1. 04 Jul, 2015 2 commits
    • Jeffrey Lee's avatar
      Enable high processor vectors/zero page relocation. OS_DynamicArea 20 fixes. · f5644f74
      Jeffrey Lee authored
        Makefile, hdr/Options - By default enable high processor vectors/zero page relocation for compatible machines, but also allow the components file to override the setting if required
        s/ChangeDyn - Fix OS_DynamicArea 20 to check the correct range for doubly mapped areas, and to correctly localise its error message
        Tested on Iyonix
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_268'
    • Robert Sprowson's avatar
      Add extra OS_DynamicArea subreason · c8f15bd6
      Robert Sprowson authored
      Subreason 20 takes a logical address and tells you which area it lies in, including system areas (ie. those returned by OS_Memory 16.
      This allows areas to change type in future without the caller needing to care where the kernel put it.
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_267'
  2. 19 Jun, 2015 1 commit
    • Jeffrey Lee's avatar
      Fix HiProcVecs build. Remove old-style PublicWS definitions. · 1aa4be26
      Jeffrey Lee authored
        s/ARMops - Added extra LTORG to fix HiProcVecs build error for some platforms
        hdr/PublicWS - Remove the old (non-Legacy_) workspace exports, and add a comment explaining how the newer Legacy_ exports should be used.
        HiProcVecs ROMs for various platforms now appear to build OK
        Untested at runtime
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_266'
  3. 14 Jun, 2015 1 commit
    • Jeffrey Lee's avatar
      Ensure IO memory is marked as non-executable · e5bd59a6
      Jeffrey Lee authored
        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.
        Tested on IGEPv5
        Fixes data corruption seen when reading from SD card
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_265'
  4. 19 Apr, 2015 2 commits
  5. 13 Apr, 2015 1 commit
    • Jeffrey Lee's avatar
      Fix aborts on Cortex-A15 when using lazy task swapping · 99b3f14a
      Jeffrey Lee authored
        s/VMSAv6 - After AMB_LazyFixUp has modified the page tables, perform a DSB + ISB to ensure the page table write has completed before we return from the abort handler.
        Tested on IGEPv5
        Fixes aborts seen in desktop, e.g. when !CloseUp is rebuilding its sprite (heavy RAM write activity delaying pagetable write?)
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_262'
  6. 29 Mar, 2015 1 commit
    • Jeffrey Lee's avatar
      Fix OS_Byte 19 getting stuck if it's called just before the screen blanker... · f17af1cf
      Jeffrey Lee authored
      Fix OS_Byte 19 getting stuck if it's called just before the screen blanker activates. Add new OS_ReadSysInfo reason code for determining IIC bus count.
        s/PMF/osbyte - For OS_Byte 19, move screen blanker check (and current GraphicsV driver check) inside the WFI loop so that the loop will terminate if the screen should blank (or current GraphicsV driver should die) before the next VSync occurs. Also move the Portable_Idle call to before the interrupt trigger - the old location meant that if the screen blanked we'd continue on to the Portable_Idle call and end up pointlessly stalling the system
        s/Middle - Add OS_ReadSysInfo 14, as a legitimate way of finding the number of IIC buses present on the system
        Tested on Pandaboard
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_261'
  7. 16 Mar, 2015 1 commit
  8. 12 Mar, 2015 1 commit
    • Robert Sprowson's avatar
      Ensure HAL cache device is initialised · abcadc7c
      Robert Sprowson authored
      The Cache_HALDevice is in the RAM clear skip table, so when there's no controller found and the kernel is doing the RAM clear it's unset, leading to probable aborts when typing *CACHE Om|Off.
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_259'
  9. 26 Feb, 2015 1 commit
    • Jeffrey Lee's avatar
      Add TouchBook to hdr:GPIODevice · 06396b24
      Jeffrey Lee authored
        hdr/GPIODevice - Added the TouchBook as an OMAP3 machine type. We don't actually create a GPIO HAL device for it, but having it defined here is useful for SDIO support in the HAL.
        Tested on TouchBook
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_258'
  10. 16 Feb, 2015 1 commit
  11. 07 Feb, 2015 1 commit
    • Jeffrey Lee's avatar
      Add workaround for Cortex-A7 errata 814220 · 94c601f8
      Jeffrey Lee authored
        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.
