I'm having no luck setting my drives to dma mode and playing dvd movies is excruciatingly slow. Copying large files sends CPU usage to 100% I've read everything I could find in the sdb, tried finding info in the howto's and SuSEHelpCenter. Sorry for the length of this mail but I've tried to include enough info to be helpful. Here's the scoop, I've got 1 dvd player, one cdrw, and 2 Seagate Barracuda IV 80gb hard drives in a new machine. The dvd is from an older box but has been used to play movies in the past w/out problems. Currently using SuSE8.1 professional) My Drives are all IDE drives and the cdrw and dvd are ATAPI. On the first IDE cable I have Seagate Hard drive #1......./dev/hda, the dvd player....................../dev/dvd (a link to /dev/scd0) On the second IDE cable is Seagate HD #2..................../dev/hdc and the cdrw......................../dev/cdrom (a link to /dev/sr0) I've tried yast2 and hdparm to enable dma on the drives but when all is said and done everything stays as it was :-( (Yast reports "Device /dev/hdx: Cannot set DMA to value 1") hdparm -d1 /dev/hda gets the following result /dev/hda: setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off) hdparm on the other devices gives similar results I'm guessing that I'm not being allowed to set dma mode because the optical devices are treated as scsi, but they won't work at all as /dev/hdb, /dev/hdd ------------------------DIAGNOSTICS FOLLOW----------------- Here is a bit of my /etc/fstab... /dev/hda5 /boot ext2 defaults 1 2 /dev/hdc5 / reiserfs defaults 1 1 /dev/cdrecorder /media/cdrecorder auto ro,noauto,user,exec 0 0 /dev/cdrom /media/cdrom auto ro,noauto,user,exec 0 0 /dev/dvd /media/dvd auto ro,noauto,user,exec 0 0 When I ran xine-check to diagnose my slow movie playing it reported: "[ hint ] Your DVD drive seems not to be attached via ATAPI. This might be due to the use of an ide-scsi emulation. If you really have a SCSI DVD drive, your SCSI controller is likely to do perfect DMA, so there's no reason to worry about this. However, if you're using ide-scsi, there is a chance that DMA is disabled for the DVD drive. Moreover, I don't know how to enable DMA in that case, so you probably have to live with some performance loss. (FIXME: check for /proc/ide, provide solution) press <enter> to continue..." Part of what hdparm reports follows: luna:/home/dh1 # /sbin/hdparm -I /dev/hda ATA device, with non-removable media Model Number: ST380021A ---------snip------------ Standards: Supported: 5 4 3 2 Likely used: 6 --------snip------------- Capabilities: LBA, IORDY(can be disabled) bytes avail on r/w long: 4 Queue depth: 1 Standby timer values: spec'd by Standard R/W multiple sector transfer: Max = 16 Current = 16 Recommended acoustic management value: 128, current value: 128 DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 *udma5 PIO: pio0 pio1 pio2 pio3 pio4 hdparm will give info for the cdrw and dvd if I call them /dev/hdd and /dev/hdb respectively but if they are named as they are shown in /etc/fstab (scsi devices sr0 and scd0) then hdparm refuses (as it doesn't support scsi?) ------------------------QUESTIONS-------------------- How can I set DMA for these 4 devices. BTW i use lilo for my boot loader, it was suggested that there might be a boot parameter (APPEND=...) that might help me. Thanks for reading this, -- dh Don't shop at GoogleGear.com!
The 03.01.02 at 17:46, David Herman wrote:
My Drives are all IDE drives and the cdrw and dvd are ATAPI. On the first IDE cable I have Seagate Hard drive #1......./dev/hda, the dvd player....................../dev/dvd (a link to /dev/scd0) On the second IDE cable is Seagate HD #2..................../dev/hdc and the cdrw......................../dev/cdrom (a link to /dev/sr0)
I've tried yast2 and hdparm to enable dma on the drives but when all is said and done everything stays as it was :-( (Yast reports "Device /dev/hdx: Cannot set DMA to value 1") hdparm -d1 /dev/hda gets the following result /dev/hda: setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off) hdparm on the other devices gives similar results
I understand that hdparm fails to set DMA on your Seagate HD, and that is a bad signal. What I would try is connecting both hardisks on IDE cable number one, and the DVD and CDrom on the second cable. I had to do that on my system to get it working faster. -- Cheers, Carlos Robinson
Carlos E. R. wrote:
The 03.01.02 at 17:46, David Herman wrote:
My Drives are all IDE drives and the cdrw and dvd are ATAPI. On the first IDE cable I have Seagate Hard drive #1......./dev/hda, the dvd player....................../dev/dvd (a link to /dev/scd0) On the second IDE cable is Seagate HD #2..................../dev/hdc and the cdrw......................../dev/cdrom (a link to /dev/sr0)
I've tried yast2 and hdparm to enable dma on the drives but when all is said and done everything stays as it was :-( (Yast reports "Device /dev/hdx: Cannot set DMA to value 1") hdparm -d1 /dev/hda gets the following result /dev/hda: setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off) hdparm on the other devices gives similar results
I understand that hdparm fails to set DMA on your Seagate HD, and that is a bad signal. What I would try is connecting both hardisks on IDE cable number one, and the DVD and CDrom on the second cable. I had to do that on my system to get it working faster.
I see that your devices are both scd0 and sr0 which are intrinsically scsi devices, unlike the hdx which non-recording cdroms are generally listed as, e.g. lr-xr-xr-- /dev/cdrom -->/dev/hdd or hdc, hde.....hdx. So, if your device reads something like lrwxrwxr-x /dev/cdrom -->/dev/sr0 I'm banking that you have the ide-scsi module loaded and therefore it is a quasi scsi device. So, the problem is the hparm will not interact with this, such is the case on my system. My CDRW is under the ide-scsi module and my CDrom is under the cd-ide module -- I can set the CDrom but the the RW. Do and lsmod and see if "ide-scsi" is indeed loaded. If so then this is the situation I have explained and hdparm will not interface with these devices because it thinks they're scsi. There are scsi programs but I haven't found one that will do things in the same manner that hdparm will YMMV. Cheers, Curtis.
