Magazyn Exec 5/CD_Magazyn_EXEC_nr_5.iso/Programy/MorphOS/morphosuser-150201.lha/notes/cybppc.notes

Text File | 2001-02-14 | 11.8 KB | 406 lines

Problems:
  o Enter the BootMenu GUI, then leave it...SysIHack won`t
    start because it checks if the gadtools.library is already
    used. The gadtools.library can`t be expunged so there`s
    no cure for this problem.
  o ProbeSCSI
    Sends a Read Defect Data(Length=0x2000) with a sc_length=$fc
    so the unit wants to transfer data but the device doesn`t know
    where to store these.
  o ACDPlay sends READ SUB-CHANNEL 0x42 SCSI command with the
    sc_CmdLength=6 but the SCSI-2 specs demand 10 Bytes for the cmd.
    The consequence is an error requester.(Friedhelm Bunk)
  o YACDP sends the following command
    0x28
    0x00
    0x00000017
    0x00
    0x0001
    0x00
    to a Matsushita with an inserted Audio CD.
    The result is that the CDRom sends the 2048 bytes and
    then 12 * 2352 audio blocks.
  o PlayCD (IDEFIX package)
    Sends a 0x47 command with 0xC cmdlength instead of 0xa
    ->Phase Error
  o Exabyte seems to have a reselection problem..for long reads(compares)
    the unit just hangs. This means the unit itself doesn`t come back from
    a Reselection and there`s nothing i can do about it. This happens inside
    of an IO job and not before any data is tranfered.

Open:
  o >8x Audio cdda and the toomuchdata error


ReleaseNotes:

- V45.10
 o Fixed a potential task startup problem

- V45.9
 o reactivated the ramwait state settings

- V45.8
 o failed OpenDevice now also clears io_Unit,io_Device
 o CloseDevice now does nothing if no io_Unit Ptr is there

- V45.7
 o fixed a taglist problem

- V45.6
 o Init Errorcheck improved

- V45.5
 o Forgot a newlist tc_MemEntry

- V45.4
 o Internal reworks

- V45.3
 o remove Exec's list function usage to inline macros

- V45.2
 o Fixed the inc partition index postfix bug
 o wrapped copymem on memcpy

- V45.1
 o Added CyberStorm SCSI HW check

- V44.67
 o Mount Gadget for each Unit so you can force a Mount even
   for DVD Rams which say they are a CDROM.

- V44.66
 o CVPPC support..
 o Removable is now start by a seperate resident
 o new bios keys..check the bios docs

- V44.65
 o "switch" openscreen mode is now saved too
 o Some text changes in the synchron/offset area to
   reflect more the technical side of SCSI:-)

- V44.64
 o resolution switch now does a fresh restart to avoid some trouble.
 o auto mode more intelligent now...auto was ignored for non RDB
   units...now it assumes YES if the unit has no RDB but if it has
   a RDB it checks for the Reselection/Sync flag in the RDB as usual.
   Synchron/Reselection isn`t used if the unit itself marks itself
   as not supporting them.

- V44.63
 o Oh well..some people really want 15khz support so here they get it in
   interlace(press SPACE)
 o Little SCSIHardware check problem fixed

- V44.62
 o Now the removable check also cares for SizeBlock and SectorPerBlock to
   detect medium changes where the layout is the same but not these 2 fields
   for some reason.
 o ALICE_LHS is now cleared(VGAOnly)...the monitor itself doesn`t need it
   but the OS itself needs it to create a real 31khz screen

- V44.61
 o SCSI/CDFS Menus off, if no SCSI hardware is there

- V44.60
 o 31khz monitor
 o 31khz bootmenu config changes

- V44.59
 o screen space optimization

- V44.58
 o internal bootmenu changes
 o PatchNCRSCSI

- V44.57
 o "right shift" is now "a"

- V44.56
 o Map sw restructured..if that helps..only god knows.
   Black Magic

- V44.55
 o Boot BIOS "right shift" is now "left shift" because of Picasso
   conflict.

- V44.54
 o Bootmenu GUI Problem with the Unit.Synchron Gadget fixed
 o patchmac`s scan now aligns the search ptr to avoid
   the V39.106 odd rt_Name start ptr problem.
   Then it hasn`t found the patch pattern

- V44.53
 o Reselection fix
   (could be reproduced with a 11 track Quake2,Philips 2600
    and the command
    "ED 00 00 00 00 00 02 00 DC 00" to write the Track TOC
    with Burnit)

- V44.52
 o ignore not legal scsi msgs after a reselection.
   Exabyte likes to send SAVE DATA msgs right after
   a reselection.

