ARGH, weird SCSI I/O errors on second/third drive on Debian/PPC
I am having some weird, weird SCSI I/O errors. I'm running Debian/PowerPC on an "Old World" (beige) Mac-- namely, a Power Mac 7300/180 (the 7300s were sold overseas, and are uncommon in the USA; they are roughly on par with the 7500s, I believe.. if not a bit higher-end) with a G3/220 CPU upgrade. The system has worked fine for a long while. However, when I tried to install some more SCSI drives into it (KNOWN WORKING drives), I get some curious I/O errors which prevent me from using the system. The system only seems to work properly when I use it with ONE hard drive. When I add more, I get these problems. The poop: * First hard drive is a 1GB Quantum * Second hard drive is a 4GB something-or-other * Third hard drive is one of those big old "boat anchor"/"freight train" 4GB drives. It fills the whole bottom bay of the Mac (twice as high as a typical "modern" hard drive). It's at the end of the SCSI chain, so I enabled the terminator on it. That is the only termination on the chain, as far as I know. The information on this drive and the (many) jumper options availble on it (nice drive!) is at: http://www.seagate.com/support/disc/scsi/st15150n.html (yes, it's a Seagate Barracuda ST15150N). The problem seems to happen whenever I'm working (reading/writing) on the second or third drive. I get I/O errors, and kernelspam like this: Dec 27 19:36:35 fortitude kernel: EXT2-fs error (device sd(8,34)): ext2_free_blocks: Freeing blocks not in datazone - block = 3717479526, count = 1 Dec 27 19:36:35 fortitude kernel: Remounting filesystem read-only Dec 27 19:36:35 fortitude kernel: EXT2-fs error (device sd(8,34)): ext2_free_blocks: Freeing blocks not in datazone - block = 620094437, count = 1 Dec 27 19:36:35 fortitude kernel: Remounting filesystem read-only cks: Freeing blocks not in datazone - block = 2460472751, count = 1 cks: Freeing blocks not in datazone - block = 1739398958, count = 1 For instance, whenever I try to mkisofs from a directory tree, after some given percentage (it will get maybe 1/3 to 1/2 the way through a disc) it will say "mkisofs: Input/output error. cannot read from (FILENAME)", and then the drive begins going "click, click, click, click, click, click, click," slowly and rhythmically (around 1.5-2 times a second, I'd say). At this point, the machine is essentially unusable, since the whole SCSI bus seems out of whack; even trying to do an ls on a directory on the FIRST drive screws up. Like this: fortitude:/home/shared/cds/cosmo# cd /var/log fortitude:/var/log# ls -l (SYSTEM JUST SITS THERE, RHYTMICALLY CLICKING AWAY) The kernel I use is 2.4.20; I compiled it myself. Other than this bug, it works perfectly. Please, could someone help me? Is this a termination issue? Is this a known bug? What in heck is wrong with my Mac? --Jessica -- J e s s i c aL e a hB l a n k
Re: ARGH, weird SCSI I/O errors on second/third drive on Debian/PPC
The drives BOTH worked fine under Mac OS. I put 'em into the Linux box and suddenly both go.. klik klik klik klik klik klik klik On Fri, 27 Dec 2002, Chris Tillman wrote: > On Fri, Dec 27, 2002 at 06:49:50PM -0700, Jessica Blank wrote: > > I am having some weird, weird SCSI I/O errors. I'm running Debian/PowerPC > > on an "Old World" (beige) Mac-- namely, a Power Mac 7300/180 (the 7300s > > were sold overseas, and are uncommon in the USA; they are roughly on par > > with the 7500s, I believe.. if not a bit higher-end) with a G3/220 CPU > > upgrade. > > > > The system has worked fine for a long while. However, when I tried to > > install some more SCSI drives into it (KNOWN WORKING drives), I get some > > curious I/O errors which prevent me from using the system. The system only > > seems to work properly when I use it with ONE hard drive. When I add more, > > I get these problems. > > > > The poop: > > > > * First hard drive is a 1GB Quantum > > * Second hard drive is a 4GB something-or-other > > * Third hard drive is one of those big old "boat anchor"/"freight train" > > 4GB drives. It fills the whole bottom bay of the Mac (twice as high as a > > typical "modern" hard drive). It's at the end of the SCSI chain, so I > > enabled the terminator on it. That is the only termination on the chain, > > as far as I know. The information on this drive and the (many) jumper > > options availble on it (nice drive!) is at: > > http://www.seagate.com/support/disc/scsi/st15150n.html (yes, it's a > > Seagate Barracuda ST15150N). > > > > The problem seems to happen whenever I'm working (reading/writing) on the > > second or third drive. I get I/O errors, and kernelspam like this: > > > > Dec 27 19:36:35 fortitude kernel: EXT2-fs error (device sd(8,34)): > > ext2_free_blocks: Freeing blocks not in datazone - block = 3717479526, > > count = 1 > > Dec 27 19:36:35 fortitude kernel: Remounting filesystem read-only > > Dec 27 19:36:35 fortitude kernel: EXT2-fs error (device sd(8,34)): > > ext2_free_blocks: Freeing blocks not in datazone - block = 620094437, > > count = 1 > > Dec 27 19:36:35 fortitude kernel: Remounting filesystem read-only > > cks: Freeing blocks not in datazone - block = 2460472751, count = 1 > > cks: Freeing blocks not in datazone - block = 1739398958, count = 1 > > > > For instance, whenever I try to mkisofs from a directory tree, after some > > given percentage (it will get maybe 1/3 to 1/2 the way through a disc) it > > will say "mkisofs: Input/output error. cannot read from (FILENAME)", and > > then the drive begins going "click, click, click, click, click, click, > > click," slowly and rhythmically (around 1.5-2 times a second, I'd say). > > > > At this point, the machine is essentially unusable, since the whole SCSI > > bus seems out of whack; even trying to do an ls on a directory on the > > FIRST drive screws up. Like this: > > > > fortitude:/home/shared/cds/cosmo# cd /var/log > > fortitude:/var/log# ls -l > > (SYSTEM JUST SITS THERE, RHYTMICALLY CLICKING AWAY) > > > > The kernel I use is 2.4.20; I compiled it myself. Other than this bug, it > > works perfectly. > > > > Please, could someone help me? Is this a termination issue? Is this a > > known bug? What in heck is wrong with my Mac? > > Whenever I got that clicking noise, I always assumed the drive was > a goner and replaced it. It works fine on another machine? > > -- J e s s i c aL e a hB l a n k
Re: ARGH, weird SCSI I/O errors on second/third drive on Debian/PPC
On Fri, 27 Dec 2002, Michael D. Crawford wrote: > Could the original drive still be terminated after you have installed the > other > drives? It is probably worse to have extra termination than to have none at > all. Make sure you've got it right. The original drive was NOT installed at the end of the chain. (The original Apple-branded CD-ROM drive was...) So I don't think termination would be an issue for this drive, no? How do I know if I got it right? I wish there was a "scsidiag" program that would say "Your termination sucks, try this; your SCSI IDs have a conflict, try this..." > > I think on my 8500 the original configuration had the boot drive in the > middle > of the bus, the CD drive at one end and the host bus adapter at the other > end. > The CD drive was terminated. > > If your CD drive is providing the termination there is probably a jumber that > you can remove. If your hard drive is terminated on the drive, there is > probably a resistor pack you can remove - although it is probably harder to > be > certain you're pulling the right thing off. > The thing that confuses me is thus: There are several termination options for the drive I placed at the end of the SCSI chain (read the Web page I referred to: http://www.seagate.com/support/disc/scsi/st15150n.html ). I do not know which ones are correct. "Terminate power from drive"? "Terminate power to SCSI bus"? I know nothing of these things. > There is a Mac OS utility called SCSI probe you might find helpful. The > version of it that Adaptec distributes is a pretty fancy version. I think it > might tell you if your bus is terminated correctly. I'm not sure but you > might > try it. Check at http://www.adaptec.com/ it should be available as a free > download. > > Mike > -- J e s s i c aL e a hB l a n k
Re: ARGH, weird SCSI I/O errors on second/third drive on Debian/PPC
On Sat, 28 Dec 2002, Michel Lanners wrote: > On 27 Dec, this message from Jessica Blank echoed through cyberspace: > > On Fri, 27 Dec 2002, Michael D. Crawford wrote: > > > >> Could the original drive still be terminated after you have installed the > >> other > >> drives? It is probably worse to have extra termination than to have none > >> at > >> all. Make sure you've got it right. > > > > The original drive was NOT installed at the end of the chain. (The > > original Apple-branded CD-ROM drive was...) So I don't think termination > > would be an issue for this drive, no? > > How did you install two more drives in the 7300? Did you remove the > CDRom? Yes... > > Anyway, your SCSI chain needs to _physically_ look like this: > > Hostlast device > | | | .| > DeviceDeviceDevice Device > > Only the last device must have termination, none of the others must have > (well, the Host does, but that's automatic). > > So, verify your SCSI cabling in the first place. All other devices > except the physically last one (i.e. at the end of the flat cable) need > to have termination switched off. Verify that with the help of the > manufacterer's jumper info. > > Then, you can try to use another drive providing termination, in case > that specific drive has problems with its termination. That means > physically rearranging drives. > > > How do I know if I got it right? I wish there was a "scsidiag" program > > that would say "Your termination sucks, try this; your SCSI IDs have a > > conflict, try this..." > > There is no other way than to open the case and have a look, sorry... > > > The thing that confuses me is thus: There are several termination options > > for the drive I placed at the end of the SCSI chain (read the Web page I > > referred to: http://www.seagate.com/support/disc/scsi/st15150n.html ). I > > do not know which ones are correct. "Terminate power from drive"? > > "Terminate power to SCSI bus"? I know nothing of these things. > > Ah, Seagate. They have a jumper to enable termination: that is putting > 'the plug on the cable'. This 'plug' needs power, and the other jumpers > specify where this power comes from. 'Term power from drive' uses the > driv'es power for this. 'Term power to bus' is independant of the > on-board terminator; it serves to provide termination power to the > reserved wire in the SCSI cable. Thank you. Which one should I use though? > > But, there is something else you need to make sure. > > The error messages you describe _may_ indicate that the filesystems you > created are larger than the actual drive. This can happen when > partitioning with some (older?) versions of mac-fdisk (aka pdisk). > I seriously doubt this. This is on a fresh install of Woody (Debian 3.0). Everyone keeps saying that I have problems with the filesystems. I've re-created the filesystems a billion times. I doubt that is it... Do you know which versions of mac-fdisk had these problems? > Try using badblocks to verify the drives are OK. Or try using 'dd' to > read the entire drive (something like dd if=/dev/sd of=/dev/null > bs=512 count=). > > If this is your problem, you need to specify the correct drive size when > creating a partition map with mac-fdisk. > > Cheers > > Michel > > - > Michel Lanners | " Read Philosophy. Study Art. > 23, Rue Paul Henkes|Ask Questions. Make Mistakes. > L-1710 Luxembourg | > email [EMAIL PROTECTED]| > http://www.cpu.lu/~mlan| Learn Always. " > -- J e s s i c aL e a hB l a n k
Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
I am having a big problem here. My nice little "CD burning server", under my bed, is crippled because the large (40GB) IDE drive in it simply does not detect. This is a KNOWN WORKING drive. I just pulled it out of a working system. I've tried compiling IDE Disk Support and the Promise controller driver (the two necessary modules, of course) as modules, and I've tried compiling them straight into the kernel. NEITHER WORKS. When I insert ide-mod and ide-disk, I get: Apr 11 10:19:28 gandalf kernel: Uniform Multi-Platform E-IDE driver Revision: 6.31 Apr 11 10:19:28 gandalf kernel: ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx Apr 11 10:19:28 gandalf kernel: PDC20246: IDE controller on PCI bus 00 dev 70 Apr 11 10:19:28 gandalf kernel: PDC20246: chipset revision 1 Apr 11 10:19:28 gandalf kernel: PDC20246: not 100%% native mode: will probe irqs later Apr 11 10:19:28 gandalf kernel: PDC20246: ROM enabled at 0x8081 Apr 11 10:19:28 gandalf kernel: PDC20246: (U)DMA Burst Bit DISABLED Primary PCI Mode Secondary PCI Mode. Apr 11 10:19:28 gandalf kernel: ide2: BM-DMA at 0x0800-0x0807, BIOS settings: hde:DMA, hdf:DMA Apr 11 10:19:28 gandalf kernel: ide3: BM-DMA at 0x0808-0x080f, BIOS settings: hdg:DMA, hdh:DMA ...however, as you can see, it does not detect the disk itself! This is very frustrating. And yes, I have already checked the connections to the disk. They are fine. :) Any help? J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