On Thursday 02 January 2003 07:14 pm, Curtis Rey wrote: -------------snip--------------
I see that your devices are both scd0 and sr0 which are intrinsically scsi devices, unlike the hdx which non-recording cdroms are generally listed as, e.g. lr-xr-xr-- /dev/cdrom -->/dev/hdd or hdc, hde.....hdx. So, if your device reads something like lrwxrwxr-x /dev/cdrom -->/dev/sr0 I'm banking that you have the ide-scsi module loaded and therefore it is a quasi scsi device. So, the problem is the hparm will not interact with this, such is the case on my system. My CDRW is under the ide-scsi module and my CDrom is under the cd-ide module -- I can set the CDrom but the the RW.
Do and lsmod and see if "ide-scsi" is indeed loaded.
Yes, just as you described. ide-scsi is loaded. If so then this
is the situation I have explained and hdparm will not interface with these devices because it thinks they're scsi. There are scsi programs but I haven't found one that will do things in the same manner that hdparm will YMMV.
Cheers, Curtis.
Thanks for your reply, I guess I will be trying Carlos' idea and place both hard drives on the same cable. I had hoped not to do this thinking that my swap partition would be faster on a seperate IDE chain (terminology?). Alternately, is there some way to change the dvd player to IDE mode? Some way to turn ide-scsi off? It seems like this was how things worked on the box I removed the drive from. Yast2 originally was able to find the dvd when it was linked to /dev/hdb, (It seems that yast2 approaches dvd's differently than is normal, ie no mount/umount when installing new packages from my dvd) but I just tried browsing a cd with /media/dvd (directory) pointing to /dev/hdb (with the appropriate fstab entry) and had no luck thanks again. -- dh Don't shop at GoogleGear.com!
About this:
/dev/cdrom -->/dev/sr0 I'm banking that you have the ide-scsi module loaded and therefore it is a quasi scsi device. So, the problem is the hparm will not interact with this, such is the case on my system.
My DVD/CDRW in the master on IDE2 (/dev/hdc), but it is recognized by suse as /dev/sr0, I have /dev/cdrom pointing to /dev/sr0... and still, if I do hdparm -d1 /dev/hdc, DMA gets activated for that drive... go figure! By doing this, I activate DMA on my DVD (even if YasT/SuSE recognices it as scsi), and my DVD movies play perfectly!... with Xine. I still do not know if this screws CD burning, since I have no tried it, so I just turn DMA off after watching a movie with hdparm -d0 /dev/hdc. __________________________________________________ Do you Yahoo!? Yahoo! Mail Plus - Powerful. Affordable. Sign up now. http://mailplus.yahoo.com
On Thursday 02 January 2003 07:55 pm, Marino Fernandez wrote: ----------snip-------------
My DVD/CDRW in the master on IDE2 (/dev/hdc), but it is recognized by suse as /dev/sr0, I have /dev/cdrom pointing to /dev/sr0... and still, if I do hdparm -d1 /dev/hdc, DMA gets activated for that drive... go figure!
By doing this, I activate DMA on my DVD (even if YasT/SuSE recognices it as scsi), and my DVD movies play perfectly!... with Xine.
That's good news, Guess I'll do some hardware reconfiguring tomorrow.
I still do not know if this screws CD burning, since I have no tried it, so I just turn DMA off after watching a movie with hdparm -d0 /dev/hdc.
-- dh Don't shop at GoogleGear.com!
The 03.01.02 at 19:47, David Herman wrote:
Thanks for your reply, I guess I will be trying Carlos' idea and place both hard drives on the same cable. I had hoped not to do this thinking that my swap partition would be faster on a seperate IDE chain (terminology?).
My original idea for separating them was because both HD support UDMA modes with the new thin ATA100 cable, bot not the CDwriter, so I assumed that mixing HD and writer on the same cable would slow down the HD. So, hdparm reports udma5 for both my hard disks, and the slower udma2 for the CDwriter and DVD reader.
Alternately, is there some way to change the dvd player to IDE mode? Some way to turn ide-scsi off? It seems like this was how things worked on the box I removed the drive from.
I have the DVD with ide-scsi, and I can set dma on with hdparm, no problem: nimrodel:~ # hdparm -d1 /dev/hdc /dev/hdc: setting using_dma to 1 (on) using_dma = 1 (on) Though it does complains a bit, after all hdparm acceses it the wrong way for an ide-scsi: nimrodel:~ # hdparm /dev/hdc /dev/hdc: HDIO_GET_MULTCOUNT failed: Input/output error IO_support = 0 (default 16-bit) unmaskirq = 0 (off) using_dma = 1 (on) keepsettings = 0 (off) readonly = 0 (off) BLKRAGET failed: Input/output error HDIO_GETGEO failed: Invalid argument That might be different in your case, of course. nimrodel:~ # hdparm /dev/hdc
Yast2 originally was able to find the dvd when it was linked to /dev/hdb, (It seems that yast2 approaches dvd's differently than is normal, ie no mount/umount when installing new packages from my dvd)
Yast mounts the dvd or cdrom on its own, and on a different directory than specified in fstab.
but I just tried browsing a cd with /media/dvd (directory) pointing to /dev/hdb (with the appropriate fstab entry) and had no luck
Well, if you have ide-scsi set, it wouldn't work that way. -- Cheers, Carlos Robinson
From what I understand anytime that you but a slower device (in throughput) the ide interface with always default to the slower speed. Makes sense since if you had a udma33 and a udma100 device the 33 speed can't run at 100 but the 100 can run at 33. I've heard that one should put the cdroms on on ide ribbon and the hdd's on another. On Friday 03 January 2003 20:06, Carlos E. R. wrote:
The 03.01.02 at 19:47, David Herman wrote:
Thanks for your reply, I guess I will be trying Carlos' idea and place both hard drives on the same cable. I had hoped not to do this thinking that my swap partition would be faster on a seperate IDE chain (terminology?).