        Tested on Raspberry Pi 2
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_257'
  12. 02 Feb, 2015 1 commit
    • Ben Avison's avatar
      Add Raspberry Pi 2 support · d6806495
      Ben Avison authored
        The Raspberry Pi ROM now joins the IOMD ROM in supporting multiple
        architectures, in this case ARMv6 and ARMv7. This has been achieved by
        creating a new machine type specific for Raspberry Pi. The old ARM11ZF
        machine type remains for builds that are ARM11-only.
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_256'
  13. 20 Jan, 2015 1 commit
    • Jeffrey Lee's avatar
      Perform extra TLB maintenance on ARMv6+. Other cache/TLB maintenance tweaks. · aca7f939
      Jeffrey Lee authored
        s/ARMops - Implement Cache_RangeThreshold for PL310 (helps AMBControl to decide what type of TLB maintenance is best). Fix MMU_ChangingEntry_PL310 doing more work than is necessary; was attempting to flush all ways for a given address tag, when really it should have only been flushing all the lines within a page and letting the cache worry about the tags/indices they correspond to.
        s/ChangeDyn, s/VMSAv6, s/AMBControl/memmap - Do extra TLB maintenance following writes to the page tables, as mandated by the ARMv6+ memory order model. Fixes frequent crashes on Cortex-A9 when running with lazy task swapping disabled (and presumably fixes other crashes too)
        s/MemInfo - Fix OS_Memory cache/uncache so that it does cache/TLB maintenance on a per-page basis instead of a global basis. Vastly improves performance when you have a large cache, but may need tweaking again in future to do a global op if large numbers of pages are being modified.
        Tested on Pandaboard
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_255'
  14. 17 Jan, 2015 1 commit
    • Jeffrey Lee's avatar
      Enable/disable HAL cache controller when enabling/disabling ARM caches · 9c55b854
      Jeffrey Lee authored
        s/VMSAv6 - Modify OS_MMUControl to ensure any HAL-based cache is disabled when either the ARM I or D cache is disabled. This emulates the behaviour of an integrated L2 cache controller.
        Tested on Pandaboard
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_254'
  15. 16 Jan, 2015 1 commit
    • Jeffrey Lee's avatar
      Escape some dollars · ece80d58
      Jeffrey Lee authored
        s/NewReset, s/Super1 - Escape some dollars contained in strings to avoid warnings from objasm
        Resulting binary unchanged
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_253'
  16. 11 Jan, 2015 1 commit
    • Jeffrey Lee's avatar
      Add ARMops for PL310 L2 cache controller · 6eb6ee2a
      Jeffrey Lee authored
        Unlike on the Cortex-A8 or Cortex-A15, the L2 cache that's used with the Cortex-A9 isn't hooked up to the standard ARMv7 CP15 cache maintenance ops. Instead, memory-mapped registers must be used to program and maintain the cache.
        Since the PL310 can't be detected automatically, this change adds support for a 'cache controller' HAL device which the HAL can use to advertise the presence of any external caches. If a cache device is registered during HAL_InitDevices the kernel will then check it against a list of known cache types and replace the appropriate ARMop routines with the alternatives for that controller.
        File changes:
        - hdr/PL310 - New header containing PL310 register listing
        - Makefile - Add export for PL310 header. Reorder exports to be alphabetical
        - hdr/HALDevice - Add cache controller device type, PL310 device
        - hdr/KernelWS - Allocate some workspace for storing a pointer to the current cache HAL device
        - s/ARMops - Add code for searching for known cache types, and implementation of PL310-specific ARMops
        - s/GetAll - Get Hdr:PL310
        - s/NewReset - Look for a cache controller after calling HAL_InitDevices
        Tested on Pandaboard
        Fixes various assorted instability issues
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_252'
  17. 09 Jan, 2015 1 commit
    • Jeffrey Lee's avatar
      Fix detection of ARMv7 minimum cache line lengths · 4843ce7e
      Jeffrey Lee authored
        s/ARMops - Replace the code to calculate the minimum cache line lengths with something much simpler which reads the values directly from the cache type register.
        The old code was buggy in two ways:
        (a) the cache size identification register stores the line length as log2(num words)-2, whereas the code throughout the kernel was expecting it to be log2(num bytes)-2
        (b) the loop is structured so that it will try and read the details of a non-existent cache level. although it doesn't read anything from CP15, it does result in the minimum cache line length values getting clobbered
        The net result of the above two bugs being that the OS would treat the CPU as if the minimum line length was just 4 bytes (although other than slowing down cache maintenance ops, this shouldn't have had any bad side-effects)
        The cache type register directly contains the minimum line lengths as log2(num bytes)-2, so by switching over to use that everything is now fine.