It was pulled out of a PeeCee Debian box. And to answer your question: Yes. My "Partition Types" setup (for the record, I am using KERNEL VERSION 2.4.20...) looks like this: x x [*] Advanced partition selection x x [ ] Acorn partition support x x [ ] Alpha OSF partition support x x [ ] Amiga partition table support x x [ ] Atari partition table support x x [*] Macintosh partition map support x x [*] PC BIOS (MSDOS partition tables) support x x [*] BSD disklabel (FreeBSD partition tables) support x x [ ] Minix subpartition support x x [ ] Solaris (x86) partition table support x x [ ] Unixware slices support x x [ ] SGI partition support x x [ ] Ultrix partition table support x x [ ] Sun partition tables support x x [ ] EFI GUID Partition support On Fri, 11 Apr 2003, Michel [ISO-8859-1] Dänzer wrote: > On Fre, 2003-04-11 at 16:24, Jessica Blank wrote: > > > > My nice little "CD burning server", under my bed, is crippled because the > > large (40GB) IDE drive in it simply does not detect. > > > > This is a KNOWN WORKING drive. I just pulled it out of a working system. > > [...] > > > Apr 11 10:19:28 gandalf kernel: PDC20246: IDE controller on PCI bus 00 dev > > 70 > > Apr 11 10:19:28 gandalf kernel: PDC20246: chipset revision 1 > > Apr 11 10:19:28 gandalf kernel: PDC20246: not 100%% native mode: will > > probe irqs later > > Apr 11 10:19:28 gandalf kernel: PDC20246: ROM enabled at 0x8081 > > Apr 11 10:19:28 gandalf kernel: PDC20246: (U)DMA Burst Bit DISABLED > > Primary PCI Mode Secondary PCI Mode. > > Apr 11 10:19:28 gandalf kernel: ide2: BM-DMA at 0x0800-0x0807, BIOS > > settings: hde:DMA, hdf:DMA > > Apr 11 10:19:28 gandalf kernel: ide3: BM-DMA at 0x0808-0x080f, BIOS > > settings: hdg:DMA, hdh:DMA > > > > ...however, as you can see, it does not detect the disk itself! > > Have you built support for whatever type the partition table on the disk > is? > > > -- > Earthling Michel Dänzer \ Debian (powerpc), XFree86 and DRI developer > Software libre enthusiast \ http://svcs.affero.net/rm.php?r=daenzer > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
On Fri, 11 Apr 2003, Benjamin Herrenschmidt wrote: > On Fri, 2003-04-11 at 16:24, Jessica Blank wrote: > > I am having a big problem here. > > > > My nice little "CD burning server", under my bed, is crippled because the > > large (40GB) IDE drive in it simply does not detect. > > > > This is a KNOWN WORKING drive. I just pulled it out of a working system. > > > > I've tried compiling IDE Disk Support and the Promise controller driver > > (the two necessary modules, of course) as modules, and I've tried > > compiling them straight into the kernel. NEITHER WORKS. > > > > When I insert ide-mod and ide-disk, I get: > > > > Apr 11 10:19:28 gandalf kernel: Uniform Multi-Platform E-IDE driver > > Revision: 6.31 > > Apr 11 10:19:28 gandalf kernel: ide: Assuming 33MHz system bus speed for > > PIO modes; override with idebus=xx > > Apr 11 10:19:28 gandalf kernel: PDC20246: IDE controller on PCI bus 00 dev > > 70 > > Apr 11 10:19:28 gandalf kernel: PDC20246: chipset revision 1 > > Apr 11 10:19:28 gandalf kernel: PDC20246: not 100%% native mode: will > > probe irqs later > > Apr 11 10:19:28 gandalf kernel: PDC20246: ROM enabled at 0x8081 > > Apr 11 10:19:28 gandalf kernel: PDC20246: (U)DMA Burst Bit DISABLED > > Primary PCI Mode Secondary PCI Mode. > > Apr 11 10:19:28 gandalf kernel: ide2: BM-DMA at 0x0800-0x0807, BIOS > > settings: hde:DMA, hdf:DMA > > Apr 11 10:19:28 gandalf kernel: ide3: BM-DMA at 0x0808-0x080f, BIOS > > settings: hdg:DMA, hdh:DMA > > > > ...however, as you can see, it does not detect the disk itself! > > Actually, I can't see as the disk probing happens later during boot... > These messages weren't FROM the boot. I already tried compiling the ide-disk and ide-mod modules into the kernel... they did not work. So, getting sick and tired of rebooting the machine around 50 times, I decided to compile them as modules, so I could quickly work on them, insmod them and rmmod them. > > This is very frustrating. > > > > And yes, I have already checked the connections to the disk. They are > > fine. :) > > Can you send me the complete dmesg log along with the version of the > kernel you are using ? > 2.4.20 and... the dmesg is attached. Also, below (in case y'all wanna read it inline). NOTE: Again, I compiled ide-mod and ide-disk AS MODULES after getting SICK AND TIRED OF REBOOTING since compiling them into the kernel itself didn't work either... at the end of this log, you can see the output generated by me insmodding ide-mod and then ide-disk, twice. As you can see, ide-mod seems to see the controller (it's a Promise Ultra DMA/33) but no disks, and ide-disk doesn't seem to do anything useful at all... HEeelpp!!! Many thanks in advance, Ben ^_^ > Ben. > J e s s i c aL e a hB l a n k LOG FOLLOWS: Memory BAT mapping: BAT2=8Mb, BAT3=8Mb, residual: 64Mb Total memory = 80MB; using 256kB for hash table (at c02c) Linux version 2.4.20 ([EMAIL PROTECTED]) (gcc version 2.95.4 20011002 (Debian prerelease)) #4 Fri Apr 11 00:00:25 EDT 2003 Found a Grand Central mac-io controller, rev: 2, mapped at 0xfdf0 PowerMac motherboard: Unknown PowerSurge Cache coherency enabled for bandit/PSX Found Bandit PCI host bridge at 0xf200. Firmware bus number: 0->0 On node 0 totalpages: 20480 zone(0): 20480 pages. zone(1): 0 pages. zone(2): 0 pages. Kernel command line: root=/dev/sda6 rw System has 32 possible interrupts GMT Delta read from XPRAM: 0 minutes, DST: off Console: colour dummy device 80x25 Calibrating delay loop... 89.49 BogoMIPS Memory: 78076k available (1516k kernel code, 960k data, 116k init, 0k highmem) Dentry cache hash table entries: 16384 (order: 5, 131072 bytes) Inode cache hash table entries: 8192 (order: 4, 65536 bytes) Mount-cache hash table entries: 2048 (order: 2, 16384 bytes) Buffer-cache hash table entries: 4096 (order: 2, 16384 bytes) Page-cache hash table entries: 32768 (order: 5, 131072 bytes) POSIX conformance testing by UNIFIX PCI: Probing PCI hardware Macintosh CUDA driver v0.5 for Unified ADB. Linux NET4.0 for Linux 2.4 Based upon Swansea University Computer Society NET3.039 Initializing RT netlink socket Starting kswapd NTFS driver v1.1.22 [Flags: R/O] i2c-core.o: i2c core module Total VRAM = 1MB 0001 Monitor sense value = 0x62b, using video mode 12 and color mode 1. Console: switching to colour frame buffer device 100x37 fb0: platinum frame buffer device input0: Macintosh mouse button emulation pty: 256 Unix98 ptys configured fd0: SWIM3 floppy controller loop: loaded (max 8 devices) Linux Tulip driver version 0.9.15-pre12 (Aug 9, 2002) PCI: Enabling device 00:0d.0 (0084 -> 0087) tulip0: no phy info, aborting mtable build eth0: Macr
