TubeNewlNoStack macros
  * FIQDebug
  * VCOstartfix
  * AssemblingArthur (n.b. still defined for safety with anything in Hdr: which uses it, but not used explicitly by the kernel)
  * MouseBufferFix
  * LCDInvert
  * LCDSupport
  * DoInitialiseMode
  * Interruptible32bitModes
  * MouseBufferManager
  * StrongARM (new CacheCleanerHack and InterruptDelay switches added to hdr/Options to cover some functionality that StrongARM previously covered)
  * SAcleanflushbroken
  * StrongARM_POST
  * IrqsInClaimRelease
  * CheckProtectionLink
  * GSWorkspaceInKernelBuffers
  * EarlierReentrancyInDAShrink
  * LongCommandLines
  * ECC
  * NoSPSRcorruption
  * RMTidyDoesNowt
  * RogerEXEY
  * StorkPowerSave
  * DebugForcedReset
  * AssembleKEYV
  * AssemblePointerV
  * ProcessorVectors
  * Keyboard_Type
  Assorted old files have also been deleted.
Admin:
  Identical binary to previous revision for IOMD & Raspberry Pi builds


Version 5.51. Tagged as 'Kernel-5_51'

Detail:
  This change gets rid of the following switches from the source (picking appropriate code paths for a 32bit HAL build):
  * HAL
  * HAL26
  * HAL32
  * No26bitCode
  * No32bitCode
  * IncludeTestSrc
  * FixR9CorruptionInExtensionSWI
  Various old files have also been removed (POST code, Arc/STB keyboard drivers, etc.)
Admin:
  Identical binary to previous revision for IOMD & Raspberry Pi builds


Version 5.49. Tagged as 'Kernel-5_49'

Detail:
  s/Kernel - Add macro for CLREX, which uses a dummy STREX on basic ARMv6 machines. Clear the exclusive monitor after issuing transient callbacks, to cope with callbacks being triggered on exit from IRQ
  s/ArthurSWIs, s/HAL, s/NewIRQs - Clear the exclusive monitor on exit from IRQ handlers & default FIQ handler
  s/VMSAv6 - Clear the exclusive monitor on entry to the data abort pre-veneer
Admin:
  Tested on Raspberry Pi
  Non-transient callback handlers, custom abort handlers, FIQ handlers, and anything else which returns directly to interrupted user code is responsible for issuing its own CLREX if the code has done something that could have left the local monitor in the exclusive state (e.g. calling a SWI counts towards this, as there's no guarantee the monitor will be open on exit from the SWI)


Version 5.35, 4.79.2.327. Tagged as 'Kernel-5_35-4_79_2_327'

Detail:
  s/ARMops, s/HAL - Move CPU feature init to after the RAM clear, to prevent the cached values being clobbered on platforms where the HAL doesn't perform the RAM clear
  s/CPUFeatures - Update/clarify comment
Admin:
  Tested on Raspberry Pi
  Fixes issue spotted by Sprow


Version 5.35, 4.79.2.320. Tagged as 'Kernel-5_35-4_79_2_320'