My original idea for separating them was because both HD support UDMA modes with the new thin ATA100 cable, bot not the CDwriter, so I assumed that mixing HD and writer on the same cable would slow down the HD.
So, hdparm reports udma5 for both my hard disks, and the slower udma2 for the CDwriter and DVD reader.
Alternately, is there some way to change the dvd player to IDE mode? Some way to turn ide-scsi off? It seems like this was how things worked on the box I removed the drive from.
I have the DVD with ide-scsi, and I can set dma on with hdparm, no problem:
nimrodel:~ # hdparm -d1 /dev/hdc
/dev/hdc: setting using_dma to 1 (on) using_dma = 1 (on)
Though it does complains a bit, after all hdparm acceses it the wrong way for an ide-scsi:
nimrodel:~ # hdparm /dev/hdc
/dev/hdc: HDIO_GET_MULTCOUNT failed: Input/output error IO_support = 0 (default 16-bit) unmaskirq = 0 (off) using_dma = 1 (on) keepsettings = 0 (off) readonly = 0 (off) BLKRAGET failed: Input/output error HDIO_GETGEO failed: Invalid argument
That might be different in your case, of course.
nimrodel:~ # hdparm /dev/hdc
Yast2 originally was able to find the dvd when it was linked to /dev/hdb, (It seems that yast2 approaches dvd's differently than is normal, ie no mount/umount when installing new packages from my dvd)
Yast mounts the dvd or cdrom on its own, and on a different directory than specified in fstab.
but I just tried browsing a cd with /media/dvd (directory) pointing to /dev/hdb (with the appropriate fstab entry) and had no luck
Well, if you have ide-scsi set, it wouldn't work that way.
-- Billboard Writer vs. Literature = Microsoft vs. Computing,
The 03.01.03 at 21:11, Curtis Rey wrote:
From what I understand anytime that you but a slower device (in throughput) the ide interface with always default to the slower speed. Makes sense since if you had a udma33 and a udma100 device the 33 speed can't run at 100 but the 100 can run at 33. I've heard that one should put the cdroms on on ide ribbon and the hdd's on another.
I don't have confirmation of that, but my educated guess is that what you say is true. My doubts arise from the fact that as the hardware can not talk to the drives simultanously, it could perhaps use different speeds for each one. But probably not. After all, udma100 needs a different cable than for udma33. -- Cheers, Carlos Robinson
A udma cable has double the wires from the old/standard version. These are, to my understanding used to help with the signal. The extra wires are in between the data wires and are for the purpose of stopping signal leak/interferance between the wires (I think they are ground wires). Anyway, a udma66/100/133 cable can transmit the signal from a udma33 drive, it is backwards compatible, and of course a udma33 cable is incapable of utilizing the 66/100/133 Mb signal. I still think that having the mixed drives will limit the spec to the lower transfer rate none the less. Cheers, Curtis On Saturday 04 January 2003 07:45, Carlos E. R. wrote:
The 03.01.03 at 21:11, Curtis Rey wrote:
From what I understand anytime that you but a slower device (in throughput) the ide interface with always default to the slower speed. Makes sense since if you had a udma33 and a udma100 device the 33 speed can't run at 100 but the 100 can run at 33. I've heard that one should put the cdroms on on ide ribbon and the hdd's on another.
I don't have confirmation of that, but my educated guess is that what you say is true. My doubts arise from the fact that as the hardware can not talk to the drives simultanously, it could perhaps use different speeds for each one. But probably not.
After all, udma100 needs a different cable than for udma33.
-- Billboard Writer vs. Literature = Microsoft vs. Computing,
Thanks to everyone for your efforts to help with my problem. On Saturday 04 January 2003 12:56 pm, Curtis Rey wrote:
A udma cable has double the wires from the old/standard version. These are, to my understanding used to help with the signal. The extra wires are in between the data wires and are for the purpose of stopping signal leak/interferance between the wires (I think they are ground wires). Anyway, a udma66/100/133 cable can transmit the signal from a udma33 drive, it is backwards compatible, and of course a udma33 cable is incapable of utilizing the 66/100/133 Mb signal.
I still think that having the mixed drives will limit the spec to the lower transfer rate none the less.
