PCI IDE Adapter for PPC/Linux
James F Dougherty
jfd at GigabitNetworks.COM
Thu Aug 23 09:30:51 EST 2001
Ok, disk works perfectly now. The problem was that my mapirq() was
mapping my PCI slot interrupt to IRQ1. When I tested the network card,
I assumed that IRQ 3 was in use, it turns out it was sharing IRQ1!
So, when I fixed mapirq(), I started getting perpetual interrupts,
and then I looked at IRQ routing on the Mousse board, and the IRQ's
are run through a PLD for IRQ0-4. The PLD had HDL code which tied the
watchdog irq from M48T59Y into IRQ3. I cleared the FT bit in the M48t59,
however, it kept interrupting. Here is the relevant HDL:
encode_ints: process (com1,pci_slot_int,inta,cpci_intb,enet_int)
begin
irq(4) <= not com1; -- and (not com3);
if reg_data2(0) = '1' then
irq(3) <= tod_wd_irq and tod_wd_rst and pci_slot_int ;
else irq(3) <= tod_wd_irq and pci_slot_int;
end if;
irq(2) <= inta and ( not cpci_intb ); -- cpci_intb is an inverted input
irq(1) <= enet_int;
end process encode_ints;
So, I cannot disable the watchdog IRQ on M48T59.... maybe the pin assignment
in the HDL code is off? I changed the HDL to read
encode_ints: process (com1,pci_slot_int,inta,cpci_intb,enet_int)
begin
irq(4) <= not com1; -- and (not com3);
irq(3) <= pci_slot_int ;
irq(2) <= inta and ( not cpci_intb ); -- cpci_intb is an inverted input
irq(1) <= enet_int;
end process encode_ints;
and reprogrammed the ISP 100LT PAL.
Needless to say, the Promise card worked the first time.
This does not explain the other broken Promise card, afaik it
suffered ESD damage.
Anyhow, Promise works great. I am even enabling the UDMA mode.
Thanks for all of the ideas I got from everyone on debugging this.
Any thoughts on the M48T59 WD IRQ would be greatly appreciated.
Thanks,
-James
>
>
> Ok, so I took the Promise card out of the test system today, and put
> it into my Windows box. I tried to install the drivers, and use a drive,
> but Windows crashed! I took the card back to Frye's and got a
> new one... it immediately detected the drive, but now, it hangs ..
>
> PDC20262: IDE controller on PCI bus 00 dev 70
> PDC20262: chipset revision 1
> PDC20262: not 100% native mode: will probe irqs later
> PDC20262: ROM enabled at 0x000d0000
> PDC20262: (U)DMA Burst Bit DISABLED Primary PCI Mode Secondary PCI Mode.
> PDC20262: FORCING BURST BIT 0x00 -> 0x01 ACTIVE
> ide0: BM-DMA at 0xbfff00-0xbfff07, BIOS settings: hda:pio, hdb:pio
> ide1: BM-DMA at 0xbfff08-0xbfff0f, BIOS settings: hdc:pio, hdd:pio
> hda: probing with STATUS(0x50) instead of ALTSTATUS(0xff)
> hda: WDC WD300AB-00BVA0, ATA DISK drive
> hdb: probing with STATUS(0x00) instead of ALTSTATUS(0xff)
> hdb: probing with STATUS(0x01) instead of ALTSTATUS(0xff)
>
> Still, this is excellent progress! I cannot believe the crap you buy at
> Frye's, even when there aren't white return stickers on them!
>
> Anyhow, I am getting interrupts, I guess the thing todo at this point
> is check the PCI bus for I/O & DMA on an analyzer, I cannot see why the
> system would hang at this point unless there was a H/W problem with
> this board. I will have to look into it... there may not be the right
> resistors stuff on this board or something with the bus loading..