Detail:
  This set of changes tackles two main issues:
  * Before mapping out a cacheable page or making it uncacheable, the OS performs a cache clean+invalidate op. However this leaves a small window where data may be fetched back into the cache, either accidentally (dodgy interrupt handler) or via agressive prefetch (as allowed for by the architecture). This rogue data can then result in coherency issues once the pages are mapped out or made uncacheable a short time later.
    The fix for this is to make the page uncacheable before performing the cache maintenance (although this isn't ideal, as prior to ARMv7 it's implementation defined whether address-based cache maintenance ops affect uncacheable pages or not - and on ARM11 it seems that they don't, so for that CPU we currently force a full cache clean instead)
  * Modern ARMs generally ignore unexpected cache hits, so there's an interrupt hole in the current OS_Memory 0 "make temporarily uncacheable" implementation where the cache is being flushed after the page has been made uncacheable (consider the case of a page that's being used by an interrupt handler, but the page is being made uncacheable so it can also be used by DMA). As well as affecting ARMv7+ devices this was found to affect XScale (and ARM11, although untested for this issue, would have presumably suffered from the "can't clean uncacheable pages" limitation)
    The fix for this is to disable IRQs around the uncache sequence - however FIQs are currently not being dealt with, so there's still a potential issue there.
  File changes:
  - Docs/HAL/ARMop_API, hdr/KernelWS, hdr/OSMisc - Add new Cache_CleanInvalidateRange ARMop
  - s/ARM600, s/VMSAv6 - BangCam updated to make the page uncacheable prior to flushing the cache. Add GetTempUncache macro to help with calculating the page flags required for making pages uncacheable. Fix abort in OS_MMUControl on Raspberry Pi - MCR-based ISB was resetting ZeroPage pointer to 0
  - s/ARMops - Cache_CleanInvalidateRange implementations. PL310 MMU_ChangingEntry/MMU_ChangingEntries refactored to rely on Cache_CleanInvalidateRange_PL310, which should be a more optimal implementation of the cache cleaning code that was previously in MMU_ChangingEntry_PL310.
  - s/ChangeDyn - Rename FastCDA_UpFront to FastCDA_Bulk, since the cache maintenance is no longer performed upfront. CheckCacheabilityR0ByMinusR2 now becomes RemoveCacheabilityR0ByMinusR2. PMP LogOp implementation refactored quite a bit to perform cache/TLB maintenance after making page table changes instead of before. One flaw with this new implementation is that mapping out large areas of cacheable pages will result in multiple full cache cleans while the old implementation would have (generally) only performed one - a two-pass approach over the page list would be needed to solve this.
  - s/GetAll - Change file ordering so GetTempUncache macro is available earlier
  - s/HAL - ROM decompression changed to do full MMU_Changing instead of MMU_ChangingEntries, to make sure earlier cached data is truly gone from the cache. ClearPhysRAM changed to make page uncacheable before flushing cache.
  - s/MemInfo - OS_Memory 0 interrupt hole fix
  - s/AMBControl/memmap - AMB_movepagesout_L2PT now split into cacheable+non-cacheable variants. Sparse map out operation now does two passes through the page list so that they can all be made uncacheable prior to the cache flush + map out.
Admin:
  Tested on StrongARM, XScale, ARM11, Cortex-A7, Cortex-A9, Cortex-A15, Cortex-A53
  Appears to fix the major issues plaguing SATA on IGEPv5


Version 5.35, 4.79.2.306. Tagged as 'Kernel-5_35-4_79_2_306'

Detail:
  s/HAL - Change RISCOS_AccessPhysicalAddress & RISCOS_ReleasePhysicalAddress (aka OS_Memory 14 & 15) to use the MMU_ChangingUncached ARMop instead of TLB_InvalidateEntry, as on ARMv6+ the MMU version ensures the write has been flushed to be visible by the TLB, while the TLB invalidate call doesn't.
  Fix RISCOS_MapInIO (aka OS_Memory 13) not detecting regions which have already been mapped in due to L1_XN flag masking issue. Also issue DSB+ISB after the page table write(s) to ensure it's visible by the TLB hardware.
Admin:
  Tested on IGEPv5


Version 5.35, 4.79.2.305. Tagged as 'Kernel-5_35-4_79_2_305'

Detail:
  * Filled in CPU tables for publicly documented ARMv8 cores (Cortex-A53,57,72).
  * Recent ARM ARMs (e.g. section B1.9.2 of the ARMv7AR ARM) permit the core to
    take an undefined instruction exception upon encountering even not-taken
    conditional undefined instructions. This option is exercised by the
    Cortex-A53, unlike all ARMv7 cores previously supported by RISC OS. This
    unfortunately trips up a lot of kernel code that adapts to different
    architectures at runtime. These have now all been replaced with branches
    over the affected code on the opposite condition.
  * Fixed bug in HAL_InvalidateCache_ARMvF: for the main body of the loop,
    which was written as though to act on the CLIDR register, r8 actually
    contained the CTR register instead.
Admin:
  Tested on Raspberry Pi 3

