From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows 98) Description of problem: IDE tape drive inside the RedHat kernels 2.4.9-6 and 2.4.9-21 does not work properly. It does work under RedHat kernel 2.4.7-10. System is Intel Pentium III 1.0GHz in Intel I815EGEW mainboard using primary IDE controller with CD-ROM as master and TAPE as slave drive using 80pin IDE cable. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. boot up 2.4.9 kernel 2. start CPIO or TAR to write data to tape 3. watch it not work. Actual Results: Errors listed in /var/adm/messages. Initially the drive status LED will remain on while the system is timing out. After the timeout the drive will start to rewind then report "end of medium found. insert next tape." Expected Results: Create tape with hard drive contents on it. Additional info: Feb 25 07:25:23 beckman kernel: ide-tape: Dumping ATAPI Identify Device tape parameters Feb 25 07:25:23 beckman kernel: ide-tape: Protocol Type: <6>ATAPI Feb 25 07:25:23 beckman kernel: ide-tape: Device Type: 1 - <6>Streaming Tape Device Feb 25 07:25:23 beckman kernel: ide-tape: Removable: Yes Feb 25 07:25:23 beckman kernel: ide-tape: Command Packet DRQ Type: <6>Accelerated DRQ Feb 25 07:25:23 beckman kernel: ide-tape: Command Packet Size: <6>12 bytes Feb 25 07:25:23 beckman kernel: ide-tape: Model: Seagate STT20000A Feb 25 07:25:23 beckman kernel: ide-tape: Firmware Revision: 8A51 Feb 25 07:25:23 beckman kernel: ide-tape: Serial Number: Feb 25 07:25:23 beckman kernel: ide-tape: Write buffer size: 372736 bytes Feb 25 07:25:23 beckman kernel: ide-tape: DMA: Yes Feb 25 07:25:23 beckman kernel: ide-tape: LBA: Yes Feb 25 07:25:23 beckman kernel: ide-tape: IORDY can be disabled: Yes Feb 25 07:25:23 beckman kernel: ide-tape: IORDY supported: Yes Feb 25 07:25:23 beckman kernel: ide-tape: ATAPI overlap supported: No Feb 25 07:25:23 beckman kernel: ide-tape: PIO Cycle Timing Category: 2 Feb 25 07:25:23 beckman kernel: ide-tape: DMA Cycle Timing Category: 2 Feb 25 07:25:23 beckman kernel: ide-tape: Single Word DMA supported modes: <6>0 <6>1 <6>2 <6>(active) <6> Feb 25 07:25:23 beckman kernel: ide-tape: Multi Word DMA supported modes: <6>0 <6>1 <6>2 <6>(active) <6> Feb 25 07:25:23 beckman kernel: ide-tape: Enhanced PIO Modes: Mode 3 Feb 25 07:25:23 beckman kernel: ide-tape: Minimum Multi-word DMA cycle per word: <6>120 ns Feb 25 07:25:23 beckman kernel: ide-tape: Manufacturer's Recommended Multi-word cycle: <6>120 ns Feb 25 07:25:23 beckman kernel: ide-tape: Minimum PIO cycle without IORDY: <6>120 ns Feb 25 07:25:23 beckman kernel: ide-tape: Minimum PIO cycle with IORDY: <6>120 ns Feb 25 07:25:23 beckman kernel: ide-tape: hdb <-> ht0: Seagate STT20000A rev 8A51 Feb 25 07:25:23 beckman kernel: ide-tape: Dumping the results of the MODE SENSE packet command Feb 25 07:25:23 beckman kernel:ide-tape: Mode Parameter Header: Feb 25 07:25:23 beckman kernel: ide-tape: Mode Data Length - 23 Feb 25 07:25:23 beckman kernel: ide-tape: Medium Type - 183 Feb 25 07:25:23 beckman kernel: ide-tape: Device Specific Parameter - 16 Feb 25 07:25:23 beckman kernel: ide-tape: Block