>
> Total memory = 64MB; using 0kB for hash table (at 00000000)
> Linux version 2.4.2_hhl20 (jfd at atlantis) (gcc version 2.95.2 19991024
(release))
> #474 Tue Aug 14 00:09:19 PDT 2001
> cpu0: MPC8240/KAHLUA : MOUSSE Platform : 64MB RAM: MPLD Rev. 1
> Sandpoint port (C) 2000, 2001 MontaVista Software, Inc. (source at mvista.com)
> IP PNP: 802.3 Ethernet Address=<0:0:10:20:30:44>
> IP PNP: mousse: eth0 IP 209.128.93.138/255.255.255.0 gateway 209.128.93.129
> server 209.128.93.133
> On node 0 totalpages: 16384
> zone(0): 16384 pages.
> zone(1): 0 pages.
> zone(2): 0 pages.
> time_init: decrementer frequency = 16.665920 MHz
> time_init: MPC8240 PCI Bus frequency = 33.331840 MHz
> Calibrating delay loop... 133.12 BogoMIPS
> Memory: 62452k available (1324k kernel code, 496k data, 88k init, 0k highmem)
> Dentry-cache hash table entries: 8192 (order: 4, 65536 bytes)
> Buffer-cache hash table entries: 4096 (order: 2, 16384 bytes)
> Page-cache hash table entries: 16384 (order: 4, 65536 bytes)
> Inode-cache hash table entries: 4096 (order: 3, 32768 bytes)
> POSIX conformance testing by UNIFIX
> PCI: Probing PCI hardware
> Linux NET4.0 for Linux 2.4
> Based upon Swansea University Computer Society NET3.039
> Initializing RT netlink socket
> Starting kswapd v1.8
> pty: 256 Unix98 ptys configured
> block: queued sectors max/low 41405kB/13801kB, 128 slots per queue
> RAMDISK driver initialized: 16 RAM disks of 131072K size 1024 blocksize
> loop: loaded (max 8 devices)
> Serial driver version 5.02 (2000-08-09) with MANY_PORTS SHARE_IRQ SERIAL_PCI
> enabled
> ttyS00 at 0xffe08080 (irq = 4) is a ST16650
> Linux Tulip driver version 0.9.13a (January 20, 2001)
> eth0: Digital DS21143 Tulip rev 65 at 0xbfff80, EEPROM not present,
00:00:10:20
> :30:44, IRQ 1.
> eth0: MII transceiver #0 config 3000 status 7829 advertising 01e1.
> NET4: Linux TCP/IP 1.0 for NET4.0
> IP Protocols: ICMP, UDP, TCP
> IP: routing cache hash table of 512 buckets, 4Kbytes
> TCP: Hash tables configured (established 4096 bind 4096)
> NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
> devfs: v0.102 (20000622) Richard Gooch (rgooch at atnf.csiro.au)
> devfs: boot_options: 0x0
> VFS: Mounted root (nfs filesystem).
> Mounted devfs on /dev
> Freeing unused kernel memory: 88k init 4k openfirmware
> INIT: version 2.78 booting
> INIT: Entering runlevel: 2
> eth0: Setting full-duplex based on MII#0 link partner capability of 45e1.
>
>
> mousse login: root
> Password:
> PAM_unix[13]: (login) session opened for user root by LOGIN(uid=0)
> Last login: Mon Aug 13 21:18:12 2001 on console
>
> mousse-(1):[11:35pm]: [/root>mount -a
> mousse-(2):[11:35pm]: [/root>cat /proc/interrupts
> CPU0
> 1: 4571 MPC8240 EPIC Level eth0
> 4: 595 MPC8240 EPIC Level serial
> BAD: 0
> mousse-(3):[11:35pm]: [/root>
>
>
>
> >
> > The fact that it works a walnut but not your board points to board
> > specific problem. Have you double checked IRQ routing?
> >
>
> It is getting PCI IRQ 3, I checked this in the sandpoint_setup file
> where the IDE configuration is, all I had todo was change the PCI device
> and vendor ID to promise (very nice :-) I also enabled a PIO printf in
> do_IRQ() and I do see it getting called.