Version 5.35, 4.79.2.304. Tagged as 'Kernel-5_35-4_79_2_304'

Detail:
  This set of changes adds support for "physical memory pools" (aka PMPs), a new type of dynamic area which allow physical pages to be claimed/allocated without mapping them in to the logical address space. PMPs have full control over which physical pages they use (similar to DAs which request specific physical pages), and also have full control over the logical mapping of their pages (which pages go where, and per-page access/cacheability control).
  Currently the OS makes use of two PMPs: one for the free pool (which now has a logical size of zero - freeing up gigabytes of logical space), and one for the RAM disc (logical size of 1MB, allowing for a physical size limited only by the amount of free memory)
  Implementing these changes has required a number of other changes to be made:
  * The CAM has been expanded from 8 bytes per entry to 16 bytes per entry, in order to allow each RAM page to store information about its PMP association
  * The system heap has been expanded to 32MB in size (from just under 4MB), in order to allow it to be used to store PMP page lists (1 word needed per page, but PMP pages may not always have physical pages assigned to them - so to allow multiple large PMPs to exist we need more than just 1 word per RAM page)
  * The &FA000000-&FBFFFFFF area of fixed kernel workspace has been shuffled around to accomodate the larger CAM, and the system heap is now located just above the RMA.
  * SoftResets code stripped out (unlikely we'll ever want to fix and re-enable it)
  * A couple of FastCDA options are now permanently on
  * Internal page flags shuffled around a bit. PageFlags_Unavailable now publicly exposed so that PMP clients can lock/unlock pages at will.
  * When OS_ChangeDynamicArea is asked to grow or shrink the free pool, it now implicitly converts it into a shrink or grow of application space (which is what would happen anyway). This simplifies the implementation; during a grow, pages (or replacement pages) are always sourced from the free pool, and during a shrink pages are always sent to the free pool.
  File changes:
  - hdr/KernelWS - Extend DANode structure. Describe CAM format. Adjust kernel workspace.
  - hdr/OSRSI6, s/Middle - Add new item to expose the CAM format
  - hdr/Options - Remove SoftResets switch. Add some PMP switches.
  - s/ARM600, s/VMSAv6 - Updated for new CAM format. Note that although the CAM stores PMP information, BangCamUpdate currently doesn't deal with updating that data - it's the caller's responsibility to do so where appropriate.
  - s/ChangeDyn - Lots of changes to implement PMP support, and to cope with the new CAM format.
  - s/HAL - Updated to cope with new CAM format, and lack of logical mapping of free pool.
  - s/MemInfo - Updated to cope with new CAM format. OS_Memory 0 updated to cope with converting PPN to PA for pages which are mapped out. OS_Memory 24 updated to decode the access permissions on a per-page basis for PMPs, and fixed its HWM usage for sparse DAs.
  - s/NewReset - Soft reset code and unused AddCamEntries function removed. Updated to cope with new CAM format, PMP free pool, PMP RAMFS
  - s/AMBControl/allocate - Update comment (RMA hasn't been used for AMBControl nodes for a long time)
  - s/AMBControl/growp, s/AMBControl/memmap, s/AMBControl/shrinkp - Update for new CAM format + PMP free pool
  - s/vdu/vdudriver - Strip out soft reset code.
Admin:
  Tested on Pandaboard
  This is just a first iteration of the PMP feature, with any luck future changes will improve functionality. This means APIs are subject to change as well.


Version 5.35, 4.79.2.284. Tagged as 'Kernel-5_35-4_79_2_284'

Detail:
  s/HAL - A typo seems to have resulted in the HiProcVecs DebuggerSpace page being mapped in as NCB instead of CB. Fix it.
Admin:
  Untested.


Version 5.35, 4.79.2.274. Tagged as 'Kernel-5_35-4_79_2_274'

Improve support for VMSAv6 cache policies & memory types. Expose raw ARMops via OS_MMUControl & cache information via OS_PlatformFeatures.

