GitLab

Version 1.62. Tagged as 'Podule-1_62'

Previous changes were untagged, but incrementing the version number gives
us a tag we can associate with the official RISC OS 5.16 release.

Version 1.61. Tagged as 'Podule-1_61'

Detail:
  Uses suffixed file extensions for compatiblity with both objasm and asasm.
Admin:
  Supplied by Peter Naulls, tested at ROOL

Version 1.60. Not tagged

Prised out the last remaining IOMD assumptions,now follows a policy of autodetection described in "Docs.Policy",which included eliminating some static tables which were produced at runtime.

The static tables are now all in s.Tables.
The podule node list is generated at run time since so many of the node
entries needed correction depending on where "SSpaceStart" is.
Though not explicitly mentioned in the PRMs,where an IO area doesn't
exist for a given IOMD varient zero is returned by Podule_ReadInfo to
denote it does not exist
Support for Podule_SetSpeed helpers added,thus IOMD abstracting the way
you set EASI cycle speeds.

Version 1.60. Tagged as 'Podule-1_60'

Network and ExtROM internal loaders now return full message strings
rather than the "Bad" token!
Presence of absence of ExtROMs/EASI/NIC/IOC and MEMC podule support is
deduced at runtime,and the logical addresses requested from the Kernel
by OS_Memory.EASI,ExtROMs and NIC remain assembly time options though
just to reduce the size of the module a bit.
An assortment of internal variables have been altered to have 0 or -1 as
their unused value rather than &80000000 as this may be a candidate for
genuine loader addresses in future.
First pass at removing IOMD specific assumptions in the module.

Version 1.59. Tagged as 'Podule-1_59'

Admin:
  Requires Library 0.72 or later.
  Requires BuildSys 3.09 or later.
  Requires Env 0.65 or later.

Version 1.58. Tagged as 'Podule-1_58'

Version 1.57. Tagged as 'Podule-1_57'

Detail:
  Modified to deal with situation in which bus 'remembers' value briefly
  Now clear bit one of Line Control Register - this bit is always pulled high
  by ethernet version
Admin:
  HdrSrc-0_97 needed for auto-detect

Version 1.56. Tagged as 'Podule-1_56'

Can now include code that auto-detects the version of the Customer F 8 board in use i.e Hardware modem or Ethernet

Detail:
  New value of FakePodule should be passed in to build the module with this code
  FakePodule = ProdType_AutoDetectF8
  Auto-Detect code writes to the Line Control Register on the modem
  (address &8800000c) and then tries to read the same data back from
  address &8800002c (the modem is only connected to three address lines so
  register map repeats every 8 words).
  This should have no effect on ethernet controller because it only starts at
  &88400000.
Admin:
  Tested on Customer F 8 modem and ethernet boards appears to work.
  Change needed for next version - since the ethernet version doesn't load the
  bus at this address the capacitance of the bus is enough to remember the valu
  making it apppear that a modem is present. Currently two reads from the same
  location are carried out and the result of the first is discarded. This seems
  to be enough to clear the bus but we should actually force a read from
  another location on the Bd bus.

Version 1.55. Tagged as 'Podule-1_55'

* Suppresses non-32-bit expansion card loaders (a 32-bit compatible loader
  must have the magic word "32OK" after the 4 entry points).
* Lots of OS_SynchroniseCodeAreas calls added. This module was far from
  StrongARM-compatible - I'm amazed it worked at all; indeed it often failed
  when softloaded.

Needs a new Hdr:NewErrors with the "Loader is not 32-bit compatible" error.

Version 1.54. Tagged as 'Podule-1_54'

Detail:
  No longer reports "Assembling Podule Manager with Fake Podule header, type 00000000"
  when it is not faking a podule.

Admin:
  No code change. Module image unchanged.

Version 1.53. Tagged as 'Podule-1_53'

Detail:
  The Customer F NC needs to pass Options from the Components file to
  ObjAsm via the Makefile. The previous version didn't do this
  correctly.
Admin:
  Tested OK.

Version 1.52. Tagged as 'Podule-1_52'

Detail:
  The Customer F 5 Ethernet NC doesn't have a podule ROM fitted, so we
  provide a fake podule header in podule manager.
  This is used if FakePodule0 is set in the machine header file
  or if FakePodule is passed in from the Makefile (FakePodule overrides
  the FakePodule0).
  It should be set to ProdType_EtherI to fake an EtherI podule but
  other values are possible, eg for MPEG).
Admin:
  Tested OK in Customer F 5 NC.

Version 1.51. Tagged as 'Podule-1_51'

Added support for ReassignedIOMDInterrupts - Podule_ReadInfo now tells expansion card 0 to use device number 11.

Added a SynchroniseCodeAreas call between reading the loader and entering it.

Version 1.50. Tagged as 'Podule-1_50'

Changed to use srccommit.

Version 1.49. Tagged as 'Podule-1_49'