>
> >
> > The fact that it works a walnut but not your board points to board
> > specific problem. Have you double checked IRQ routing?
> >
>
> Well, when I disable IDE, my IRQ's work fine for the onboard Dec 21143
> and the NS16552 .... I tested a BCM5701 NIC card with the PCI slot
> though so I am at a loss ... perhaps the Calibrate delay is wrong?
> If the IDE timing is off, it might for a toss. What about the bogomips
> reading, does this look right for a 200Mhz Kahlua? I'm out of ideas...
> thoughts?
>
> >
> > Where did you get your hhl-2.4.2 kernel? Try using the one in our
> > Journeyman Edition, available on our ftp site. You can see the config
> > options I turned on for using the Promise card in that kernel.
>
> I did that, from your ftp site. I then modified it to run on the
> Kahlua Mousse board (see also http://www.vooha.com/mousse.html).
>
>
> > >
> > > Why can't I simply run fdisk on /dev/hda and set this up? -- because
> > there
> > > is no /dev/hda or /dev/ide/xx devices showing up with devfsd
> > > running!
>
> I love devfsd ---- I've always loved devfsd ... (smiles) :-)
>
> Oh yeah, I installed all of the software in ppc82xx_apps from
> ftp.mvista.com in Journeyman under ppc from the cd distribution.
> Still, when it boots up, it only runs a getty. It is not starting
> the network nor is it mounting entries in fstab. Seems like none
> of the rc scripts work ..
>
>
> Anyhow, thanks Guys, more later ...
> -James
>
>
>
> > > Also, how come it is not running in Native mode? Does anyone know
> > > what the changes are to the pdc driver to enable this?
> > >
> > > BTW, I would gladly pay MVista for support, If I had the budget for this
> > > sort of thing ... unfortunately, I don't at this point in time :-(
> >
> > np.
> >
> > >
> > >
> > > Thoughts?
> > >
> > >
> > Thanks, -James
> > >
> > >
> > >
> > >> The Promise Ultra-66 is known to work on several ppc platforms in
> > hhl-2.4.2.
> > >> Do you have the DMA options turned on in the kernel?
> > >>
> > >> btw, if you are a MV customer contact support as this something we
> > support on the walnut and several other ppc platforms.
> > >>
> > >>
> > >> James F Dougherty wrote:
> > >>
> > >>
> > >>> (LOL) .... I have the Promise Ultra-66..... since it's
> > >>> not working, this is what prompted me to look for an
> > >>> alternative ..
> > >>>
> > >>> I have ported MontaVista HHL (Linux 2.4.2) to
> > >>> a custom processor board, and have been
> > >>> trying to run fdisk to setup a mountable root
> > >>> drive.
> > >>>
> > >>> Kernel is 2.4.2, with the Promise driver compiled in.
> > >>> I boot up, and it finds the card, and sometimes hdb
> > >>> is showing up in DMA mode ...
> > >>>
> > >>>
> > >>> pty: 256 Unix98 ptys configured
> > >>> block: queued sectors max/low 41386kB/13795kB, 128 slots per queue
> > >>> RAMDISK driver initialized: 16 RAM disks of 524288K size 1024 blocksize
> > >>> Uniform Multi-Platform E-IDE driver Revision: 6.31
> > >>> ide: Assuming 33MHz system bus speed for PIO modes; override with
> > idebus=xx
> > >>> PDC20262: IDE controller on PCI bus 00 dev 70
> > >>> PDC20262: chipset revision 1
> > >>> PDC20262: not 100% native mode: will probe irqs later
> > >>> PDC20262: ROM enabled at 0x000d0000
> > >>> PDC20262: (U)DMA Burst Bit DISABLED Primary PCI Mode Secondary PCI
> > Mode.