        Tested on BB-xM, Pandaboard
        Fixes issue spotted by Willi Theiss
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_251'
  18. 21 Dec, 2014 2 commits
    • Robert Sprowson's avatar
      Delegate L2 (and below) cache init at power on/reset to the HAL · 16c00596
      Robert Sprowson authored
      Historically the kernel looked after all aspects of cache control since they were common across all ARMs. However, not all cache controllers are created equal, and sometimes more complex initialisation steps are needed than fit the generic coprocessor ops - for example the PL310 attached to a Cortex-A9 has memory mapped control registers.
      Rather than clutter the kernel with one shot init code for every cache controller invented, we delegate that step to the HAL in HAL_Init. This is only a few hundred instructions later than where it was already being set. The kernel remains responsible for subsequent maintenance, this is just init which is being handed off.
      A quick survey of the Cortex-A TRMs shows:
      A5 - optional, for example ARM's PL310, ref TRM section 8.1.7.
      A7 - optional, C bit of SCTLR, ref TRM section 1.1.
      A8 - L2EN bit of ACTLR, note this bit has been recycled for other uses on other cores, ref TRM section 8.3.
      A9 - not integrated, ARM's PL310 uses bit 0 of control register 1, ref PL310 TRM section 3.1.1.
      A12 - see A17
      A15 - integrated, C bit of SCTLR, ref TRM section 7.2.3.
      A17 - integrated, bit 18 of L2CTLR & C bit of SCTLR, ref TRM section 7.2.
      and while we've got the TRMs open, back fill the CPU id register table.
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_250'
    • Robert Sprowson's avatar
      Use symbol for GraphicsV op · 13123813
      Robert Sprowson authored
  19. 06 Dec, 2014 2 commits
    • Robert Sprowson's avatar
      Untangle some buses · aaddca77
      Robert Sprowson authored
       More clearly describe the intent of the different bus types.
       Reclassify the OMAP interconnect as either a system bus or peripheral bus, noting their names from the datasheet.
       Reclassify the GPMC as an expansion bus.
       Add AMBA 3's AXI bus, which (along with the existing AHB/APB) is what iMx6 uses. Delete redundant iMx6 interconnect type.
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_249'
    • Robert Sprowson's avatar
      Fixes to HAL memory info calls and docs · 75de3830
      Robert Sprowson authored
       Several places in this code called the HAL or other ATPCS defined functions like memset() and hoped that the overall result was V clear. If any of them accidentally set V (for example a CMP that straddles 0x80000000) you ended up trying to look up an international error at the address of the reason code to OS_Memory.
       Now, explicitly clear V in the non error cases where an ATPCS function was called.
       Change the HAL_PhysInfo call to expect a physical ROM size back as an inclusive range, to match the RAM range subreason code. Add 1 to correct for this. A value of 0 & 0 is taken to mean "no physical ROM" as before.
       Document that 255 means "no IOMD" or "no VIDC", that's what the HALs have been using since year dot.
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_248'
  20. 13 Nov, 2014 1 commit
    • Robert Sprowson's avatar
      Add a means to write NetStnCMOS in a HAL world · 10a86092
      Robert Sprowson authored
      With ProtectStationID turned on there are no routes to writing the Econet station (or bottom octect of the IP address), a function previously fulfilled by the SetStation utility which pokes the hardware directly and doesn't fit into a HAL model.
      Add a new subreason to OS_NVMemory to perform this role. This SWI appeared for RISC OS 5.00, and errors unsupported subreasons, so there's a means of run tim selecting its use by checking the platform class and trying the SWI. All RISC OS 5 based platforms can always be upgraded to this version, since they're all still being maintained.
      hdr/Options: move the switch with the other options from osinit.s
      i2cutils.c: new subreason
      Ditch the 'ObsoleteNC1CMOS' switch, if it was obsolete for NC1, it's certainly obsolete now.
      Ditch unmaintained messages files for Morris4/Omega/Ursula projects.
      Tested on a Risc PC.