Yes I agree with you, I remember reading that as well (now that you pointed it out). Seems that there are conflicting Ideas about what should work, but I agree that the drive speed isue is probably greater than the time needed to access swap... As far as I know my cables are UDMA100 capable but I may have to try changing them next. (soon I'll have more money in drive cables than I do in drives :-( ) I've read the numerous mails in this and the concurrent CD-R(W) threads but my problems continue. Here's what I have tried. Swapped my cdrw and dvd from their original positions, now I have: On the first IDE cable: 80gb IDE hard drive (udma100 capable) - linked to /dev/hda 80gb IDE hard drive (udma100 capable) - linked to /dev/hdb On the second IDE cable: ATAPI DVD-ROM - linked to /dev/hdc ATAPI CDRW - linked to /dev/sr1 Added to my lilo.conf: (as suggested by Joe Morris) append = "hdc=ide-cd hdd=ide-scsi" /var/log/bootmsg shows: <4>Kernel command line: auto BOOT_IMAGE=win4lin ro root=345 hdc=ide-cd hdd=ide-scsi <4>ide_setup: hdc=ide-cd <4>ide_setup: hdd=ide-scsi ----------snip---------- <4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive ------------snip----------- <4>VP_IDE: chipset revision 6 <4>VP_IDE: not 100%% native mode: will probe irqs later <4>VP_IDE: Unknown VIA SouthBridge, contact Vojtech Pavlik (my board is an msi KT3 Ultra2, via kt333 chipset) <4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive <4>ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 <4>ide0: probed IRQ 14 failed, using default. <4>ide1 at 0x170-0x177,0x376 on irq 15 <4>ide1: probed IRQ 15 failed, using default. (I don't know if these failures mean anything BUT it seems significant) (In a message from Carlos Robinson showing part of his boot.log DMA seemed to be set up at this point (snip follows) (<4>ICH2: not 100%% native mode: will probe irqs later <4> ide0: BM-DMA ... , BIOS settings: hda:DMA, hdb:DMA <4> ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:DMA, hdd:DMA) ----------snip------------- <6>scsi1 : SCSI host adapter emulation for IDE ATAPI devices <4> Vendor: LITE-ON Model: LTR-48125W Rev: VS06 <4> Type: CD-ROM ANSI SCSI revision: 02 <7>sd_attach() ----------snip------------- Setting up IDE DMA mode Force IDE DMA mode on: hda hdb done So The drives are accessible but DMA mode (doesnt really seem to get set) -- dh Don't shop at GoogleGear.com!
Ok, I have the same problems with my latest Maxtor drive in ide2 (doesn't matter which interconnect) and the issue revolves around the specific firmware that comes with the Maxtor HD. I have to turn off the dma for that specific settings because the irq and/or partitions fail (at least most of the time) and this solved the issue. Also, in my case, I have the HD on #2 interconnect on ide1 and this is mostly likely what the problem is. I really should change the settings in my fstab, the shutdown and change the interconnects to the onboard promise connects - but I digress. So, the problem isn't so much about setting things with hdparm as it is with the boot process trying to enable DMA at start up and failing. I have to use the boot parameter "ide=nodma" and then I set the dma in the kde CC/yast for that specific HD to off and the others to on. Though this is optimal or preferred it does solve my problem and I no longer suffer for lost IRQ's or bad patition readings. It's about the Maxtor drive and they way it is manufactured. I will probably stick with IBM or others, but there's never any real garauntee - the but the present Maxtor doesn't make be happy (it's a slim module and I suspect it may be along, if not actually, made for laptops - at least that was my impression with I really open the box and looked at the HD for the 1st time). HTH, Curtis On Saturday 04 January 2003 16:07, David Herman wrote:
Thanks to everyone for your efforts to help with my problem.
On Saturday 04 January 2003 12:56 pm, Curtis Rey wrote:
A udma cable has double the wires from the old/standard version. These are, to my understanding used to help with the signal. The extra wires are in between the data wires and are for the purpose of stopping signal leak/interferance between the wires (I think they are ground wires). Anyway, a udma66/100/133 cable can transmit the signal from a udma33 drive, it is backwards compatible, and of course a udma33 cable is incapable of utilizing the 66/100/133 Mb signal.
I still think that having the mixed drives will limit the spec to the lower transfer rate none the less.
Yes I agree with you, I remember reading that as well (now that you pointed it out). Seems that there are conflicting Ideas about what should work, but I agree that the drive speed isue is probably greater than the time needed to access swap... As far as I know my cables are UDMA100 capable but I may have to try changing them next. (soon I'll have more money in drive cables than I do in drives :-( )
I've read the numerous mails in this and the concurrent CD-R(W) threads but my problems continue.
Here's what I have tried.
Swapped my cdrw and dvd from their original positions, now I have: On the first IDE cable: 80gb IDE hard drive (udma100 capable) - linked to /dev/hda 80gb IDE hard drive (udma100 capable) - linked to /dev/hdb On the second IDE cable: ATAPI DVD-ROM - linked to /dev/hdc ATAPI CDRW - linked to /dev/sr1
Added to my lilo.conf: (as suggested by Joe Morris) append = "hdc=ide-cd hdd=ide-scsi"
/var/log/bootmsg shows: <4>Kernel command line: auto BOOT_IMAGE=win4lin ro root=345 hdc=ide-cd hdd=ide-scsi <4>ide_setup: hdc=ide-cd <4>ide_setup: hdd=ide-scsi ----------snip---------- <4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive ------------snip----------- <4>VP_IDE: chipset revision 6 <4>VP_IDE: not 100%% native mode: will probe irqs later <4>VP_IDE: Unknown VIA SouthBridge, contact Vojtech Pavlik (my board is an msi KT3 Ultra2, via kt333 chipset) <4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive <4>ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 <4>ide0: probed IRQ 14 failed, using default. <4>ide1 at 0x170-0x177,0x376 on irq 15 <4>ide1: probed IRQ 15 failed, using default. (I don't know if these failures mean anything BUT it seems significant) (In a message from Carlos Robinson showing part of his boot.log DMA seemed to be set up at this point (snip follows) (<4>ICH2: not 100%% native mode: will probe irqs later <4> ide0: BM-DMA ... , BIOS settings: hda:DMA, hdb:DMA <4> ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:DMA, hdd:DMA) ----------snip------------- <6>scsi1 : SCSI host adapter emulation for IDE ATAPI devices <4> Vendor: LITE-ON Model: LTR-48125W Rev: VS06 <4> Type: CD-ROM ANSI SCSI revision: 02 <7>sd_attach() ----------snip------------- Setting up IDE DMA mode
Force IDE DMA mode on: hda hdb done
So The drives are accessible but DMA mode (doesnt really seem to get set)
-- Billboard Writer vs. Literature = Microsoft vs. Computing,
On 01/05/2003 06:07 AM, David Herman wrote:
Here's what I have tried.