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
On Fri, 11 Apr 2003, Benjamin Herrenschmidt wrote: > On Fri, 2003-04-11 at 17:48, Jessica Blank wrote: > > > These messages weren't FROM the boot. I already tried compiling the > > ide-disk and ide-mod modules into the kernel... they did not work. So, > > getting sick and tired of rebooting the machine around 50 times, I decided > > to compile them as modules, so I could quickly work on them, insmod them > > and rmmod them. > > IDE as module is known brain dead. Compile them on in the kernel and > send me the whole dmesg output with the ide-disk compile in. Gyaah. Round and round and round we go... ^_^;;; Oh, another thing that pisses me off about 2.4.20.. WHERE THE HELL DID THEY PUT HFS SUPPORT!?!??! I cannot find the HFS option ANYWHERE under Filesystems. I CAN find it on 2.2.25... what's wrong with this picture!? I KNOW the driver is in there; I see hfs.c... but I don't see the config option. Ben, hopefully your patch will fix this annoyance. I'm sure it will, no? :) > > > > > This is very frustrating. > > > > > > > > And yes, I have already checked the connections to the disk. They are > > > > fine. :) > > > > > > Can you send me the complete dmesg log along with the version of the > > > kernel you are using ? > > > > > > > 2.4.20 and... the dmesg is attached. Also, below (in case y'all wanna read > > it inline). NOTE: Again, I compiled ide-mod and ide-disk AS MODULES after > > getting SICK AND TIRED OF REBOOTING since compiling them into the kernel > > itself didn't work either... at the end of this log, you can see the > > output generated by me insmodding ide-mod and then ide-disk, twice. As you > > can see, ide-mod seems to see the controller (it's a Promise Ultra > > DMA/33) but no disks, and ide-disk doesn't seem to do anything useful at > > all... > > Can you try my patches ? (2.4.20-ben10 patch available on kernel.org) > and send me a log with those ? > Certainly! I'd love to. What directory izzit? Can you provide a full URL? I remember you; I used your kernel on my G4 Cube way back when. You rule. :) > Ben. > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
Ben: Found your patches, installing 'em, compiling IDE into the kernel itself (i.e. not as a module). Will send dmesg if that dunna fix my problem. Thanks! On Fri, 11 Apr 2003, Benjamin Herrenschmidt wrote: > On Fri, 2003-04-11 at 17:48, Jessica Blank wrote: > > > These messages weren't FROM the boot. I already tried compiling the > > ide-disk and ide-mod modules into the kernel... they did not work. So, > > getting sick and tired of rebooting the machine around 50 times, I decided > > to compile them as modules, so I could quickly work on them, insmod them > > and rmmod them. > > IDE as module is known brain dead. Compile them on in the kernel and > send me the whole dmesg output with the ide-disk compile in. > > > > > This is very frustrating. > > > > > > > > And yes, I have already checked the connections to the disk. They are > > > > fine. :) > > > > > > Can you send me the complete dmesg log along with the version of the > > > kernel you are using ? > > > > > > > 2.4.20 and... the dmesg is attached. Also, below (in case y'all wanna read > > it inline). NOTE: Again, I compiled ide-mod and ide-disk AS MODULES after > > getting SICK AND TIRED OF REBOOTING since compiling them into the kernel > > itself didn't work either... at the end of this log, you can see the > > output generated by me insmodding ide-mod and then ide-disk, twice. As you > > can see, ide-mod seems to see the controller (it's a Promise Ultra > > DMA/33) but no disks, and ide-disk doesn't seem to do anything useful at > > all... > > Can you try my patches ? (2.4.20-ben10 patch available on kernel.org) > and send me a log with those ? > > Ben. > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
On Fri, 11 Apr 2003, Benjamin Herrenschmidt wrote: > On Fri, 2003-04-11 at 17:54, Jessica Blank wrote: > > On Fri, 11 Apr 2003, Benjamin Herrenschmidt wrote: > > > > > On Fri, 2003-04-11 at 17:48, Jessica Blank wrote: > > > > > > > These messages weren't FROM the boot. I already tried compiling the > > > > ide-disk and ide-mod modules into the kernel... they did not work. So, > > > > getting sick and tired of rebooting the machine around 50 times, I > > > > decided > > > > to compile them as modules, so I could quickly work on them, insmod them > > > > and rmmod them. > > > > > > IDE as module is known brain dead. Compile them on in the kernel and > > > send me the whole dmesg output with the ide-disk compile in. > > > > Gyaah. Round and round and round we go... ^_^;;; > > > > Oh, another thing that pisses me off about 2.4.20.. WHERE THE HELL DID > > THEY PUT HFS SUPPORT!?!??! I cannot find the HFS option ANYWHERE under > > Filesystems. I CAN find it on 2.2.25... what's wrong with this picture!? I > > KNOW the driver is in there; I see hfs.c... but I don't see the config > > option. > > Hrm... It seem to be here... You may need CONFIG_EXPERIMENTAL Dear God, HFS is _STILL_ experimental!? > > > Ben, hopefully your patch will fix this annoyance. I'm sure it will, no? > > :) > > > > > > > > > > > This is very frustrating. > > > > > > > > > > > > And yes, I have already checked the connections to the disk. They > > > > > > are > > > > > > fine. :) > > > > > > > > > > Can you send me the complete dmesg log along with the version of the > > > > > kernel you are using ? > > > > > > > > > > > > > 2.4.20 and... the dmesg is attached. Also, below (in case y'all wanna > > > > read > > > > it inline). NOTE: Again, I compiled ide-mod and ide-disk AS MODULES > > > > after > > > > getting SICK AND TIRED OF REBOOTING since compiling them into the kernel > > > > itself didn't work either... at the end of this log, you can see the > > > > output generated by me insmodding ide-mod and then ide-disk, twice. As > > > > you > > > > can