- V44.51
 o Some maprom tweaks..maybe that removes
   the problems some people have.

- V44.50
 o Forgotten freegadgets in the Bootmenu
 o Added CDFS Cache(only useful for the P5 cdfs) support
 o fixed a forgotten ram config problem

- V44.49
 o CDRDBScan bootmenu gadget
 o Fix to the cd devicename update

- V44.48
 o changed CDP5 to CDFS

- V44.47
 o Added CDrive filesystem support
 o Rework GUI again and added CDRom Mount/Boot support
 o initialize the devicename into a fse Node

- V44.46
 o Bootmenu tweaks

- V44.45
 o Blizzard Board detection was broken
 o MaxUnit in ControllerID wasn`t correct for the Blizzard

- V44.44
 o Bootmenu keycheck code workaround for key synclost situation

- V44.43
 o new Bootmenu memory entries

- V44.42
 o more intelligent unit scan to find slow startup drives

- V44.41
 o Disable Unit,boot,rescan,scsiconfig fails with a unit which
   can change it`s sector size(for example a MO)..fixed
 o Reworked the way Inquiry works...maybe this also fixes 0 totalsector problems
 o Made Bootmenu 040 aware
 o If the loaded filesystem had more than 1 hunk the bptr in the seglist was
   no bptr..caused an enforcer hit with version filesystem:

- V44.40
 o More than 1 Filesystem problem fixed
 o Unit Dostype cache to avoid unnecessary IO

- V44.39
 o BootMenu Init reworked

- V44.38
 o LED reworked..now controlled by the script..luckily that works too now
   and this hopefully fixes the buserror nightmare.
 o For Fusion owners increased the chipram border in PatchMac to 0x4000.

- V44.37
 o Buserror fog adventure...
 o buserror doesn`t happen with the 68040..
 o buserror doesn`t happen with all 68060 boards..(not with mine anymore)
 o buserrors don`t come as often when i added some more NOPs..
   from 10min to 60min frequency for Klaus
 o buserror reason found by logicanalyser.
   The write value in DRIVER_SetLed and DRIVER_ClearLed is equal to the
   weird 4th byte in the DNAD(internal scripts dma pointer) which we
   thought caused the buserror. The chip basicly run amok and put this
   led value into some internal register it used as its own dma pointer.
 o LED COMPLETELY DISABLED..

- V44.36
 o Little debug output changes
 o CMD Buffer length now 32 instead of 12 which already collides with new 16byte cmds
 o switch off HandShake Timeout(which is already on max) after the initial scan
   for el lamo devices which take that long to complete its instruction.
   The Handshake timeout was used to detect a bus hang and try to recover...
 o "Save" -> "Use"

- V44.35
 o Lun Number gadget update problem

- V44.34
 o Maxoffset 15(fixes Atlas2 problem)
   Shit manual.
 o Fixed the listview display problem(instead of Unit x only Unit 0)
   Disable/Enable setting were shown after a fresh reboot.

- V44.33
 o 68040/68060 disable check for some bootmenu fields
 o busid prefs activated
 o maprom prefs activated
 o fixed a missing FreeScreenDrawInfo

- V44.32
 o Flash Prefs write working now

- V44.31
 o Added BootMenu
 o Thrown out the old prefs system(have to disable it in Unitcontrol too)
 o fixed some little max sync period value problem

- V44.30
 o Playing around with the fifos...maybe that`s the cause for the buserror.
   (Walking through the fog sucks)
 o Check for termination on the wide bus side..if that fails..well..completely
   disable reselection

- V44.29
 o more intelligent dostype comparison in the CheckDeviceNode routine
   which tests if a current partition`s geometry is equal to a one
   already in the system so that filesystem can be used.
   The old check simply compared the whole DOSTYPEID.
   The new one checks if it`s the same and if not it checks if it`s
   almost the same where it ignores the last byte of the dostype
   which normally represents a filesystem type number a filesystem
   supports.(FFS supports 5 different versions for example)
 o Fixed a removable problem with different disks that have
   the same geometry and same DeviceName. Now it works like it should.
   (MrDaniel)
 o I wasn`t able to reproduce the Buserror anymore..

