Help! :-)

I'm trying to upgrade my player to a 128gb SSD (the KingSpec unit that MLord recommended). I'm trying to copy the FIDs from drive1 (the old drive) to drive0 (the new drive), and it ran for hours just fine, but now it's displaying error messages in the copy operation.

Here's what I did, and below that, the output:

- Installed the new SSD drive by itself, with the jumpers set to Master.
- Installed builder_bigdisk_v6.upgrade, freshly downloaded from http://rtr.ca/bigdisk on that drive, seemed to go OK, but I didn't get to the terminal prompt quickly enough to see if it was prompting me with a "disk was already built" message after the installation. I'm not clear on whether we need to press enter or not at that point any more.
- Installed car2_v2b13_hijack.upgrade, freshly downloaded from http://rtr.ca/bigdisk on that drive, it was successful and the empeg booted into the player software as expected.
- Installed the old drive back on the hard disk cable with its jumper in the slave position.
- At the serial terminal prompt:

Q
cd /
df drive0
df drive1

(confirmed that drive0 looked like the new 128gb SSD drive and drive1 looked like the original 40gb Fujitsu drive)

rw
rwm
cp -auvfx /drive1/fids /drive0/

It copied FIDs for several hours succesfully, and was still copying when I got back from going out to dinner. This is the current screen output (I have removed hundreds of lines of successful FID copying from the scrollback history):

(...)
/drive1/fids/b571 -> /drive0/fids/b571
/drive1/fids/b580 -> /drive0/fids/b580
/drive1/fids/b581 -> /drive0/fids/b581
/drive1/fids/b590 -> /drive0/fids/b590
/drive1/fids/b591 -> /drive0/fids/b591
/drive1/fids/b5a0 -> /drive0/fids/b5a0
/drive1/fids/b5a1 -> /drive0/fids/b5a1
/drive1/fids/b5b0 -> /drive0/fids/b5b0
/drive1/fids/b5b1 -> /drive0/fids/b5b1
/drive1/fids/b5c0 -> /drive0/fids/b5c0
/drive1/fids/b5c1 -> /drive0/fids/b5c1
/drive1/fids/b5d0 -> /drive0/fids/b5d0
/drive1/fids/b5d1 -> /drive0/fids/b5d1
/drive1/fids/b5e0 -> /drive0/fids/b5e0
hda: status timeout: status=0xd0
hda: no DRQ after issuing WRITE
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
hda: drive not ready for command
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
end_request: I/O error, dev 03:04 (hda), sector 0
hda: drive not ready for command
hda: status timeout: status=0xd0
hda: drive not ready for command
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
hda: drive not ready for command
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
end_request: I/O error, dev 03:04 (hda), sector 8
hda: drive not ready for command
hda: status timeout: status=0xd0
hda: drive not ready for command
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
hda: drive not ready for command
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
end_request: I/O error, dev 03:04 (hda), sector 1048672
hda: drive not ready for command
hda: status timeout: status=0xd0
hda: drive not ready for command
EXT2-fs error (device ide0(3,4)): ext2_write_inode: unable to read inode block - inode=2108, block=131084
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
hda: drive not ready for command
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
end_request: I/O error, dev 03:04 (hda), sector 27000832
hda: drive not ready for command
hda: status timeout: status=0xd0
hda: drive not ready for command
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
hda: drive not ready for command
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0
end_request: I/O error, dev 03:04 (hda), sector 27014384
hda: drive not ready for command
hda: status timeout: status=0xd0
hda: drive not ready for command
ide0: reset timed-out, status=0xd0
hda: status timeout: status=0xd0


Unfortunately I don't have scrollback of the beginning of the terminal session which would show the boot sequence, seems Putty doesn't keep everything in the scrollback. Ah well.

I'm expecting that I should terminate the copy operation, and do a "ro" and "rom", to protect the contents of the original drive before I try to start over again. But before I start over again, does anyone have any other advice?

Thanks!

Also, ctrl-C isn't working for me right now, and going into the Putty menu and saying "Special command: Break" is also not working.

Neither is ctrl-X, ctrl-Z, or ctrl-d...

I managed to get the terminal to "unlock" and go back to the shell by unplugging the new drive and plugging it back in. Then it let me type RO and ROM. The old drive seems to still be working fine. The new one, I'm going to try a reformat again.