see, ide-mod seems to see the controller (it's a Promise Ultra > > > > DMA/33) but no disks, and ide-disk doesn't seem to do anything useful at > > > > all... > > > > > > Can you try my patches ? (2.4.20-ben10 patch available on kernel.org) > > > and send me a log with those ? > > > > > > > Certainly! I'd love to. What directory izzit? Can you provide a full URL? > > pub/linux/kernel/people/benh, then get the -ben10 patch against 2.4.20 > > > I remember you; I used your kernel on my G4 Cube way back when. You rule. > > :) > > Heh, well, I'm the maintainer for the PowerMac support, so there are > great chances my kernels work better than standard ones on those > machines ;) Though I'm trying to sync regulary with the official > tree. > What ticks me off is the utter lack of attention paid to non-x86 platforms by Linus and crew. Shouldn't your superb patches be integrated regularly into the main (Linus) kernel tree? :/ Oh well. OK, off to patch I go. > Cheers, > > Ben. > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
Well... okay. This is interesting. Ben, your kernel patches simply don't like my machine. Maybe I'm doing something wrong, but I've configured around a billion trillion bajillion kernels in my life (since age 15) and I see nothing missing. Plain old non-Ben-patched 2.4.20 kernels seem to boot fine on this thing. However, your patched kernel seems to shunt itself immediately into Runlevel 0 or Runlevel 6, since the last thing I see when I boot it up (sorry, the messages scroll by so fast I can't catch them all, and no, they DO NOT end up in /var/log/messages nor /var/log/syslog) is "Setting the System Clock using the Hardware Clock as reference..." ... which is generated by hwclock.sh / hwclockfirst.sh, two startup scripts (in /etc/init.d) which are only loaded by runlevels 0 and 6... This is not making me happy. :( Any ideas what I might be doing wrong? Again, for reference, this machine is a Mac 7200. Oh, and I even tried removing all the PCI cards, effectively returning the Mac to "stock" condition (no IDE, just SCSI; no Tulip NIC, just the built-in MESH)... no go. Same results! Any help? On Fri, 11 Apr 2003, Benjamin Herrenschmidt wrote: > On Fri, 2003-04-11 at 16:24, Jessica Blank wrote: > > I am having a big problem here. > > > > My nice little "CD burning server", under my bed, is crippled because the > > large (40GB) IDE drive in it simply does not detect. > > > > This is a KNOWN WORKING drive. I just pulled it out of a working system. > > > > I've tried compiling IDE Disk Support and the Promise controller driver > > (the two necessary modules, of course) as modules, and I've tried > > compiling them straight into the kernel. NEITHER WORKS. > > > > When I insert ide-mod and ide-disk, I get: > > > > Apr 11 10:19:28 gandalf kernel: Uniform Multi-Platform E-IDE driver > > Revision: 6.31 > > Apr 11 10:19:28 gandalf kernel: ide: Assuming 33MHz system bus speed for > > PIO modes; override with idebus=xx > > Apr 11 10:19:28 gandalf kernel: PDC20246: IDE controller on PCI bus 00 dev > > 70 > > Apr 11 10:19:28 gandalf kernel: PDC20246: chipset revision 1 > > Apr 11 10:19:28 gandalf kernel: PDC20246: not 100%% native mode: will > > probe irqs later > > Apr 11 10:19:28 gandalf kernel: PDC20246: ROM enabled at 0x8081 > > Apr 11 10:19:28 gandalf kernel: PDC20246: (U)DMA Burst Bit DISABLED > > Primary PCI Mode Secondary PCI Mode. > > Apr 11 10:19:28 gandalf kernel: ide2: BM-DMA at 0x0800-0x0807, BIOS > > settings: hde:DMA, hdf:DMA > > Apr 11 10:19:28 gandalf kernel: ide3: BM-DMA at 0x0808-0x080f, BIOS > > settings: hdg:DMA, hdh:DMA > > > > ...however, as you can see, it does not detect the disk itself! > > Actually, I can't see as the disk probing happens later during boot... > > > This is very frustrating. > > > > And yes, I have already checked the connections to the disk. They are > > fine. :) > > Can you send me the complete dmesg log along with the version of the > kernel you are using ? > > Ben. > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > On Sat, 2003-04-12 at 17:14, Jessica Blank wrote: > > Well... okay. This is interesting. > > > > Ben, your kernel patches simply don't like my machine. Maybe I'm doing > > something wrong, but I've configured around a billion trillion bajillion > > kernels in my life (since age 15) and I see nothing missing. > > > > Plain old non-Ben-patched 2.4.20 kernels seem to boot fine on this thing. > > However, your patched kernel seems to shunt itself immediately into > > Runlevel 0 or Runlevel 6, since the last thing I see when I boot it up > > (sorry, the messages scroll by so fast I can't catch them all, and no, > > they DO NOT end up in /var/log/messages nor /var/log/syslog) is > > "Setting the System Clock using the Hardware Clock as reference..." > > ... which is generated by hwclock.sh / hwclockfirst.sh, two startup > > scripts (in /etc/init.d) which are only loaded by runlevels 0 and 6... > > Make sure you have CONFIG_PPC_RTC and _not_ CONFIG_RTC and let me > know. If the kernel tries to access the legacy RTC chip on the ISA > bus, it will hang. Yeah, that's how I did it. Actually, I was currently trying (just for grins and giggles) to compile it with BOTH. (ctrl-Cs that one.. thanks for the warning). Point is, why is the box going to runlevel 0 or 6 in the first place? Those are the only runlevels on my box where it runs the RTC scripts... :/ I am currently trying my umpteenth 2.4.20-ben10 kernel compile, this time with the workaround for 601 bugs (my box has a 601 processor) enabled. However, I doubt that will fix it, since I didn't need to enable it for plain old stock 2.4.20 to boot :/ Is there anything I can provide to you to help? > > > This is not making me happy. :( > > > > Any ideas what I might be doing wrong? > > > > Again, for reference, this machine is a Mac 7200. > > > > Oh, and I even tried removing all the PCI cards, effectively returning the > > Mac to "stock" condition (no IDE, just SCSI; no Tulip NIC, just the > > built-in MESH)... no go. Same results! > > > > Any help? > > > > On Fri, 11 Apr 2003, Benjamin Herrenschmidt wrote: > > > > > On Fri, 2003-04-11 at 16:24, Jessica Blank wrote: > > > > I am having a big problem here. > > > > > > > > My nice little "CD burning server", under my bed, is crippled because > > > > the > > > > large (40GB) IDE drive in it simply does not detect. > > > > > > > > This is a KNOWN WORKING drive. I just pulled it out of a working