Swapped my cdrw and dvd from their original positions, now I have: On the first IDE cable: 80gb IDE hard drive (udma100 capable) - linked to /dev/hda 80gb IDE hard drive (udma100 capable) - linked to /dev/hdb On the second IDE cable: ATAPI DVD-ROM - linked to /dev/hdc ATAPI CDRW - linked to /dev/sr1
The CDRW would be /dev/sr0
Added to my lilo.conf: (as suggested by Joe Morris) append = "hdc=ide-cd hdd=ide-scsi"
This was only if you used lilo. Otherwise, if GRUB, IIRC it is menu.lst (I only have 8.0 and use lilo).
/var/log/bootmsg shows: <4>Kernel command line: auto BOOT_IMAGE=win4lin ro root=345 hdc=ide-cd hdd=ide-scsi <4>ide_setup: hdc=ide-cd <4>ide_setup: hdd=ide-scsi ----------snip---------- <4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive ------------snip----------- <4>VP_IDE: chipset revision 6 <4>VP_IDE: not 100%% native mode: will probe irqs later <4>VP_IDE: Unknown VIA SouthBridge, contact Vojtech Pavlik (my board is an msi KT3 Ultra2, via kt333 chipset)
I think this is your problem. I know there was a VIA Southridge patch in one of the later kernels. Maybe you should try a newer kernel.
<4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive <4>ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 <4>ide0: probed IRQ 14 failed, using default. <4>ide1 at 0x170-0x177,0x376 on irq 15 <4>ide1: probed IRQ 15 failed, using default. (I don't know if these failures mean anything BUT it seems significant)
IRQ 14 is your primary IDE controller, IRQ 15 is your secondary IDE controller. I think you have found the source of your problem. I don't believe your controller is supported in your kernel. -- Joe & Sesil Morris New Tribes Mission Email Address: Joe_Morris@ntm.org Web Address: http://www.mydestiny.net/~joe_morris Registered Linux user 231871 God said, I AM that I AM. I say, by the grace of God, I am what I am.
On Saturday 04 January 2003 04:54 pm, Joe Morris (NTM) wrote:
On 01/05/2003 06:07 AM, David Herman wrote:
Here's what I have tried. -----------snip------------- ATAPI CDRW - linked to /dev/sr1
The CDRW would be /dev/sr0
Yeah I grokked that the first time I actually tried to access the drive, Thanks :-) (I must be learning something)
Added to my lilo.conf: (as suggested by Joe Morris) append = "hdc=ide-cd hdd=ide-scsi"
This was only if you used lilo. Otherwise, if GRUB, IIRC it is menu.lst (I only have 8.0 and use lilo).
Still using lilo here
/var/log/bootmsg shows: ------------snip----------- <4>VP_IDE: chipset revision 6 <4>VP_IDE: not 100%% native mode: will probe irqs later <4>VP_IDE: Unknown VIA SouthBridge, contact Vojtech Pavlik (my board is an msi KT3 Ultra2, via kt333 chipset)
I think this is your problem. I know there was a VIA Southridge patch in one of the later kernels. Maybe you should try a newer kernel.
I was afraid of that, but glad to know that I may be getting closer to a solution.
<4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive <4>ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 <4>ide0: probed IRQ 14 failed, using default. <4>ide1 at 0x170-0x177,0x376 on irq 15 <4>ide1: probed IRQ 15 failed, using default. (I don't know if these failures mean anything BUT it seems significant)
IRQ 14 is your primary IDE controller, IRQ 15 is your secondary IDE controller. I think you have found the source of your problem. I don't believe your controller is supported in your kernel.
Thanks for taking the time to help, I'll look into the kernel patch as time allows. At least I can stop trying the same things over and over. -- dh Don't shop at GoogleGear.com!
The 03.01.04 at 14:07, David Herman wrote:
As far as I know my cables are UDMA100 capable but I may have to try changing them next. (soon I'll have more money in drive cables than I do in drives :-( )
I don't remember right now, but I think they are detected by the bios during the boot, or reported somewhere. At least, you can not use the high udma modes if the cable is not capable. Ah, yes, the boot.msg has it: <6>hda: 117231408 sectors (60022 MB) w/2048KiB Cache, CHS=7297/255/63, UDMA(100)
Swapped my cdrw and dvd from their original positions, now I have: On the first IDE cable: 80gb IDE hard drive (udma100 capable) - linked to /dev/hda 80gb IDE hard drive (udma100 capable) - linked to /dev/hdb On the second IDE cable: ATAPI DVD-ROM - linked to /dev/hdc ATAPI CDRW - linked to /dev/sr1
<4>Kernel command line: auto BOOT_IMAGE=win4lin ro root=345 hdc=ide-cd hdd=ide-scsi <4>ide_setup: hdc=ide-cd <4>ide_setup: hdd=ide-scsi ----------snip---------- <4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive ------------snip----------- <4>VP_IDE: chipset revision 6 <4>VP_IDE: not 100%% native mode: will probe irqs later <4>VP_IDE: Unknown VIA SouthBridge, contact Vojtech Pavlik (my board is an msi KT3 Ultra2, via kt333 chipset)
Ralph also has VP_IDE, and he has problems as well. Mine is "ICH2"
<4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive <4>ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 <4>ide0: probed IRQ 14 failed, using default. <4>ide1 at 0x170-0x177,0x376 on irq 15 <4>ide1: probed IRQ 15 failed, using default. (I don't know if these failures mean anything BUT it seems significant)
He has the same problem here than you. One question: do you see any thing about ACPI in the boot.msg log? ACPI (Advanced Configuration and Power Interface) is also related to IRQ handling on the new boards, and is also related to APIC. Many people complained of computer freeze problems (including me) and one of the proposals was to disable ACPI; but that may affect how the IRQs are detected and/or handled. In my case, ACPI is enabled, but not APIC (not fully) - if I do, I loose USB. And, DMA and IRQ are related.