Okay, on the second try of running the builder image on the new SSD drive, I was able to get the terminal window up at the very beginning of the process, and this is what it recorded:


it seems not. Let fly the Penguins of Linux!

e000 v1.04
Copying kernel...
Calling linux kernel...
Uncompressing Linux..................................... done, booting the kernel.
Linux version 2.2.17-rmk5-np17-empeg55-hijack-v508 ([email protected]) (gcc version 2.95.3 20010315 (release)) #2 Fri Jan 9 16:06:35 EST 2009
Processor: Intel StrongARM-1100 revision 11
Checking for extra DRAM:
c0c00000: wrote ffffffff, read 00000003
NetWinder Floating Point Emulator V0.94.1 (c) 1998 Corel Computer Corp.
empeg-car player (hardware revision 7, serial number 60000013) 12MB DRAM
Command line: mem=12m
Calibrating delay loop... 207.67 BogoMIPS
Memory: 10944k/12M available (996k code, 20k reserved, 324k data, 4k init)
Dentry hash table entries: 2048 (order 2, 16k)
Buffer cache hash table entries: 16384 (order 4, 64k)
Page cache hash table entries: 4096 (order 2, 16k)
POSIX conformance testing by UNIFIX
Linux NET4.0 for Linux 2.2
Based upon Swansea University Computer Society NET3.039
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP
TCP: Hash tables configured (ehash 16384 bhash 16384)
IrDA (tm) Protocols for Linux-2.2 (Dag Brattli)
Starting kswapd v 1.5
SA1100 serial driver version 4.27 with no serial options enabled
ttyS00 at 0xf8010000 (irq = 15) is a SA1100 UART
ttyS01 at 0xf8050000 (irq = 17) is a SA1100 UART
ttyS02 at 0xf8030000 (irq = 16) is a SA1100 UART
Signature is 67706d65 'empg'
Found custom animation at offset 0x9bb80
Tuner: loopback=0, ID=0
show_message("Hijack v508 by Mark Lord")
empeg display initialised.
empeg dsp audio initialised
empeg dsp mixer initialised
empeg dsp initialised
empeg audio-in initialised, CS4231A revision a0
empeg remote control/panel button initialised.
empeg usb initialised, PDIUSBD12 id 1012
empeg state support initialised 0089/88c1 (save to d0005900).
empeg RDS driver initialised
empeg power-pic driver initialised
RAM disk driver initialized: 16 RAM disks of 4096K size
empeg single channel IDE
Probing primary interface...
ide_data_test: wrote 0x0000 read 0x0000
ide_data_test: wrote 0xffff read 0x0000
ide_data_test: wrote 0xaaaa read 0x0000
ide_data_test: wrote 0x5555 read 0x0000
hda: KSD-PA25.6-128MS, ATA DISK drive
ide_data_test: wrote 0x0000 read 0x044a
ide_data_test: wrote 0xffff read 0x3fff
ide_data_test: wrote 0xaaaa read 0x044a
ide_data_test: wrote 0x5555 read 0x3fff
hda: KSD-PA25.6-128MS, ATA DISK drive
ide_data_test: wrote 0x0000 read 0x044a
ide_data_test: wrote 0xffff read 0x3fff
ide_data_test: wrote 0xaaaa read 0x044a
ide_data_test: wrote 0x5555 read 0x3fff
hda: KSD-PA25.6-128MS, ATA DISK drive
ide_data_test: wrote 0x0000 read 0x044a
ide_data_test: wrote 0xffff read 0x3fff
ide_data_test: wrote 0xaaaa read 0x044a
ide_data_test: wrote 0x5555 read 0x3fff
hda: KSD-PA25.6-128MS, ATA DISK drive
ide_data_test: wrote 0x0000 read 0x044a
ide_data_test: wrote 0xffff read 0x3fff
ide_data_test: wrote 0xaaaa read 0x044a
ide_data_test: wrote 0x5555 read 0x3fff
hda: KSD-PA25.6-128MS, ATA DISK drive
ide_data_test: wrote 0x0000 read 0x044a
ide_data_test: wrote 0xffff read 0x3fff
ide_data_test: wrote 0xaaaa read 0x044a
ide_data_test: wrote 0x5555 read 0x3fff
hda: KSD-PA25.6-128MS, ATA DISK drive
ide0 at 0x000-0x007,0x038 on irq 6
hda: KSD-PA25.6-128MS, 121082MB w/1kB Cache, CHS=15435/255/63
empeg-flash driver initialized
smc chip id/revision 0x3349
smc9194.c:v0.12 03/06/96 by Erik Stahlman ([email protected])

SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC 00:02:d7:0c:00:0d
Partition check:
hda: hda1 < hda5 hda6 > hda2 hda3 hda4
RAMDISK: ext2 filesystem found at block 0
RAMDISK: Loading 320 blocks [1 disk] into ram disk... done.
VFS: Mounted root (ext2 filesystem).
empeg-pump v0.03 (19980601)
Press Ctrl-A to enter pump...VFS: Mounted root (ext2 filesystem) readonly.
change_root: old root has d_count=1
Trying to unmount old root ... okay
Freeing unused kernel memory: 4k initBuilder image
show_message("Builder image")
Setting up swapspace version 0, size = 16412672 bytes
Adding Swap: 16028k swap-space (priority -1)
Setting up swapspace version 0, size = 24670208 bytes
Adding Swap: 24092k swap-space (priority -2)
Tuning hda5
show_message("Tuning hda5")
tune2fs 1.17, 26-Oct-1999 for EXT2 FS 0.5b, 95/08/09
Setting maximal mount count to -1
Setting interval between check 0 seconds
Making hda4 filesysteshow_message("Making hda4 filesystem")
m
mke2fs 1.14, 9-Jan-1999 for EXT2 FS 0.5b, 95/08/09
ext2fs_check_if_mount: No such file or directory while determining whether /dev/hda4 is mounted.
Linux ext2 filesystem format
Filesystem label=
484352 inodes, 30969303 blocks
0 blocks (0.00%) reserved for the super user
First data block=0
Block size=4096 (log=2)
Fragment size=4096 (log=2)
946 block groups
32768 blocks per group, 32768 fragments per group
512 inodes per group
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
4096000, 7962624, 11239424, 20480000, 23887872,

Writing inode tables: done
Writing superblocks and filesystem accounting information: done
Tuning hda4
show_message("Tuning hda4")
tune2fs 1.17, 26-Oct-1999 for EXT2 FS 0.5b, 95/08/09
Setting maximal mount count to -1
Setting interval between check 0 seconds
Mounting hda4
show_message("Mounting hda4")
Making directories
show_message("Making directories")
Remounting ro
show_message("Remounting ro")
hda4 completed
show_message("hda4 completed")
Zeroing hda3
show_message("Zeroing hda3")
/bin/cat: write error: No space left on device
Done!
show_message("Done!")
bash-2.01#

That last set of messages look good.

I'm not sure how it managed to get itself into a huff mid-copy there, but the whole thing would be more reliable and a lot faster if you could just hook the two drives up to a PC running Linux (bootable CD/USB) and do the copy from there. USB connections are fine for this.

Even the one near the end about "/bin/cat: write error: No space left on device"? If you say that's OK too, then I'll try again this weekend.

From the messages, it looks like the issue was entirely on the SSD side of things, correct? The HDA and IDE0 messages seem to indicate that.



I might try that if I can't get it to work from within the player. I'd have to find and purchase some IDE-to-USB converters though. Right now my only external disk connectors are SATA.

Yes, nothing wrong at all about the "No space" message. The builder attempts to write an infinite amount of zeros to the partition, and unfortunately the partition isn't large enough to hold them all.

So does the infinite zeroes thing also constitute a write test, indicating that the partition is in working order?

Thanks so much for all your help.

I suppose it does, more or less. Not the intent, but a side-effect.

Should you ever feel the need for such, the UGREEN USB IDE Adapter (USB 3.0) sold by Amazon is excellent for the task. One can also get by nicely with sub-$10 IDE/USB enclosures sold by dx.com.