system. > > > > > > > > I've tried compiling IDE Disk Support and the Promise controller driver > > > > (the two necessary modules, of course) as modules, and I've tried > > > > compiling them straight into the kernel. NEITHER WORKS. > > > > > > > > When I insert ide-mod and ide-disk, I get: > > > > > > > > Apr 11 10:19:28 gandalf kernel: Uniform Multi-Platform E-IDE driver > > > > Revision: 6.31 > > > > Apr 11 10:19:28 gandalf kernel: ide: Assuming 33MHz system bus speed for > > > > PIO modes; override with idebus=xx > > > > Apr 11 10:19:28 gandalf kernel: PDC20246: IDE controller on PCI bus 00 > > > > dev > > > > 70 > > > > Apr 11 10:19:28 gandalf kernel: PDC20246: chipset revision 1 > > > > Apr 11 10:19:28 gandalf kernel: PDC20246: not 100%% native mode: will > > > > probe irqs later > > > > Apr 11 10:19:28 gandalf kernel: PDC20246: ROM enabled at 0x8081 > > > > Apr 11 10:19:28 gandalf kernel: PDC20246: (U)DMA Burst Bit DISABLED > > > > Primary PCI Mode Secondary PCI Mode. > > > > Apr 11 10:19:28 gandalf kernel: ide2: BM-DMA at 0x0800-0x0807, BIOS > > > > settings: hde:DMA, hdf:DMA > > > > Apr 11 10:19:28 gandalf kernel: ide3: BM-DMA at 0x0808-0x080f, BIOS > > > > settings: hdg:DMA, hdh:DMA > > > > > > > > ...however, as you can see, it does not detect the disk itself! > > > > > > Actually, I can't see as the disk probing happens later during boot... > > > > > > > This is very frustrating. > > > > > > > > And yes, I have already checked the connections to the disk. They are > > > > fine. :) > > > > > > Can you send me the complete dmesg log along with the version of the > > > kernel you are using ? > > > > > > Ben. > > > > > > > J e s s i c aL e a hB l a n k > -- > Benjamin Herrenschmidt <[EMAIL PROTECTED]> > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
I was thinking of the init=/bin/sh idea. Will try. On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > On Sat, 2003-04-12 at 17:28, Jessica Blank wrote: > > > > Make sure you have CONFIG_PPC_RTC and _not_ CONFIG_RTC and let me > > > know. If the kernel tries to access the legacy RTC chip on the ISA > > > bus, it will hang. > > > > Yeah, that's how I did it. > > > > Actually, I was currently trying (just for grins and giggles) to compile > > it with BOTH. (ctrl-Cs that one.. thanks for the warning). > > > > Point is, why is the box going to runlevel 0 or 6 in the first place? > > Those are the only runlevels on my box where it runs the RTC scripts... :/ > > I never really bothered about the init scripts, but it always > run that script for me as well when booting to level 2... > > > I am currently trying my umpteenth 2.4.20-ben10 kernel compile, this time > > with the workaround for 601 bugs (my box has a 601 processor) enabled. > > However, I doubt that will fix it, since I didn't need to enable it for > > plain old stock 2.4.20 to boot :/ > > > > Is there anything I can provide to you to help? > > Well, afaik, this kernel works on 7200... so it's difficult to say > remotely what's up. Tried booting with init=/bin/sh to see if you > get a prompt at all ? > > Ben. > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
Interesting notes: init=/bin/sh DOES get me to a shell! And everything works fine. HOWEVER: Booting WITHOUT init=/bin/sh (even with the 601 bug thingy turned on, which I really doubted would solve the problem anyhow), I get the same problem... oh, and for some reason the keyboard is also screwed up. Instead of: qwertyuiop I get: -=\=/ wqdgsh No kidding. The last things I see on the screen when I get the problem are: (And yes I am having to TYPE this stuff by hand, since the system is hard-locked): (At the top, it scans for and finds the hard drive and CD-ROM drive) then: Vendor: SEAGATE Model: ST32171N Rev: 0484 Type: Direct-Access ANSI SCSI revision: 02 Vendor: YAMAHAModel: CRW4416S Rev: 1.0f Type: CD-ROM ANSI SCSI revision: 02 Attached scsi disk sda at scsi0, channel 0, id 1, lun 0 SCSI device sda: 4223444 512-byte hdwr secotrs (2162MB) Partition check: /dev/scsi/host0/bus0/target1/lun0: [mac] p1 p2 p3 p4 p5 p6 Attached scsi CD-ROM sr0 at scsi0, channel 0, id 2, lun 0 sr0: scsi3-mmc drive: 16x/16x writer cd/rw xa/form2 cdda tray Uniform CD-ROM driver Revision: 3.12 dmasound_pmac: Awacs/Screamer Codec Mfct: 1 Rev 2 DMA sound driver: Interrupt initialization failed Macintosh non-volatile memory driver v1.0 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 8192 bind 8192) NET4: Unix domain sockets 1.0/SMP for Linux NET4.0. NET4: Appletalk 0.18a for Linux NET4.0 adb: starting probe task... VFS: Mounted root (ext2 filesystem) readonly. Freeing unused kernel memory: 140k init 8k chrp 8k prep ADB keyboard at 2, handler 1 adb: finished probe task... INIT: version 2.84 booting Loading /etc/console/boottime.kmap.gz Activating swap. EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended. System time was Mon Aug 27 20:42:28 UTC 1956. Setting the System Clock using the Hardware Clock as reference... (and there she hangs) On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > On Sat, 2003-04-12 at 17:28, Jessica Blank wrote: > > > > Make sure you have CONFIG_PPC_RTC and _not_ CONFIG_RTC and let me > > > know. If the kernel tries to access the legacy RTC chip on the ISA > > > bus, it will hang. > > > > Yeah, that's how I did it. > > > > Actually, I was currently trying (just for grins and giggles) to compile > > it with BOTH. (ctrl-Cs that one.. thanks for the warning). > > > > Point is, why is the box going to runlevel 0 or 6 in the first place? > > Those are the only runlevels on my box where it runs the RTC scripts... :/ > > I never really bothered about the init scripts, but it always > run that script for me as well when booting to level 2... > > > I am currently trying my umpteenth 2.4.20-ben10 kernel compile, this time > > with the workaround for 601 bugs (my box has a 601 processor) enabled. > > However, I doubt that will fix it, since I didn't need to enable it for > > plain old stock 2.4.20 to boot :/ > > > > Is there anything I can provide to you to help? > > Well, afaik, this kernel works on 7200... so it's difficult to say > remotely what's up. Tried booting with init=/bin/sh to see if you > get a prompt at all ? > > Ben. > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > On Sat, 2003-04-12 at 19:59, Jessica Blank wrote: > > Interesting notes: > > > > init=/bin/sh DOES get me to a shell! And everything works fine. > > > > HOWEVER: Booting WITHOUT init=/bin/sh (even with