> > >>> PDC20262: FORCING BURST BIT 0x00 -> 0x01 ACTIVE
> > >>> ide0: BM-DMA at 0xbfff00-0xbfff07, BIOS settings: hda:pio, hdb:pio
> > >>> ide1: BM-DMA at 0xbfff08-0xbfff0f, BIOS settings: hdc:pio, hdd:pio
> > >>> udf: registering filesystem
> > >>> loop: loaded (max 8 devices)
> > >>> Serial driver version 5.02 (2000-08-09) with MANY_PORTS SHARE_IRQ
> > SERIAL_PCI
> > >>> enabled
> > >>>
> > >>>
> > >>> mousse-(8):[11:29pm]: [/proc/ide>cat drivers
> > >>> ide-cdrom version 4.59
> > >>> ide-disk version 1.10
> > >>> mousse-(9):[11:29pm]: [/proc/ide>cat pdc202xx
> > >>>
> > >>> PDC20262 Chipset.
> > >>> ------------------------------- General Status
> > ---------------------------------
> > >>> Burst Mode : enabled
> > >>> Host Mode : Tri-Stated
> > >>> Bus Clocking : 100 External
> > >>> IO pad select : 10 mA
> > >>> Status Polling Period : 15
> > >>> Interrupt Check Status Polling Delay : 15
> > >>> --------------- Primary Channel ---------------- Secondary Channel
> > -------------
> > >>> disabled disabled
> > >>> 66 Clocking enabled enabled
> > >>> Mode MASTER Mode MASTER
> > >>> Error Error
> > >>> --------------- drive0 --------- drive1 -------- drive0 ----------
> > drive1 ------
> > >>> DMA enabled: yes yes yes yes
> > >>> DMA Mode: NOTSET NOTSET NOTSET
> > NOTSET
> > >>> PIO Mode: NOTSET NOTSET NOTSET
> > NOTSET
> > >>> mousse-(10):[11:29pm]: [/proc/ide>
> > >>>
> > >>>
> > >>> Now, when I run fdisk .... it either cannot find the device (when
> > >>> compiled kernel with INCLUDE_DEVFS
> > >>>
> > >>> mousse-(17):[11:31pm]: [/>devfsd /dev
> > >>> Started device management daemon for /dev
> > >>> mousse-(18):[11:31pm]: [/>fdisk /dev/hda
> > >>> modprobe: Can't locate module /dev/hda
> > >>> modprobe: Can't locate module /dev/hda
> > >>>
> > >>> Unable to open /dev/hda
> > >>> mousse-(19):[11:31pm]: [/>fdisk /dev/hdb
> > >>> modprobe: Can't locate module /dev/hdb
> > >>> modprobe: Can't locate module /dev/hdb
> > >>>
> > >>> Unable to open /dev/hdb
> > >>> mousse-(20):[11:31pm]: [/>
> > >>>
> > >>> Or, when I run with a standard dev filesystem and the device
> > >>> files created with scripts/MAKEDEV.ide it gives me an error,
> > >>> cant find module block-3 blah blah blah...
> > >>>
> > >>>
> > >>> Any ideas?
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>> A good shot is always the Promise cards. I know at least the 66 used
> > >>>> to work, and I'm reasonably sure the 100 does. You'll need 2.4 for it
> > >>>> though, and no booting off it.
> > >>>>
> > >>>> On Sun, Aug 12, 2001 at 11:49:14PM -0700, James F Dougherty wrote:
> > >>>>
> > >>>>
> > >>>>> I guess my question is which drivers run in native mode
> > >>>>> (e.g. do not require an x86 ROM BIOS).
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>>>> Anyone have any suggestions for a PCI IDE card to
> > >>>>>> use for PPC/Linux? Something one could buy down
> > >>>>>> at a Frye's would be great.
> > >>>>>>
> > >>>>>>
> > >>>>>> Many thanks in advance.