Detail:
  Docs/HAL/ARMop_API - Document two new ARMops: Cache_Examine and IMB_List
  hdr/KernelWS - Shuffle workspace round a bit to allow space for the two new ARMops. IOSystemType now deleted (has been deprecated and fixed at 0 for some time)
  s/ARM600 - Cosmetic changes to BangCam to make it clearer what's going on. Add OS_MMUControl 2 (get ARMop) implementation.
  s/ARMops - Switch out different ARMop implementations and XCB tables depending on MMU model - helps reduce assembler warnings and make it clearer what code paths are and aren't possible. Add implementations of the two new ARMops. Simplify ARM_Analyse_Fancy by removing some tests which we know will have certain results. Use CCSIDR constants in ARMv7 ARMops instead of magic numbers. Update XCB table comments, and add a new table for VMSAv6
  s/ChangeDyn - Define constant for the new NCB 'idempotent' cache policy (VMSAv6 normal, non-cacheable memory)
  s/HAL - Use CCSIDR constants instead of magic numbers. Extend RISCOS_MapInIO to allow the TEX bits to be specified.
  s/Kernel - OS_PlatformFeatures 33 (read cache information) implementation (actually, just calls through to an ARMop)
  s/MemInfo - Modify VMSAv6 OS_Memory 0 cache/uncache implementation to use the XCB table instead of modifying L2_C directly. This allows the cacheability to be changed without affecting the memory type - important for e.g. unaligned accesses to work correctly. Implement cache policy support for OS_Memory 13.
  s/Middle - Remove IOSystemType from OS_ReadSysInfo 6.
  s/VMSAv6 - Make sure BangCam uses the XCB table for working out the attributes of temp-uncacheable pages instead of manipulating L2_C directly. Add OS_MMUControl 2 implementation.
  s/AMBControl/memmap - Update VMSAv6 page table pokeing to use XCB table
  s/PMF/osinit - Remove IOSystemType reference, and switch out some pre-HAL code that was trying to use IOSystemType.
Admin:
  Tested on Iyonix, ARM11, Cortex-A7, -A8, -A9, -A15
  Note that contrary to the comments in the source the default NCB policy currently maps to VMSAv6 Device memory type (as per previous kernel versions). This is just a temporary measure, and it will be switched over to Normal, non-cacheable once appropriate memory barriers have been added to the affected IO code.


Version 5.35, 4.79.2.273. Tagged as 'Kernel-5_35-4_79_2_273'

Detail:
  s/HAL - The VMSAv6/v7 memory model allows speculative instruction fetches from any memory (including device/strongly-ordered), unless the memory is marked as non-executable. So to prevent interference with read-sensitive devices we must make sure all appropriate IO memory is marked as non-executable.
Admin:
  Tested on IGEPv5
  Fixes data corruption seen when reading from SD card


Version 5.35, 4.79.2.265. Tagged as 'Kernel-5_35-4_79_2_265'

Detail:
  s/ARMops, s/HAL - Errata 814220 states that the Cortex-A7 set/way cache maintenance operations violate the usual operation ordering rules, such that an L2 maintenance operation which is started after an L1 operation may actually complete before it, causing data corruption if the L1 data was to be evicted to the L2 entry. Implement the suggested workaround of performing a DSB when switching cache levels, rather than just at the end of the combined L1+L2 group of operations.
Admin:
  Tested on Raspberry Pi 2


Version 5.35, 4.79.2.257. Tagged as 'Kernel-5_35-4_79_2_257'

Following hot on the heels of revision 1.1.2.39, when there's more than one block in existance the shuffle up loop trashes v3 & v4, which we need in the calculation just below.
Could just use other registers in the shuffle loop, but we only have ip free at that point, so be lazy and just reload & reextract the flags.
Tested on a softload Kinetic, now the RAM speed flags look sensible and the RAM clear doesn't fall off the end.