the 601 bug thingy turned > > on, which I really doubted would solve the problem anyhow), I get the same > > problem... oh, and for some reason the keyboard is also screwed up. > > Instead of: > > The keyboard problem is a config problem as well. Either you enabled > ADB keycodes and are using linux keycodes, or the opposite. > > ADB keycodes is deprecated now. Eh? Why does the keyboard WORK with init=/bin/sh, but NOT WORK when booting as per usual? That makes zero sense... > > > Setting the System Clock using the Hardware Clock as reference... > > So the problem is really with the real time clock, or with the > Cuda driver. It works fine on the 8500 I have here (approximately > same hardware), though. Can you disable both CONFIG_RTC and > CONFIG_PPC_RTC ? That will prevent sync'ing the clock but should > allow it to boot. Trying right now. Kernel compiles take a while on a 601/90... even with cache... > > Also, is your kernel compile with CONFIG_SMP ? (I beleive it isn't > but still...) > I don't think so. >checks< No. And just in case you were gonna ask that next, YES, I AM compiling with the Machine Type as "CHRP/PowerMac/PReP" ;) And with the Processor Type as "6xx/7xx/64xx/8260" ;) And with AltiVec Support off. And with Thermal Management Support off. An with CPU Frequency scaling off. And with MPC8260 CPM Support off. :> > Ben. > > > (and there she hangs) > > > > On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > > > > > On Sat, 2003-04-12 at 17:28, Jessica Blank wrote: > > > > > > > > Make sure you have CONFIG_PPC_RTC and _not_ CONFIG_RTC and let me > > > > > know. If the kernel tries to access the legacy RTC chip on the ISA > > > > > bus, it will hang. > > > > > > > > Yeah, that's how I did it. > > > > > > > > Actually, I was currently trying (just for grins and giggles) to compile > > > > it with BOTH. (ctrl-Cs that one.. thanks for the warning). > > > > > > > > Point is, why is the box going to runlevel 0 or 6 in the first place? > > > > Those are the only runlevels on my box where it runs the RTC scripts... > > > > :/ > > > > > > I never really bothered about the init scripts, but it always > > > run that script for me as well when booting to level 2... > > > > > > > I am currently trying my umpteenth 2.4.20-ben10 kernel compile, this > > > > time > > > > with the workaround for 601 bugs (my box has a 601 processor) enabled. > > > > However, I doubt that will fix it, since I didn't need to enable it for > > > > plain old stock 2.4.20 to boot :/ > > > > > > > > Is there anything I can provide to you to help? > > > > > > Well, afaik, this kernel works on 7200... so it's difficult to say > > > remotely what's up. Tried booting with init=/bin/sh to see if you > > > get a prompt at all ? > > > > > > Ben. > > > > > > > > > -- > > > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > > > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > > > > > > J e s s i c aL e a hB l a n k > -- > Benjamin Herrenschmidt <[EMAIL PROTECTED]> > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > > > Eh? Why does the keyboard WORK with init=/bin/sh, but NOT WORK when > > booting as per usual? That makes zero sense... > > It does make senes. With init=/bin/sh, the keyboard works with > the kernel built-in keymap. Later on, your userland scripts are > loading a keymap, which is probably designed for the "wrong" > keycode type, and thus produce the incorrect results. Ahh. God bless userland scripts in /etc/*.d screwing things up. :/ > > If you have CONFIG_ADB_KEYCODES set, then either clear it or > boot with keyboard_sends_linux_keycodes=1 on the cmdline (you > can also change it via /proc/sys/dev/mac_hid) > I could find no CONFIG_ADB_KEYCODES. (I rgrep -r -i'd for it). I am presuming you meant CONFIG_ADB_KEYBOARD, and proceeding along those lines... > > > > Setting the System Clock using the Hardware Clock as reference... > > > > > > So the problem is really with the real time clock, or with the > > > Cuda driver. It works fine on the 8500 I have here (approximately > > > same hardware), though. Can you disable both CONFIG_RTC and > > > CONFIG_PPC_RTC ? That will prevent sync'ing the clock but should > > > allow it to boot. > > > > Trying right now. Kernel compiles take a while on a 601/90... even with > > cache... > > Heh, yup > > > > > > > Also, is your kernel compile with CONFIG_SMP ? (I beleive it isn't > > > but still...) > > > > > > > I don't think so. >checks< No. And just in case you were gonna ask that > > next, YES, I AM compiling with the Machine Type as "CHRP/PowerMac/PReP" ;) > > You wouldn't have gone far without this one ;) Hee. I bet. I never tried mussing that one up ;) > > > And with the Processor Type as "6xx/7xx/64xx/8260" ;) And with AltiVec > > Support off. And with Thermal Management Support off. An with CPU > > Frequency scaling off. And with MPC8260 CPM Support off. :> > > Ok, well, those above don't really matter. It really looks like you are > either running CONFIG_RTC instead of CONFIG_PPC_RTC or I broke something > in the Cuda driver on non-SMP (guess what ? my 8500 is SMP :) I'll check > the later case here. > Was ist das "Cuda"? > Ben. > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > > > I could find no CONFIG_ADB_KEYCODES. (I rgrep -r -i'd for it). I am > > presuming you meant CONFIG_ADB_KEYBOARD, and proceeding along those > > lines... > > Yup. > > > > Was ist das "Cuda"? > > Cuda is the "system controller" chip of those old machines, it does, > among other, the real time clock, the poweroff/reboot, some i2c busses, > and the ADB controller. > Ahh. Color me stupid, but what is i2c again? > Ben. > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
Wonkalicious. Now it boots just fine... when I disabled both RTC and PPC_RTC (or RTC_PPC or whatever it was), the thing boots just fine O_o However... after having disabled ADB_KEYBOARD, the keyboard does NOTHING! Like, zilch, zero, zip, nada... I can't even log in. I can, of course, ctrl-openapple-reset ;) I think I'm gonna go dip into the keyboard settings >sigh< Dmesg is attached... gonna test my IDE hardware now. On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > > > I could find no CONFIG_ADB_KEYCODES. (I rgrep -r -i'd for it). I am > > presuming you meant CONFIG_ADB_KEYBOARD, and proceeding along those > > lines... > > Yup. > > > > Was ist das "Cuda"? > > Cuda is the "system controller" chip of those old machines, it does, > among other, the real time clock, the poweroff/reboot, some i2c busses, > and the ADB controller. > > Ben. > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