(In a message from Carlos Robinson showing part of his boot.log DMA seemed to be set up at this point (snip follows) (<4>ICH2: not 100%% native mode: will probe irqs later <4> ide0: BM-DMA ... , BIOS settings: hda:DMA, hdb:DMA <4> ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:DMA, hdd:DMA)
Yes, but not quite: that only report the bios settings. I have to enable DMA using hdparm later for the dvd, is not enabled when I need it. In yast, I can configure dma for the hard disks, but not for the ide-scsi drives, they don't appear.
Setting up IDE DMA mode
Force IDE DMA mode on: hda hdb done
That's during boot, or later?
So The drives are accessible but DMA mode (doesnt really seem to get set)
What happens if you issue "hdparm -d1 /dev/hdc /dev/hdd"? In your first message you said it failed, but as you have changed the connections, maybe it varies :-? If the command fails, check what it says for hdb and hda as well, for comparison. -- Cheers, Carlos Robinson
Carlos E. R. wrote:
The 03.01.04 at 14:07, David Herman wrote:
As far as I know my cables are UDMA100 capable but I may have to try changing them next. (soon I'll have more money in drive cables than I do in drives :-( )
I don't remember right now, but I think they are detected by the bios during the boot, or reported somewhere. At least, you can not use the high udma modes if the cable is not capable. Ah, yes, the boot.msg has it:
<6>hda: 117231408 sectors (60022 MB) w/2048KiB Cache, CHS=7297/255/63, UDMA(100)
Swapped my cdrw and dvd from their original positions, now I have: On the first IDE cable: 80gb IDE hard drive (udma100 capable) - linked to /dev/hda 80gb IDE hard drive (udma100 capable) - linked to /dev/hdb On the second IDE cable: ATAPI DVD-ROM - linked to /dev/hdc ATAPI CDRW - linked to /dev/sr1
<4>Kernel command line: auto BOOT_IMAGE=win4lin ro root=345 hdc=ide-cd hdd=ide-scsi <4>ide_setup: hdc=ide-cd <4>ide_setup: hdd=ide-scsi ----------snip---------- <4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive ------------snip----------- <4>VP_IDE: chipset revision 6 <4>VP_IDE: not 100%% native mode: will probe irqs later <4>VP_IDE: Unknown VIA SouthBridge, contact Vojtech Pavlik (my board is an msi KT3 Ultra2, via kt333 chipset)
Ralph also has VP_IDE, and he has problems as well. Mine is "ICH2"
<4>hda: ST380021A, ATA DISK drive <4>hdb: ST380021A, ATA DISK drive <4>hdc: HITACHI DVD-ROM GD-5000, ATAPI CD/DVD-ROM drive <4>hdd: LITE-ON LTR-48125W, ATAPI CD/DVD-ROM drive <4>ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 <4>ide0: probed IRQ 14 failed, using default. <4>ide1 at 0x170-0x177,0x376 on irq 15 <4>ide1: probed IRQ 15 failed, using default. (I don't know if these failures mean anything BUT it seems significant)
He has the same problem here than you. One question: do you see any thing about ACPI in the boot.msg log? ACPI (Advanced Configuration and Power Interface) is also related to IRQ handling on the new boards, and is also related to APIC. Many people complained of computer freeze problems (including me) and one of the proposals was to disable ACPI; but that may affect how the IRQs are detected and/or handled. In my case, ACPI is enabled, but not APIC (not fully) - if I do, I loose USB. And, DMA and IRQ are related.
(In a message from Carlos Robinson showing part of his boot.log DMA seemed to be set up at this point (snip follows) (<4>ICH2: not 100%% native mode: will probe irqs later <4> ide0: BM-DMA ... , BIOS settings: hda:DMA, hdb:DMA <4> ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:DMA, hdd:DMA)
Yes, but not quite: that only report the bios settings. I have to enable DMA using hdparm later for the dvd, is not enabled when I need it. In yast, I can configure dma for the hard disks, but not for the ide-scsi drives, they don't appear.
Setting up IDE DMA mode
Force IDE DMA mode on: hda hdb done
That's during boot, or later?
So The drives are accessible but DMA mode (doesnt really seem to get set)
What happens if you issue "hdparm -d1 /dev/hdc /dev/hdd"? In your first message you said it failed, but as you have changed the connections, maybe it varies :-? If the command fails, check what it says for hdb and hda as well, for comparison.
Just a comment which may be irrelevant because I have not been following this thread closely. Are the HDs sitting in cradles? I had a hell of a time trying to figure out why I was having big problems with my new HD until I realised that while I had the correct 80-wire cable connecting the motherboard to the cradle body the cable INSIDE the cradle in which the HD is located had the old 40-wire cable :-(. I replaced the old cradle with a new one and the problem disappeared. Cheers. -- Sound that shatters silence is called noise. Sound that enhances silence is called music.
Hi All, I am trying to get Suse 8.1 to be up and running on a old P-166, now at boot time it complains something of ACPI/APCI. gets installed well and good but does not take a perfectly working 3c59x ethernet adapter and complaints at startup of ifup: Interface name invalid.(I did this posting also..) What is new is that Now the moment i plug this HDD on a new P-III motherboard it works fine, with the same network card..anything I need to enable on this old P-1 motherboard / disable something ? I am refurbishing old PCs with Linux but apparently I cannot proceed further if there is no networking..:-(... Any help is gr8.. (I Tried all the make module etc.etc.. tried a RTL card and a SureCom card also but to fail..) Cheers Vishal ..