> > >>>>>> -James
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>> --
> > >>>> Daniel Jacobowitz Carnegie Mellon University
> > >>>> MontaVista Software Debian GNU/Linux Developer
> > >>>>
> > >>>>
> > >>>
> > >>>
> > >>
> > >>
> >
> >
> >
> > >> >> support on the walnut and several other ppc platforms.
> > >>
> > >>
> > >> James F Dougherty wrote:
> > >>
> > >>
> > >>> (LOL) .... I have the Promise Ultra-66..... since it's not
> > >>> working, this is what prompted me to look for an alternative ..
> > >>>
> > >>> I have ported MontaVista HHL (Linux 2.4.2) to a custom processor
> > >>> board, and have been trying to run fdisk to setup a mountable
> > >>> root drive.
> > >>>
> > >>> Kernel is 2.4.2, with the Promise driver compiled in. I boot up,
> > >>> and it finds the card, and sometimes hdb is showing up in DMA
> > >>> mode ...
> > >>>
> > >>>
> > >>> pty: 256 Unix98 ptys configured block: queued sectors max/low
> > >>> 41386kB/13795kB, 128 slots per queue RAMDISK driver
> > >>> initialized: 16 RAM disks of 524288K size 1024 blocksize Uniform
> > >>> Multi-Platform E-IDE driver Revision: 6.31 ide: Assuming 33MHz
> > >>> system bus speed for PIO modes; override with idebus=xx PDC20262:
> > >>> IDE controller on PCI bus 00 dev 70 PDC20262: chipset revision
> > >>> 1 PDC20262: not 100% native mode: will probe irqs later PDC20262:
> > >>> ROM enabled at 0x000d0000 PDC20262: (U)DMA Burst Bit DISABLED
> > >>> Primary PCI Mode Secondary PCI Mode. PDC20262: FORCING BURST
> > >>> BIT 0x00 -> 0x01 ACTIVE ide0: BM-DMA at 0xbfff00-0xbfff07, BIOS
> > >>> settings: hda:pio, hdb:pio ide1: BM-DMA at 0xbfff08-0xbfff0f,
> > >>> BIOS settings: hdc:pio, hdd:pio udf: registering filesystem loop:
> > >>> loaded (max 8 devices) Serial driver version 5.02 (2000-08-09)
> > >>> with MANY_PORTS SHARE_IRQ SERIAL_PCI enabled
> > >>>
> > >>>
> > >>> mousse-(8):[11:29pm]: [/proc/ide>cat drivers ide-cdrom version
> > >>> 4.59 ide-disk version 1.10 mousse-(9):[11:29pm]: [/proc/ide>cat
> > >>> pdc202xx
> > >>>
> > >>> PDC20262 Chipset. ------------------------------- General Status
> > >>> --------------------------------- Burst Mode
> > >>> : enabled Host Mode :
> > >>> Tri-Stated Bus Clocking : 100 External IO
> > >>> pad select : 10 mA Status Polling
> > >>> Period : 15 Interrupt Check Status Polling Delay
> > >>> : 15 --------------- Primary Channel ----------------
> > >>> Secondary Channel ------------- disabled
> > >>> disabled 66 Clocking enabled
> > >>> enabled Mode MASTER Mode MASTER Error
> > >>> Error --------------- drive0 ---------
> > >>> drive1 -------- drive0 ---------- drive1 ------ DMA enabled:
> > >>> yes yes yes yes DMA
> > >>> Mode: NOTSET NOTSET NOTSET
> > >>> NOTSET PIO Mode: NOTSET NOTSET
> > >>> NOTSET NOTSET mousse-(10):[11:29pm]: [/proc/ide>
> > >>>
> > >>>
> > >>> Now, when I run fdisk .... it either cannot find the device (when
> > compiled kernel with INCLUDE_DEVFS
> > >>>
> > >>> mousse-(17):[11:31pm]: [/>devfsd /dev
> > >>> Started device management daemon for /dev
> > >>> mousse-(18):[11:31pm]: [/>fdisk /dev/hda
> > >>> modprobe: Can't locate module /dev/hda
> > >>> modprobe: Can't locate module /dev/hda
> > >>>
> > >>> Unable to open /dev/hda
> > >>> mousse-(19):[11:31pm]: [/>fdisk /dev/hdb
> > >>> modprobe: Can't locate module /dev/hdb
> > >>> modprobe: Can't locate module /dev/hdb
> > >>>
> > >>> Unable to open /dev/hdb
> > >>> mousse-(20):[11:31pm]: [/>
> > >>>
> > >>> Or, when I run with a standard dev filesystem and the device
> > >>> files created with scripts/MAKEDEV.ide it gives me an error,
> > >>> cant find module block-3 blah blah blah...