- V44.28
 o corrected the readtoomuch length output bug i introduced lately with
   some script changes
 o fixed the cybermap problem
 o max sync offset set to 16(manual is a bit unsecure about this)
   (The symbios web page says 16)

- V44.27
 o Activated script ram usage now. Maybe that removes
   these weird problems.(It didn`t...sigh)

- V44.26
 o Hopefully this awful buserror problem fixed.
   I couldn`t reproduce it with a heavy diskspeed test anymore
   but that doesn`t mean much.(well..actually 2 days afterwards
   it came back)

- V44.25
 o IOERR_OPENFAIL missed in some lowmemory error case
 o Fixed HDToolBox scan problem...oh well.

- V44.24
 o Small changes in the sync bookkeeping
 o Fixed Reselection hang problem

- V44.23
 o Didn`t init the internal sync,wide parameters after a SCSIReset,
   so unitcontrol still showed the old though it was already in async.

- V44.22
 o Corrected DCNTL_COM init

- V44.21
 o My little LED Devil...

- V44.20
 o WORM RDB load thrown out.
 o Phase Mismatch -> SCSI Reset now
   Force disconnect and Reset MSG is too riscy

- V44.19
 o Fixed some LED problem...sigh..we should really have done
   some simple BUS BUSY Led system.

- V44.18
 o Compiled with OPT

- V44.17
 o Handshake xx second enforcer hit interval with no termination.
   Fixed.

- V44.16
 o Reselection crash/actual problem fixed

- V44.15
 o Used Handshake Timeout to escape deadlocks
 o After a new Exabyte 8500 Firmware update, the Inquiry Data Phase
   deadlock vanished...strange.
   It worked with the old firmware at the Mk2 and 4091
 o Fixed SCSI timing problems(UMAX Scanner works now and probably
   also the Exabyte without the update)..strange days
 o default reselection off(My little friend the Philips CDD2600)

- V44.14
 o Changed some requester msgs

- V44.13
 o Changed something in the Pool system to adapt to changed struct sizes
 o Enhanced Error output(more informations)

- V44.12
 o Fixed Autosense problem..uses SCSIF_OLDAUTOSENSE and not the bit
   which are different, so sense was always 4 bytes long.
   (MakeCD problem)

- V44.11
 o Fixed some beta release no reselection set bug
 o Reworked the sense system
 o Fixed 2 mungwall hits which fixes the AFS problem

- V44.10
 o Beta Release

- V44.9
 o Improved No Termination Check
 o Optimized the removable notify
 o Fixed some removable enforcer hit

- V44.8
 o Forgot the START Unit cmd...should fix some coldboot startup problems
   (Seagate 9G)
 o Removable task problem with notify..
 o Lun problem with Identify msg

- V44.7
 o the automatic wide init was disactivated..changed that

- V44.6
 o More error check requesters+recovering
 o Reactivated UnitArray check for new Unitcontrol
 o Reactivated BusReset(it`s more dangerous now)
 o Activated WideResidue

- V44.5
 o Added a cludge to fix probescsi`s bug where it does this:
   Read Defect Data(Length=0x2000) with a sc_length=$fc so the unit wants to
   transfer data but the device doesn`t know where to store these.
   Oh well...this isn`t worth it.
   This cludge will break when a disconnect occurs...
 o More sophisticated bug requester output
 o smakefile problem which caused that RDB_Removable wasn`t up to date always which
   may caused enforcer hits.

- V44.4
 o Instead of a global termination check i added the check to OpenDevice which fails
   with HFERR_Phase if the termination isn`t correct. I should create a HFERR_Termination.

- V44.3
 o The removable task isn`t woken up when you rescan and a new unit is found.
   Has only an effect if you don`t have a recognized removable drive there.

- V44.2
 o Added IO_GetBusID
 o Added IO_GetMaxUnit
 o Added IO_GetMaxLun
 o Changed the IO_Prefs functions to care for more than 8 units
 o Fixed filesystem hunk loader`s more than 1 hunk bug...
   has no effect on FFS because it only has 1 hunk

- V44.1

 o reactivated Synchron,Wide and Reselection
 o used the optimizer(but disabled the modes which cause the code to crash)
 o fixed the hang problem(really weird thing)

- V44.0

 o Initial Beta release