* Vishal Khanna
Hi All, I am trying to get Suse 8.1 to be up and running on a old P-166, now at boot time it complains
You, sir, hijacked the thread! Please adhere to accepted procedure. -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org icq#173753138
On Sunday 05 January 2003 06:09 pm, Carlos E. R. wrote: ---------snip-----------
I don't remember right now, but I think they are detected by the bios during the boot, or reported somewhere. At least, you can not use the high udma modes if the cable is not capable. Ah, yes, the boot.msg has it:
<6>hda: 117231408 sectors (60022 MB) w/2048KiB Cache, CHS=7297/255/63, UDMA(100)
The corresponding part of my bootlog follows... <6>hda: 156301488 sectors (80026 MB) w/2048KiB Cache, CHS=9729/255/63 <4>hdb: safely enabled flush <6>hdb: 156301488 sectors (80026 MB) w/2048KiB Cache, CHS=9729/255/63 So it appears that in your case UDMA(100) is available while my boot.msg indicates it is not available on my machine (possibly due to cables?) Is this a correct interpretation?
------------snip----------- <4>VP_IDE: chipset revision 6 <4>VP_IDE: not 100%% native mode: will probe irqs later <4>VP_IDE: Unknown VIA SouthBridge, contact Vojtech Pavlik (my board is an msi KT3 Ultra2, via kt333 chipset)
Ralph also has VP_IDE, and he has problems as well. Mine is "ICH2" ---------------snip-------------
<4>ide1: probed IRQ 15 failed, using default. (I don't know if these failures mean anything BUT it seems significant)
He has the same problem here than you. One question: do you see any thing about ACPI in the boot.msg log? ACPI (Advanced Configuration and Power Interface) is also related to IRQ handling on the new boards, and is also related to APIC. Many people complained of computer freeze problems (including me) and one of the proposals was to disable ACPI; but that may affect how the IRQs are detected and/or handled. In my case, ACPI is enabled, but not APIC (not fully) - if I do, I loose USB. And, DMA and IRQ are related.
There are numerous references to ACPI and APIC I 'll have to try changing my bios settings, who knows, maybe it will help... ----------snip------------
Setting up IDE DMA mode
Force IDE DMA mode on: hda hdb done
That's during boot, or later?
Its about half way down my boot.msg, A slightly longer snip follows... Boot logging started on /dev/tty1(/dev/console) at Sun Jan 5 13:17:21 2003 Setting up IDE DMA mode Force IDE DMA mode on: hda hdb hdc done Configuring serial ports... ttyS0 at 0x03f8 (irq = 4) is a 16550A ttyS1 at 0x02f8 (irq = 3) is a 16550A Configured serial ports
So The drives are accessible but DMA mode (doesnt really seem to get set)
What happens if you issue "hdparm -d1 /dev/hdc /dev/hdd"? In your first message you said it failed, but as you have changed the connections, maybe it varies :-? If the command fails, check what it says for hdb and hda as well, for comparison.
Still appears to fail. I added "append = "hdc=ide-cd hdd=ide-scsi"" to my lilo.conf Here are the messages returned By my current setup: luna:/home/dh1 # hdparm -d1 /dev/hda /dev/hdb /dev/hdc /dev/hdd /dev/hda: (80gb ide harddrive) setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off) /dev/hdb: (80gb ide harddrive) setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off) /dev/hdc: (ATAPI dvd/cd player) setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off) /dev/hdd: (ATAPI CD-RW w ide-scsi emulation) setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off) luna:/home/dh1 # Thanks for your continued efforts with this problem, I'll keep you posted. -- dh Don't shop at GoogleGear.com!
The 03.01.05 at 22:52, David Herman wrote:
<6>hda: 117231408 sectors (60022 MB) w/2048KiB Cache, CHS=7297/255/63, UDMA(100)
The corresponding part of my bootlog follows... <6>hda: 156301488 sectors (80026 MB) w/2048KiB Cache, CHS=9729/255/63 <4>hdb: safely enabled flush <6>hdb: 156301488 sectors (80026 MB) w/2048KiB Cache, CHS=9729/255/63 So it appears that in your case UDMA(100) is available while my boot.msg indicates it is not available on my machine (possibly due to cables?) Is this a correct interpretation?
Could be. At least, it doesn't report it, so it could be the cables, or could be that thing about the south bridge Geoffrey has posted here.
handled. In my case, ACPI is enabled, but not APIC (not fully) - if I do, I loose USB. And, DMA and IRQ are related.
There are numerous references to ACPI and APIC I 'll have to try changing my bios settings, who knows, maybe it will help...
Try to upgrade the kernel from suse first, just in case.
Its about half way down my boot.msg, A slightly longer snip follows...
Boot logging started on /dev/tty1(/dev/console) at Sun Jan 5 13:17:21 2003 Setting up IDE DMA mode Force IDE DMA mode on: hda hdb hdc done
Yes, that's a script that reads from /etc/sysconfig/hardware (DEVICES_FORCE_IDE_DMA_ON="")
What happens if you issue "hdparm -d1 /dev/hdc /dev/hdd"? In your first message you said it failed, but as you have changed the connections, maybe it varies :-? If the command fails, check what it says for hdb and hda as well, for comparison.
Still appears to fail. I added "append = "hdc=ide-cd hdd=ide-scsi"" to my lilo.conf
Did you notice an email on the list about that? It appears you have to add ide-cd to the initial ramdisk, or you get ide-scsi on all cdroms.