Version 5.35, 4.79.2.244. Tagged as 'Kernel-5_35-4_79_2_244'

Detail:
  s/HAL - Change ClearPhysRAM to always map in memory as cacheable + bufferable instead of only on StrongARM, as it's an optimisation that can help other platforms as well.
Admin:
  Tested on BB-xM, StrongARM RiscPC


Version 5.35, 4.79.2.231. Tagged as 'Kernel-5_35-4_79_2_231'

When Subtractv1v2fromRAMtable is called to remove a region that results in one of the RAM blocks being split in the middle the resulting size was incorrect.
The shuffle up loop was reusing v6 as an iterator not realising that it's needed to calculate the size of the 2nd half later, the error introduced was the difference between the physical address where PhysRamTable is located and the block being split - these could be a long way apart for example when there are two SDRAM banks.
Even if the PhysRamTable is nearby (eg. 1 SDRAM bank) the result would be some weird sized entries which ultimately mean some dynamic area address space is "leaked".

Fixed by swapping to v7, and for symmetry also adjusted the shuffle down loop to match.

Version 5.35, 4.79.2.224. Tagged as 'Kernel-5_35-4_79_2_224'

Add OS_Memory 24 implementation. Change OS_ValidateAddress to use it. Fix kernel leaving the physical access MB in a messy state. Try and protect against infinite abort loops caused by bad environment handlers.

Detail:
  s/MemInfo - Added an implementation of ROL's OS_Memory 24 call. Unlike the old OS_ValidateAddress call, this call should successfully report the presence of all memory areas known to the kernel. It should also correctly indicate which parts of a sparse DA are mapped in, unlike the old OS_ValidateAddress implementation.
  s/ChangeDyn - Update dynamic area handling to construct a lookup table for mapping logical addresses to dynamic areas; this is used by OS_Memory 24 to quickly locate which DA(s) hit a given region
  s/AMBControl/main - Make sure lazy task swapping is marked as disabled when AMB_LazyMapIn is {FALSE} - required so that OS_Memory 24 will give application space the correct flags
  s/ArthurSWIs - Switch OS_ValidateAddress over to using OS_Memory 24, as per ROL. For compatibility, Service_ValidateAddress is still issued for any areas which the kernel doesn't recognise (currently, OS_Memory 24 doesn't issue any service calls itself)
  s/Convrsions - ADR -> ADRL to keep things happy
  s/HAL - Fix L2PT page allocation and RAM clear to release the physical access region once they're done with it
  s/Kernel - Make the error dispatcher validate the error handler code ptr & error buffer using OS_Memory 24 before attempting to use them. If they look bad, reset to default. Should prevent getting stuck in an infinite abort loop in some situations (e.g. as was the case with ticket 279). The system might not fully recover, but it's better than a hard crash.
  s/Middle - Rework data/prefetch/etc. abort handlers so that DumpyTheRegisters can validate the exception dump area via OS_Memory 24 before anything gets written to it. Should also help to prevent some infinite abort loops. Strip 26bit/pre-HAL code to make things a bit more readable.
  hdr/KernelWS - Update comment
Admin:
  Tested on BB-xM, Raspberry Pi


Version 5.35, 4.79.2.222. Tagged as 'Kernel-5_35-4_79_2_222'

Detail:
  s/ARMops - If extended pages aren't supported, make sure we use a PCBTrans table which doesn't use L2_X, otherwise the AP flags for some of the sub-pages will be corrupted when the PCB flags get merged in. Add some more comments to the PCBTrans tables so it's easier to see what the different columns are.
  s/ARM600 - Fix BangCam to use extended pages if they're supported; otherwise (assuming ARMops has selected the right PCBTrans table) we'll end up corrupting the AP flags again
  s/HAL - Fix ConstructCAMfromPageTables using the wrong register for ZeroPage when looking up MMU_PCBTrans. Correct a few comments.
Admin:
  Tested on Iyonix
  Page table examination now shows that all subpages have the correct (i.e. identical) AP flags. Previously some pages would have incorrect access (e.g. every 4th subpage in some FileCore disc map/dir buffer DAs were writable in user mode)
  ARMops fix will presumably mean extended pages will now work correctly on IOP 80200, as before it would have been using small pages with corrupt AP flags


Version 5.35, 4.79.2.221. Tagged as 'Kernel-5_35-4_79_2_221'