Cheers

Performing a retry (with the master/slave order reversed from my prior attempt), I'm getting this occasionally which is different from before:

/drive0/fids/b2e1 -> /drive1/fids/b2e1
/drive0/fids/b2f0 -> /drive1/fids/b2f0
/drive0/fids/b2f1 -> /drive1/fids/b2f1
/drive0/fids/b300 -> /drive1/fids/b300
/drive0/fids/b301 -> /drive1/fids/b301
/drive0/fids/b310 -> /drive1/fids/b310
hdb: status timeout: status=0xd0
hdb: no DRQ after issuing WRITE
ide0: reset: success
/drive0/fids/b311 -> /drive1/fids/b311
/drive0/fids/b320 -> /drive1/fids/b320
/drive0/fids/b321 -> /drive1/fids/b321
/drive0/fids/b330 -> /drive1/fids/b330
/drive0/fids/b331 -> /drive1/fids/b331
/drive0/fids/b340 -> /drive1/fids/b340
/drive0/fids/b301 -> /drive1/fids/b301
/drive0/fids/b310 -> /drive1/fids/b310
hdb: status timeout: status=0xd0
hdb: no DRQ after issuing WRITE
ttyS1: 1 input overrun(s)
ide0: reset:uccess
/drive0/fids/b311 -> /drive1/fids/b311
/drive0/fids/b320 -> /drive1/fids/b320
/drive0/fids/b321 -> /drive1/fids/b321
/drive0/fids/b341 -> /drive1/fids/b341
/drive0/fids/b350 -> /drive1/fids/b350

That's all okay. Not perfect, but no data loss or anything.
The "no DRQ after issuing WRITE" is something that has been observed with other SSDs on this kernel too. I might give you something to beef up the debug info there so we can nail/fix that for good.

Curiously, I cannot find anywhere in the kernel that might print this exact message from your log above:

hdb: status timeout: status=0xd0

The message looks like it should instead read thusly:

hdb: stat_timo: status=0xd0

Did you hand copy the messages, or cut and paste?
Or am I more old and feeble than I feel?

Okay, Hijack v521 is now released. Changes:

• Fixed Hijack version number (v520 claimed to be "v519").
• Increase DRQ timeout from 5 to 10 seconds.
• Read IDE status register after writing device control reg.

EDIT: Actually, I think I spotted a different IDE protocol issue.. having a look now to see if it is real or imagined.. false alarm.

Okay Tony: Here (attached) is a pre-release version of hijack, with an extra tweak included. Dunno if this will make a difference here or not, but it doesn't seem to break anything on the SSD I'm using here (64GB version of the same brand/make).

All I've done is add hardware memory barrier instructions into the IDE command issuance code. Modern kernels do this everywhere, but the need for such had not been determined back in the empeg days. These prevent register writes/reads from getting performed out of the intended sequence by the compiler and/or CPU.

The copy operation seems to have completed overnight without the issues from before. But Putty doesn't keep the full scrollback so I'm not 100 percent certain that it really worked, but I'm going to move forward as if it had worked. At least the tail end of the output looked correct the next morning:

/drive0/fids/18080 -> /drive1/fids/18080
/drive0/fids/18081 -> /drive1/fids/18081
/drive0/fids/180a0 -> /drive1/fids/180a0
/drive0/fids/_00018 -> /drive1/fids/_00018
/drive0/fids/_00018/090 -> /drive1/fids/_00018/090
/drive0/fids/_00018/091 -> /drive1/fids/_00018/091
/drive0/fids/_00018/0a1 -> /drive1/fids/_00018/0a1
/drive0/fids/_00018/0b0 -> /drive1/fids/_00018/0b0
/drive0/fids/_00018/0b1 -> /drive1/fids/_00018/0b1
/drive0/fids/_00018/0c0 -> /drive1/fids/_00018/0c0
/drive0/fids/_00018/0c1 -> /drive1/fids/_00018/0c1
/drive0/fids/_00018/0d0 -> /drive1/fids/_00018/0d0
/drive0/fids/_00018/0d1 -> /drive1/fids/_00018/0d1
/drive0/fids/_00018/0e0 -> /drive1/fids/_00018/0e0
/drive0/fids/_00018/0e1 -> /drive1/fids/_00018/0e1
/drive0/fids/_00018/100 -> /drive1/fids/_00018/100
empeg:/empeg/bin#

I'll also install that pre-release Hijack.

Copied directly from Putty.

So, though we're both getting old and feeble, I believe it's unrelated to what was in the messages.