> > >>>
> > >>>
> > >>> Any ideas?
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>> A good shot is always the Promise cards. I know at least the 66 used
> > >>>> to work, and I'm reasonably sure the 100 does. You'll need 2.4 for it
> > >>>> though, and no booting off it.
> > >>>>
> > >>>> On Sun, Aug 12, 2001 at 11:49:14PM -0700, James F Dougherty wrote:
> > >>>>
> > >>>>
> > >>>>> I guess my question is which drivers run in native mode
> > >>>>> (e.g. do not require an x86 ROM BIOS).
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>>>> Anyone have any suggestions for a PCI IDE card to
> > >>>>>> use for PPC/Linux? Something one could buy down
> > >>>>>> at a Frye's would be great.
> > >>>>>>
> > >>>>>>
> > >>>>>> Many thanks in advance.
> > >>>>>> -James
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>> --
> > >>>> Daniel Jacobowitz Carnegie Mellon University
> > >>>> MontaVista Software Debian GNU/Linux Developer
> > >>>>
> > >>>>
> > >>>
> > >>>
> > >>
> > >>
> >
> >
> >
> > >>> >>> compiled kernel with INCLUDE_DEVFS
> > >>>
> > >>> mousse-(17):[11:31pm]: [/>devfsd /dev Started device management
> > >>> daemon for /dev mousse-(18):[11:31pm]: [/>fdisk /dev/hda modprobe:
> > >>> Can't locate module /dev/hda modprobe: Can't locate module
> > >>> /dev/hda
> > >>>
> > >>> Unable to open /dev/hda mousse-(19):[11:31pm]: [/>fdisk /dev/hdb
> > modprobe:
> > >>> Can't locate module /dev/hdb modprobe: Can't locate module
> > >>> /dev/hdb
> > >>>
> > >>> Unable to open /dev/hdb mousse-(20):[11:31pm]: [/>
> > >>>
> > >>> Or, when I run with a standard dev filesystem and the device files
> > >>> created with scripts/MAKEDEV.ide it gives me an error, cant
> > >>> find module block-3 blah blah blah...
> > >>>
> > >>>
> > >>> Any ideas?
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>> A good shot is always the Promise cards. I know at least the
> > >>>> 66 used to work, and I'm reasonably sure the 100 does.
> > >>>> You'll need 2.4 for it though, and no booting off it.
> > >>>>
> > >>>> On Sun, Aug 12, 2001 at 11:49:14PM -0700, James F Dougherty
> > >>>> wrote:
> > >>>>
> > >>>>
> > >>>>> I guess my question is which drivers run in native mode (e.g.
> > >>>>> do not require an x86 ROM BIOS).
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>>>> Anyone have any suggestions for a PCI IDE card to use for
> > >>>>>> PPC/Linux? Something one could buy down at a Frye's would
> > >>>>>> be great.
> > >>>>>>
> > >>>>>>
> > >>>>>> Many thanks in advance. -James
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>> -- Daniel Jacobowitz Carnegie Mellon
> > >>>> University MontaVista Software
> > >>>> Debian GNU/Linux Developer
> > >>>>
> > >>>>
> > >>>
> > >>>
> > >>
> > >>
> >
> >
> >
>
>
** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/
More information about the Linuxppc-dev
mailing list