Here are the messages returned By my current setup:
luna:/home/dh1 # hdparm -d1 /dev/hda /dev/hdb /dev/hdc /dev/hdd
/dev/hda: (80gb ide harddrive) setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off)
Mm, all drives reject dma, so it probably is not the scsi emulation or the drives, but the motherboard-bios-software combination. It doesn't know how to handle your IDE, yet. -- Cheers, Carlos Robinson
-------------------SNIP---------------------- So after hours wasted and many trials and tribulations my problems came down to a combination of A) Kernel not recognizing my via southbridge (msi KT3Ultra2 motherboard) 1) the kernel on the boxed set of SuSE linux 8.1 Pro, (not recognizing southbridge) 2) the kernel installed for SuSE 8.1 which was installed by win4lin (not recognizing southbridge) I tried my 2 hard drives on separate IDE channels, this worked as far as having an operating computer but would not allow DMA for either drive. At least it didn't crash. I tried both drives on the same IDE Cable, This led to so many crashes that my reiserFS became totally corrupt and unfixable, The crashes were unpredictable occurring anywhere from 10 minutes after boot up to 14 hours after boot. Being that I multiboot other OS' (as well as the rebooting I was doing trying to get DMA working)I didn't immediately recognize the source of the problem, and was not able to fix the fs w/ the available tools. I was able to salvage my /home directory , then re-installed SuSE. (Both hard drives are back on separate channels. Although this is not ideal speedwise (they share cables w/ a dvd and a cdrw), I needed to get up and running quickly so my e-mail wouldn't overflow (mainly this list ;-). I may try them back on the same cable as time allows. I then upgraded to the most recent stable kernel for 8.1 (2.4.19-115) and can now set DMA for both of my hard drives and my dvd (which started the whole adventure). I also downloaded the sources for SuSE kernel2.4.19-115, patched them for win4lin and compiled. This also works for DMA. currently /dev/hda seagateHD /dev/hdb dvd player /dev/hdc seagateHD /dev/cdrecorder linked to /dev/sr1 eventually I will probably place the 2 HD's on the same cable again but this works for now. Overall speed isn't as big of issue for me as the high CPU usage caused by disk activity without DMA Basically it came down to the kernel not recognizing the via southbridge as mentioned by Geoff Horn. Thanks to all who worked on this w/ me. Particularly Curtis and Carlos Hopefully this will help someone else. -- dh Don't shop at GoogleGear.com!
The 03.01.18 at 13:59, David Herman wrote:
I tried both drives on the same IDE Cable, This led to so many crashes that my reiserFS became totally corrupt and unfixable, The crashes were unpredictable occurring anywhere from 10 minutes after boot up to 14
Oh, my!
I then upgraded to the most recent stable kernel for 8.1 (2.4.19-115) and can now set DMA for both of my hard drives and my dvd (which started the whole adventure).
Good! I'm happy you got it working. -- Cheers, Carlos Robinson
I haven't been following this thread so please make allowances if the following has already been covered. I had problems with setting dma on with my MSI 6593 motherboard (KT3 Ultra2-C) with SuSE 8.1 Pro. I also got that message:
<4>VP_IDE: Unknown VIA SouthBridge, contact Vojtech Pavlik (my board is an msi KT3 Ultra2, via kt333 chipset).
In my case the VIA chip was the 8235. In the driver file via82cxxxc it was listed under FUTURE_BRIDGES. I found a patch from Vojtech Pavlik, applied it to via82cxxx.c and lo! dma "on" worked. Since then I have upgraded to the latest SuSE 8.1 kernel and drivers all is still well (k_athlon-2.4.19-167). And the 8235 has been moved out of FUTURE_BRIDGES. I hope that wasn't just so much static! -- Geoff Horn
Abou this
Do and lsmod and see if "ide-scsi" is indeed loaded. If so then this is the situation I have explained and hdparm will not interface with these devices
I have ide-scsi loaded and in my case hdparm works... again go figure!... but I can watch DVD fine I guess I'd better shut up. But thanks for all of your shared wisdom... this is great for a bigtime newby in linuxland like me... whoa! my laptop roars with suse 8.1 loaded! __________________________________________________ Do you Yahoo!? Yahoo! Mail Plus - Powerful. Affordable. Sign up now. http://mailplus.yahoo.com
On Thursday 02 January 2003 07:02 pm, Carlos E. R. wrote:
The 03.01.02 at 17:46, David Herman wrote:
My Drives are all IDE drives and the cdrw and dvd are ATAPI. On the first IDE cable I have Seagate Hard drive #1......./dev/hda, the dvd player....................../dev/dvd (a link to /dev/scd0) On the second IDE cable is Seagate HD #2..................../dev/hdc and the cdrw......................../dev/cdrom (a link to /dev/sr0)
I've tried yast2 and hdparm to enable dma on the drives but when all is said and done everything stays as it was :-( (Yast reports "Device /dev/hdx: Cannot set DMA to value 1") hdparm -d1 /dev/hda gets the following result /dev/hda: setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off) hdparm on the other devices gives similar results
I understand that hdparm fails to set DMA on your Seagate HD, and that is a bad signal. What I would try is connecting both hardisks on IDE cable number one, and the DVD and CDrom on the second cable. I had to do that on my system to get it working faster.
Thanks that idea did occur to me as I was writing my lengthy mail. hopefully I'll have time to try that tomorrow, I'll let the list know if it works. Thanks for your reply -- dh Don't shop at GoogleGear.com!
On Thursday 02 January 2003 07:02 pm, Carlos E. R. wrote:
I understand that hdparm fails to set DMA on your Seagate HD, and that is a bad signal. What I would try is connecting both hardisks on IDE cable number one, and the DVD and CDrom on the second cable. I had to do that on my system to get it working faster.
So I moved the disks as sggested but still no luck setting DMA w/ hdparm output follows: luna:/home/dh1 # hdparm -d1 /dev/hda /dev/hda: setting using_dma to 1 (on) HDIO_SET_DMA failed: Operation not permitted using_dma = 0 (off) I'm beginning to feel a bit frustrated. It seems that changing cables may be my next step. Thanks for your help -- dh Don't shop at GoogleGear.com!
participants (9)
-
Basil Chupin -
Carlos E. R. -
Curtis Rey -
David Herman -
Geoffrey Horn -
Joe Morris (NTM) -
Marino Fernandez -
SuSEnixER -
Vishal Khanna