Thanks for all your help, Mark! And thanks for the new kernel!

After remounting the drives RO/ROM and loading up Emplode, it tries to rebuild its databases and gives me a warning saying that the length of playlist "Out of Time" does not match what was expected.

When I look at that playlist in Emplode (it is the R.E.M. album), I see that it has three fewer songs now than what I would have expected. Looking at the FID numbers in Emplode for the song files, I see that we're indeed in the range of numbers that got the Status Timeout errors:

-- 0xb2a0 Playlist: "Out of Time"
01 0xb2d0 Radio Song
02 0xb2e0 Losing my Religion
03 0xb2f0 Low

xx xxxxxx MISSING: Near Wild Heaven
xx xxxxxx MISSING: Endgame
xx xxxxxx MISSING: Shiny Happy People

07 0xb330 Belong
08 0xb340 Half a World Away
09 0xb350 Texarkana
10 0xb360 Country Feedback
11 0xb370 Me in Honey

I'm probably going to just delete and re-upload the whole album from scratch, just to be sure, and then that should take care of it.

Thanks again!

I think I've got everything copied and it's all working. Thanks again for your help!

One last question

I have a note here which (I think) is telling me that I have a "fixed" copy of the "escher" visual that won't crash the player.

I want to copy that file from the old drive, currently mounted as master drive0, to the slave drive, currently mounted as slave drive1. The slave drive has already had the 2.0 Beta 13 player software installed onto it from an .upgrade file.

The file I want to copy lives in empeg/lib/visuals/escher.raw. I want to copy it from the old drive to the new drive, but I can't see any folder named "empeg" as a subdirectory of "drive1" anywhere. I can type cd /empeg and it gets me to what I suppose is the player software folder on drive0, but I don't know how to get to the corresponding location on drive1.

Clearly I'm not understanding the drive/directory structure on the player despite years of messing with it.

I'm pretty sure that file resides on the "root filesystem" (/), rather than on the fids filesystems (/drive0, /drive1).

So it should just exist at /empeg/lib/visuals/escher.raw on the player, and can be overwritten by first doing "rw" to make the partition writable, followed by "ro" after copying over the new file.

.. so if you have both drives connected to the empeg, with the old drive as "/dev/sdb", you can do this:

I'm assuming it wants a "-t ext2" on the command line somewhere there?

Nope, it says:

And I'm getting more error messages trying to back my way out of this situation:

Okay, what does this show:

cat /proc/partitions

If /dev/sdb5 (or /dev/sdc5 perhaps) does appear in /proc/partitions then just modify that line as follows:

mount -t ext2 /dev/sdb5 /t

The mount command on the empeg seems more than overly fussy about the order of parameters..

Thanks so much for the help, Mark!

I'll try those things when I get home tonight.

I also realized that I could have skipped all of this complexity by merely doing the following:
- Attach the old drive by itself as master.
- On the old drive, copy the desired file into drive0/var
- Attach the new drive as master and the old drive as slave
- Copy the desired file from drive1/var

I'm assuming that proc/partitions is the file that controls which drives and partitions are mounted at startup, correct? If so, and if it's pointing at a nonexistent partition, what is the best direct way to repair the file? From my scrollback, my attempts to do it with "umount /t" failed.

Thanks again!

With the new drive as Master and the old drive as Slave, it says:

Hm. Looking at the old hard drive (connected as Master) maybe I'm wrong about that file. Maybe it wasn't a fixed copy of the file, but rather, a customized version of the file that I don't need any more. If that's the case, I don't need to make the copy. And now, I'm not even seeing the error message "mount: can't find /drive1 in /etc/fstab or /etc/mtab" any more, so I think I'm probably done here actually with nothing left to do.

Thanks again for all your help! We'll see how this SSD does in the long run.

Holy wow database rebuilds and drive syncs are fast on that SSD.

Yeah, sync/rebuilds are all about seeking around through thousands of tiny little tags files. And SSDs take effectively zero time for seeks! Compare that to mechanical drives, where seeking is what they spend almost all of their time doing.

The /proc/partitions thing just shows the Linux kernel's view of what drives and partitions it knows about in the system -- those that actually exist. It doesn't mount them or anything, that's up to the system scripts or the end user.

Cheers