Err. Ignore the bit about the dmesg. I gotta get my keyboard working first ;) On Sat, 12 Apr 2003, Jessica Blank wrote: > Wonkalicious. Now it boots just fine... when I disabled both RTC and > PPC_RTC (or RTC_PPC or whatever it was), the thing boots just fine O_o > > However... after having disabled ADB_KEYBOARD, the keyboard does NOTHING! > Like, zilch, zero, zip, nada... I can't even log in. I can, of course, > ctrl-openapple-reset ;) > > > I think I'm gonna go dip into the keyboard settings >sigh< > > Dmesg is attached... gonna test my IDE hardware now. > > On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > > > > > > I could find no CONFIG_ADB_KEYCODES. (I rgrep -r -i'd for it). I am > > > presuming you meant CONFIG_ADB_KEYBOARD, and proceeding along those > > > lines... > > > > Yup. > > > > > > > Was ist das "Cuda"? > > > > Cuda is the "system controller" chip of those old machines, it does, > > among other, the real time clock, the poweroff/reboot, some i2c busses, > > and the ADB controller. > > > > Ben. > > > > > > -- > > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > > > J e s s i c aL e a hB l a n k > J e s s i c aL e a hB l a n k
Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
Now, all of this begs the question... since my Mac started working again after I disabled BOTH RTC settings... yet previously, I'd been using 2.4 kernels (non-Ben-patched/virgin/stock/unaltered/Linus's-tree/whatever-you-call-it) with the PPC RTC option on, and it booted fine... ...what did you do to alter it, Ben? ^_^ I would be curious.. can you look over the diff(s) of the relevent file(s) and see what might be causing this? Also, I am not quite sure if your 8500 and my 7200 have "approximately the same hardware" as you mentioned... yes, they're both Old World, and yes, they share the same complement of SWIM/MESH/etc. chips, but if I recall correctly the 7200 (and 8200 too, I think?) had some hardware wonkiness which made them somewhat different from the other Macs... no? At a bare minimum, they lack the neat CPU upgrade slot which most of the other PCI OldWorld Macs had... hence forcing the makers of Mac CPU upgrades to release a special 7200/8200-only module, or release none at all... Or is that the only notable hardware difference between the 7200s and the other 7x00/8x00 machines? On Sat, 12 Apr 2003, Jessica Blank wrote: > Wonkalicious. Now it boots just fine... when I disabled both RTC and > PPC_RTC (or RTC_PPC or whatever it was), the thing boots just fine O_o > > However... after having disabled ADB_KEYBOARD, the keyboard does NOTHING! > Like, zilch, zero, zip, nada... I can't even log in. I can, of course, > ctrl-openapple-reset ;) > > > I think I'm gonna go dip into the keyboard settings >sigh< > > Dmesg is attached... gonna test my IDE hardware now. > > On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > > > > > > I could find no CONFIG_ADB_KEYCODES. (I rgrep -r -i'd for it). I am > > > presuming you meant CONFIG_ADB_KEYBOARD, and proceeding along those > > > lines... > > > > Yup. > > > > > > > Was ist das "Cuda"? > > > > Cuda is the "system controller" chip of those old machines, it does, > > among other, the real time clock, the poweroff/reboot, some i2c busses, > > and the ADB controller. > > > > Ben. > > > > > > -- > > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > > > J e s s i c aL e a hB l a n k > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > J e s s i c aL e a hB l a n k
SUCCESS!! Was Re: Mac 7200 (PPC) sees Promise controller.. BUT NO DRIVES
My box sees the drives now!! Now, why can't I get RTC to work? On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > On Sat, 2003-04-12 at 19:59, Jessica Blank wrote: > > Interesting notes: > > > > init=/bin/sh DOES get me to a shell! And everything works fine. > > > > HOWEVER: Booting WITHOUT init=/bin/sh (even with the 601 bug thingy turned > > on, which I really doubted would solve the problem anyhow), I get the same > > problem... oh, and for some reason the keyboard is also screwed up. > > Instead of: > > The keyboard problem is a config problem as well. Either you enabled > ADB keycodes and are using linux keycodes, or the opposite. > > ADB keycodes is deprecated now. > > > Setting the System Clock using the Hardware Clock as reference... > > So the problem is really with the real time clock, or with the > Cuda driver. It works fine on the 8500 I have here (approximately > same hardware), though. Can you disable both CONFIG_RTC and > CONFIG_PPC_RTC ? That will prevent sync'ing the clock but should > allow it to boot. > > Also, is your kernel compile with CONFIG_SMP ? (I beleive it isn't > but still...) > > Ben. > > > (and there she hangs) > > > > On Sat, 12 Apr 2003, Benjamin Herrenschmidt wrote: > > > > > On Sat, 2003-04-12 at 17:28, Jessica Blank wrote: > > > > > > > > Make sure you have CONFIG_PPC_RTC and _not_ CONFIG_RTC and let me > > > > > know. If the kernel tries to access the legacy RTC chip on the ISA > > > > > bus, it will hang. > > > > > > > > Yeah, that's how I did it. > > > > > > > > Actually, I was currently trying (just for grins and giggles) to compile > > > > it with BOTH. (ctrl-Cs that one.. thanks for the warning). > > > > > > > > Point is, why is the box going to runlevel 0 or 6 in the first place? > > > > Those are the only runlevels on my box where it runs the RTC scripts... > > > > :/ > > > > > > I never really bothered about the init scripts, but it always > > > run that script for me as well when booting to level 2... > > > > > > > I am currently trying my umpteenth 2.4.20-ben10 kernel compile, this > > > > time > > > > with the workaround for 601 bugs (my box has a 601 processor) enabled. > > > > However, I doubt that will fix it, since I didn't need to enable it for > > > > plain old stock 2.4.20 to boot :/ > > > > > > > > Is there anything I can provide to you to help? > > > > > > Well, afaik, this kernel works on 7200... so it's difficult to say > > > remotely what's up. Tried booting with init=/bin/sh to see if you > > > get a prompt at all ? > > > > > > Ben. > > > > > > > > > -- > > > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > > > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > > > > > > J e s s i c aL e a hB l a n k > -- > Benjamin Herrenschmidt <[EMAIL PROTECTED]> > J e s s i c aL e a hB l a n k