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_247'
  21. 07 Nov, 2014 1 commit
  22. 04 Nov, 2014 1 commit
    • Jeffrey Lee's avatar
      Make OS_GSTrans be more sensible about what it treats as system variables · 2e79f8be
      Jeffrey Lee authored
        s/Arthur2 - OS_GSTrans now uses the same rules as OS_SetVarValue when deciding whether text within angle brackets is a valid system variable name or not. In particular spaces in the middle of a variable name are no longer considered valid, so expressions such as "*If 0<1 AND 1>0 then echo true" now have the expected result
        Also replaced the magic constant used for the name buffer length with a proper symbolic value, and tweaked its handling a bit in order to increase the maximum permissible variable name length from 253 chars to 255 (although OS_SetVarVal allows longer)
        Tested on Iyonix
        Fixes issue reported on forums:
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_245'
  23. 27 Oct, 2014 1 commit
    • Robert Sprowson's avatar
      Another fix to split_block · bc954604
      Robert Sprowson authored
      Following hot on the heels of revision, 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, Tagged as 'Kernel-5_35-4_79_2_244'
  24. 25 Oct, 2014 1 commit
  25. 22 Oct, 2014 1 commit
  26. 18 Oct, 2014 1 commit
    • Jeffrey Lee's avatar
      Update HAL device ID allocations · c95a9fff
      Jeffrey Lee authored
        hdr/HALDevice - Added some extra HAL device IDs
        Tested by building BCM2835 ROM
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_241'
  27. 15 Oct, 2014 1 commit
    • John Ballance's avatar
      Re enable compile with kernel built from current tree. · 0564d81c
      John Ballance authored
      Recent kernal changes appear to have enabled lazy task swapping, which brought
      up a data alignment abort whilst compiling the source tree using a rom compiled
      from this tree. Simple change added
      to AMB_MakeHonestLA and PN routines to avoid this.
        (highlight level of testing that has taken place)
        (bugfix number if appropriate)
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_240'
  28. 08 Oct, 2014 1 commit
    • Robert Sprowson's avatar
      Header defs for Pi Compute module and IIC · fe5436ee
      Robert Sprowson authored
      OSEntries.h: added a C structure for RISCOS_IICOpV.
      GPIODevice.hdr: allocate Pandora and Pi Compute module sub device nos.
      HALDevice.hdr/Options.hdr: tabs expanded, capitalised abbreviations.
      Not tagged.
  29. 01 Oct, 2014 1 commit
    • Robert Sprowson's avatar
      Fix for spurious IIC access when probing · 02832075
      Robert Sprowson authored
      When HAL_NVMemoryType reports NVMemoryFlag_MaybeIIC the kernel tries to probe a number of common/known addresses on startup, however the result of the probe is stored around line 1346 without a value value for zero page in R2.
      This is sufficiently early on that the default data abort handler (from when probing the ARM's abort model) is still in place so the stores are silently skipped.
      Due to the RAM clear the NVRamBase (and size) are 0, which later on in ValChecksum result in a zero length IIC probe to address &01.
      Now, R2 is initialised.
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_239'
  30. 26 Sep, 2014 1 commit
  31. 18 Sep, 2014 1 commit
  32. 16 Sep, 2014 1 commit
  33. 15 Sep, 2014 1 commit
    • Jeffrey Lee's avatar
      Fix IIC bus information being wiped by RAM clear · 13e1fc5c
      Jeffrey Lee authored
        hdr/KernelWS - Enlarge the SkippedTables area to encompass IICBus_Base
        s/PMF/IIC - Manually set IICBus_Status of each bus to 0 within IICInit
        Bug was introduced in Kernel-5_35-4_79_2_168 when IIC initialisation was moved to earlier in the ROM init sequence, but has gone unnoticed due to it only really affecting the high-level API (and none of the relevant HALs were relying on the kernel for the RAM clear)
        Tested on BB-xM with kernel RAM clear
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_236'
  34. 29 Aug, 2014 1 commit
    • Jeffrey Lee's avatar
      Add more HAL USB definitions to hdr/HALEntries. Add debug option for calling... · 371f701c
      Jeffrey Lee authored
      Add more HAL USB definitions to hdr/HALEntries. Add debug option for calling HangWatch_Dump on serious errors.
        hdr/HALEntries - Added definition of the struct returned by HAL_USBControllerInfo
        hdr/Options, s/Middle - Added the option to call HangWatch_Dump on serious errors
        Tested on BB-xM
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_235'
  35. 17 Aug, 2014 1 commit
    • John Ballance's avatar
      Added hdr.options line to set up max RAMFS size. · 66c83bb2
      John Ballance authored
        s.NewReset line 1304 amended to use this variable. It was
        hard codede to 128MB as a compromise between address space reservation and size.
        Tested OK to expand to 511MB in task manager. At 512MB the taskmanager window fails.
      Version 5.35, Tagged as 'Kernel-5_35-4_79_2_234'
  36. 31 Jul, 2014 1 commit