Descriptor Length - 0 Feb 25 07:25:23 beckman kernel: ide-tape: Capabilities and Mechanical Status Page: Feb 25 07:25:23 beckman kernel: ide-tape: Page code - 42 Feb 25 07:25:23 beckman kernel: ide-tape: Page length - 18 Feb 25 07:25:23 beckman kernel: ide-tape: Read only - No Feb 25 07:25:23 beckman kernel: ide-tape: Supports reverse space - Yes Feb 25 07:25:23 beckman kernel: ide-tape: Supports erase initiated formatting - No Feb 25 07:25:23 beckman kernel: ide-tape: Supports QFA two Partition format - Yes Feb 25 07:25:23 beckman kernel: ide-tape: Supports locking the medium - No Feb 25 07:25:23 beckman kernel: ide-tape: The volume is currently locked - No Feb 25 07:25:23 beckman kernel: ide-tape: The device defaults in the prevent state - No Feb 25 07:25:23 beckman kernel: ide-tape: Supports ejecting the medium - No Feb 25 07:25:23 beckman kernel: ide-tape: Supports error correction - Yes Feb 25 07:25:23 beckman kernel: ide-tape: Supports data compression - No Feb 25 07:25:23 beckman kernel: ide-tape: Supports 512 bytes block size - Yes Feb 25 07:25:23 beckman kernel: ide-tape: Supports 1024 bytes block size - No Feb 25 07:25:23 beckman kernel: ide-tape: Supports 32768 bytes block size / Restricted byte count for PIO transfers - No Feb 25 07:25:23 beckman kernel: ide-tape: Maximum supported speed in KBps - 1000 Feb 25 07:25:23 beckman kernel: ide-tape: Continuous transfer limits in blocks - 108 Feb 25 07:25:23 beckman kernel: ide-tape: Current speed in KBps - 1000 Feb 25 07:25:23 beckman kernel: ide-tape: Buffer size - 331776 Feb 25 07:25:23 beckman kernel: ide-tape: Adjusted block size - 512 Feb 25 07:25:23 beckman kernel: ide-tape: hdb <-> ht0: 1000KBps, 6*54kB buffer, 9720kB pipeline, 110ms tDSC, DMA Feb 25 07:25:23 beckman kernel: ide-tape: Reached idetape_chrdev_open Feb 25 07:26:29 beckman kernel: hdb: timeout waiting for DMA Feb 25 07:26:29 beckman kernel: ide_dmaproc: chipset supported ide_dma_timeout func only: 14 Feb 25 07:26:29 beckman kernel: hdb: status timeout: status=0xd0 { Busy } Feb 25 07:26:29 beckman kernel: hdb: drive not ready for command Feb 25 07:26:29 beckman kernel: hdb: ATAPI reset complete Feb 25 07:26:29 beckman kernel: ide-tape: ht0: I/O error, pc = a, key = 2, asc = 4, ascq = 1 Feb 25 07:26:32 beckman last message repeated 180 times Feb 25 07:26:32 beckman kernel: ide-tape: ht0: I/O error, pc = 10, key = 2, asc = 4, ascq = 1 Feb 25 07:26:32 beckman kernel: ide-tape: ht0: I/O error, pc = 1b, key = 2, asc = 3a, ascq = 0 Feb 25 07:26:32 beckman kernel: ide-tape: ht0: I/O error, pc = 1, key = 2, asc = 4, ascq = 1
Found out that if you add "ide=nodma" to the kernel line, the 2.4.9 kernels work correctly.
The "ide=nodma" is a little too harsh, because it kills DMA on IDE disks too. A better workaround is to use hdparm -d0 /dev/hdb. This is an ongoing problem somewhere in bowels of IDE framework. Apparently, the old 2.2 IDE did not hit hw too hard. Last note - I suggest to use ide-scsi and st instead of ide-tape. The latter is a poorly implemented driver, with bugs crawling all over it. *** This bug has been marked as a duplicate of 54517 ***