Messages in this thread | | | Date | Mon, 11 Jan 1999 08:40:14 +0200 (IST) | From | Gadi Oxman <> | Subject | Re: ide-tape/ide-card broken in 2.2pre3 and greater |
| |
On Sun, 10 Jan 1999, Andre M. Hedrick wrote:
> Follow-up:: > > The drive that failed resulted in a file_mark_error, but the larger > drive has been successfull with repeatable backups. After checking the > media back in dos, there was a bad-block. An attempted re-format with > CBD 4.X on the Colorado 5GB failed and wiped the media to unusable. > > Second note: > > #if 0 > drive->dsc_overlap = 1; > #else > drive->dsc_overlap = (drive->id->capability & 0x20) ? 1 : 0; > #endif > > We have discussed this issue in the past. > The test maybe wrong, but the result is success; therefore, since > dsc_overlap does not appear to behave nicely (based on imperical tests). > > Would 'drive->dsc_overlap = 0;' seem practable? I still have not read all > of the referrence material you have suggested.
First we disable DMA by default on IDE chipsets back in August, and now we think of disabling dsc_overlap.. which will just take away most of my work on optimizing the ide-tape driver for no good reason.
That's not the way to solve problems. We must first diagnose and fully understand the real cause of a problem. Only then we might try to fix it.
Gadi
> Other test observed that I can not setup xfer_rates at init time on > a new chipset that is non-bootable. I get a status value of 0x01 > when using Dave Millers set_xfer stripped from cmd646.c with > 'drive->dsc_overlap = 1;' and the system (SMP) hangs hard; however, > with it set to 0, I have full success in tuning the tape drive. > We get a series of stat values 0xd0 then 0x50 between programming > commands. > > I know that what DM has written works nicely, since I can enable > DMA mode 2 via chipset tuning on both tape drives, the driver does > acknowledge the fact, and hdparm does correctly report the dma modes. > > On Sat, 9 Jan 1999, Andre M. Hedrick wrote: > > > Gadi, > > > > I looked at what you did with ide-floppy and repeated with > > > > /* > > * Some tape drives require a long irq timeout > > */ > > #if 0 > > #define IDETAPE_WAIT_CMD 60 > > #else > > #define IDETAPE_WAIT_CMD (5 * WAIT_CMD) > > #endif > > > > > > and I have two tapes under SMP streaming, opps, maybe one now. > > > > What is the preferred wait timeout? > > Cheers, > Andre Hedrick > The (NEW) Linux IDE guy > The APC UPS Specialist for Linux > > http://www.dyer.vanderbilt.edu/server/udma/ > http://www.dyer.vanderbilt.edu/server/apcupsd/apcupsd-3.5.2.bin.tar.gz > > You just need a bigger hammer, or learn how to swing the one you have better. > (C) me..... > > > > > > > > >
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |