HighPoint 374

What's the current status of HPT 374 support? Is it working in any kernel version?

I know HP provides binaries, but that's no good, I need to use other kernel builds.

2.4.20-8 detects the card and provides the devices but data corruptions happens all the time. Using the HP drivers everything works fine (note that access is via /dev/sd* instead of /dev/hd*)...except that I can't replace the kernel.

Thanks.

Works fine with 2.4.21 and 2.6.0-testX, at least. I guess 2.4.22 should work too, but I haven't tried it.

Måns Rullgård
[email protected]

Unfortunately, in my experience the included kernel driver doesn't work reliable at all with my hardware either. I have tried all possible kernel versions up to and including 2.4.23-pre5 without success.

The Highpoint proprietary drivers work fine in my case as well, but I'm not too happy about the fact that I simply can't access the SMART stats of my drivers at all for example.

I have tested the kernel hpt374 drivers with all possible kernel configurations with and without ACPI/APM, local apic, io-apic and they don't seem to change anything in my case. I just get the following error message sooner or later and the whole system hangs.

Sep 26 23:07:42 alderan kernel: blk: queue c0466908, I/O limit 4095Mb
(mask 0xffffffff)
Sep 26 23:07:42 alderan kernel: hde: dma_timer_expiry: dma status == 0x20
Sep 26 23:07:42 alderan kernel: hde: timeout waiting for DMA
Sep 26 23:07:42 alderan kernel: hde: timeout waiting for DMA
Sep 26 23:07:42 alderan kernel: blk: queue c0466d5c, I/O limit 4095Mb
(mask 0xffffffff)
Sep 26 23:07:42 alderan kernel: hdg: dma_timer_expiry: dma status == 0x20
Sep 26 23:07:42 alderan kernel: hdg: timeout waiting for DMA
Sep 26 23:07:42 alderan kernel: hdg: timeout waiting for DMA

It might be that my problems are somehow related to the motherboards I've been using, as the first one is "Epox 8K9A3+ (Via KT400 chipset)" while the another one is "Epox 4PCA3+ (Intel 875p chipset)". But they are 100% reproduceble with multiple brands of disk-drives.

Regards,
Tomi Orava

This is the EXACT same things I see in my logs.

Mine is an Asus CUSL2, Pentium III motherboard with 512 Mb.

Very interesting. The other type of errors I have received (the last time with 2.4.23-pre4) were:

Sep 14 14:29:49 alderan kernel: hde: set_drive_speed_status: status=0xff {
Busy }
Sep 14 14:29:49 alderan kernel: blk: queue c0492188, I/O limit 4095Mb
(mask 0xffffffff)
Sep 14 14:29:49 alderan kernel: hde: set_drive_speed_status: status=0xff {
Busy }
Sep 14 14:29:49 alderan kernel: hde: dma_timer_expiry: dma status == 0x21
Sep 14 14:29:49 alderan kernel: hde: error waiting for DMA
Sep 14 14:29:49 alderan kernel: hde: dma timeout retry: status=0x50 {
DriveReady SeekComplete }
Sep 14 14:29:49 alderan kernel:
Sep 14 14:29:49 alderan kernel: hdg: set_drive_speed_status: status=0xff {
Busy }
Sep 14 14:29:49 alderan kernel: blk: queue c04925dc, I/O limit 4095Mb
(mask 0xffffffff)
Sep 14 14:29:49 alderan kernel: hdg: dma_timer_expiry: dma status == 0x00
Sep 14 14:29:49 alderan kernel: hdg: timeout waiting for DMA
Sep 14 14:29:49 alderan kernel: hdg: timeout waiting for DMA
Sep 14 14:29:49 alderan kernel: hdg: (__ide_dma_test_irq) called while not
waiting
Sep 14 14:29:49 alderan kernel: blk: queue c0492a30, I/O limit 4095Mb
(mask 0xffffffff)
Sep 14 14:29:49 alderan kernel: hdk: set_drive_speed_status: status=0xff {
Busy }
Sep 14 14:29:49 alderan kernel: blk: queue c0492e84, I/O limit 4095Mb
(mask 0xffffffff)
Sep 14 14:29:49 alderan kernel: hdk: dma_timer_expiry: dma status == 0x01
Sep 14 14:29:49 alderan kernel: hdk: set_drive_speed_status: status=0xff {
Busy }
Sep 14 14:29:49 alderan kernel: hdk: error waiting for DMA
Sep 14 14:29:49 alderan kernel: hdk: dma timeout retry: status=0x50 {

DriveReady SeekComplete }

Ok, so it doesn't sound that the fault is only with Epox mb's. I sincerely hope that someone with more knowledge about ide-stuff could figure out something to overcome this particular problem ... Currently my disk-drivers are made by 2*samsung (SV8004H) and 2*Samsung(SV1604N), in case that changes anything.

BTW. Have you noticed in what situation the lock up happens ? As all my drivers are mirrored in pairs (ie. Raid1) the system might run stable for 2-3 days before a lockup unless there is some heavy transfers going on. In case any of the mirrors required syncing when bootingg up the system, the system will lockup immediately.

What does /proc/interrupts show in your case ?
Mine is:
/proc/interrupts:

CPU0

  0:  138788842    IO-APIC-edge  timer
  1:     328481    IO-APIC-edge  keyboard
  2:          0          XT-PIC  cascade
  8:   21357900    IO-APIC-edge  rtc
 14:      14205    IO-APIC-edge  ide0
 16:  104139080   IO-APIC-level  radeon@PCI:1:0:0
 17:  136188388   IO-APIC-level  ide2, ide3, hpt374
 19:   98331044   IO-APIC-level  eth0
 21:   70689058   IO-APIC-level  ehci_hcd, usb-uhci, usb-uhci, usb-uhci
 22:   10181857   IO-APIC-level  VIA8233
NMI:          0
LOC:  138780432
ERR:          0
MIS:     109612

And have you tried with ACPI on/off and io-apic on/off ?

Regards,
Tomi Orava

[email protected]

Same here.

I'm using two seagates (different models), one Samsung and one Maxtor, this happens in all 4.
I'm not using any kind of RAID.

If the filesystems are mounted read only it seems to run stable (even though I do see those errors)....if I mount r/w, any write is likely to make the system crash rather soon.

Using the HPT drivers:

CPU0

  0:     118661          XT-PIC  timer
  1:         14          XT-PIC  keyboard
  2:          0          XT-PIC  cascade
  3:       2698          XT-PIC  eth1
  5:       1133          XT-PIC  hpt374
  7:          0          XT-PIC  usb-uhci
  8:          1          XT-PIC  rtc
  9:       1098          XT-PIC  usb-uhci, eth0
 12:         28          XT-PIC  PS/2 Mouse
 14:      69807          XT-PIC  ide0
 15:         54          XT-PIC  ide1
NMI:          0
ERR:          0

Using the regular kernel IDE drivers

CPU0

  0:       9318          XT-PIC  timer
  1:         12          XT-PIC  keyboard
  2:          0          XT-PIC  cascade
  3:        151          XT-PIC  eth1
  5:        130          XT-PIC  ide2, ide3, ide4, ide5
  7:          0          XT-PIC  usb-uhci
  8:          1          XT-PIC  rtc
  9:        573          XT-PIC  usb-uhci, eth0
 12:         28          XT-PIC  PS/2 Mouse
 14:       9163          XT-PIC  ide0
 15:         54          XT-PIC  ide1
NMI:          0
ERR:          0

No, to be honest I didn't even think of this. You think it could make a difference? Given the fact that the card works correctly with the HPT drivers, pretty much everything that does not relate directly to the IDE drivers seems ruled out as the cause...

Hmm, if I checked correctly, your motherboard (Asus CUSL2) doesn't seem to include a HPT374 based (integrated) IDE-controller. Are you using some add-on PCI-card for IDE ? If yes, are you using older style Parallel-ATA or newer Serial-ATA interface ?

Based on previous questions and answers on this list, those settings might very well affect you, as they affect the way interrupts are handled in your system. Unfortunately, they haven't fixed my problems no matter what I combination I have tried.

Regards,
Tomi Orava

[email protected]

Tomi,

I'm using HighPoint's RocketRaid 404 card (HPT374 based PCI card).

In 2.4.21 and 2.4.22 it's working great for me. I'm using the "experimental" IDE Raid with two disks on a HPT 374 controller with the drivers that come with the kernel.

Joel Smith

I have tried these versions in the past as well without success. However, I don't use HPT-raid features at all ie. I'm using the disks as JBOD. What hardware do you have and have you enabled ACPI/local-apic/io-apic ? What brand & model of disk-drives you are using with HPT374 controller ? And finally what does the /proc/interrupts show for you ?

There really must be some explanation why some of us are having really huge problems with HPT374-contollers while for others it's working just fine. I haven't exactly heard anyone been too succesfull for example with Epox 8K9A3+ motherboard even on this mailing-list based on previous questions seen here.

Regards,
Tomi Orava

I'm using a RocketRAID 1540 SATA card (hpt374 based) in an Alpha system. It has no such thing as ACPI. The disks are four Seagate Barracuda 7200.7 running software raid5. My /proc/interrupts:

  1:      11440          XT-PIC   keyboard
  2:          0          XT-PIC   cascade
  8:  179530685             RTC  +timer
 14:        198          XT-PIC  +ide0
 15:         11          XT-PIC  +ide1
 22:          0             SRM   timer-cascade
 24:          0             SRM   usb-ohci
 25:    2327086             SRM  +CMI8738-MC6
 26:    4198981             SRM   ide2,  ide3,  ide4,  ide5
 28:      92591             SRM   usb-ohci
 32:   21577855             SRM   ehci-hcd

ERR: 0

The hpt374 is on IRQ 26.

Måns Rullgård
[email protected]

Ok, that might be one reason why it's working for you but not for me. If I've understood correctly, people who have problems with HPT374 are using the integrated Parallel-ATA interface instead of SATA.

I'd be really interested to hear if anyone has a working system with kernel included drivers & HPT374-controller integrated in motherboard and using PATA-drives ?

Regards,
Tomi Orava

Tomy,

The

hdparm -m0 device

seems to have fixed the problem for me. I'll try increasing the number in the following days and run extensive tests, but for now, it's quite enough.

BTW your email server doesn't seem to like my address and refuses to deliver any mail, if you aren't running it maybe you should tell the admin that he's blocking Spain's largest ISP for some reason?

----- Original Message -----
From: "Tomi Orava" <[email protected]>
To: "Måns Rullgård" <[email protected]>
Cc: <[email protected]>
Sent: Sunday, October 19, 2003 21:14
Subject: Re: HighPoint 374

the

That is *very* interesting. I quess that's about the only hdparm option that I have never tried before. I'm really interested in hearing if that really fixes the problem for you and for good. The thing is that the machine with HPT374-chip is running as a server so I'm not very eager to use it as a test bed (unless it's absolutely necessary). Please, if it's possible for you, try to copy say two big files from one disk to another at the same time couple of times in order to see if your machine is able to handle it ... I have been able to run mine 2-3 days in the past without any problems if there are _no_ major disk transfers going on.

The question of course is why does that hdparm option seem to fix the problem in this case ? Is it perhaps a bug in HPT374-driver or some lower IDE-layer ? (Just quessing ...)

I'm sorry, I don't accept mails coming from dsl/cable/modem-pools due to high amount of spam. I will accept mails coming from your ISP's mail gateway though.

Regards,
Tomi Orava

[email protected]

Tomi,

Copied one large DVD image from one to drive to another a couple of times, no problems whatsoever. Compared the file each time, everything looks good.

As for the email, it's not my ISP gateway but mine personally. By blocking the entire Telefonica IP pool you are denying access to a huge userbase, as Telefonica is, in the end, the ISP for almost everyone here. But well, it's your decision :-)

----- Original Message -----
From: "Tomi Orava" <[email protected]>
To: "Carlos Fernandez Sanz" <[email protected]>
Cc: <[email protected]>
Sent: Sunday, October 19, 2003 21:39
Subject: Re: HighPoint 374

in

The chip on the SATA controller is the same as the PATA version. The SATA interface is handled by external bridge chips. I don't really see how it would make any difference if the hpt374 is on a PCI board on on the motherboard. The Alpha system is quite different from your average PC in some ways, though. That could explain some differences.

Måns Rullgård
[email protected]

Tomi,

I also copied you in my email to Carlos. But my email got bounced.

You might be interested in the problems with Seagate 120 GB drive thread.

Multiple controllers and hard drive mentioned.

Cheers,
Dave

I have an Epox 8k5a3+ with 2 Maxtor drives (80GB) connected. It't running fine, from 2.4.18-ac* kernels (currently 2.4.23-pre7-pac2). I haven't seen any problems, the disks are mirrored (md driver), ext3, reiserfs, xfs and jfs filesystems are on the disks. (I have very bad experiences with HTP372 and 80GB Maxtor drives)

Geller Sandor <[email protected]>

The disks are both 80 GB Western Digital 7200RPM/8 MB cache (WD800JB). FWIW, the motherboard is an Abit IT7-Max 2 v. 2.0, and I'm currently using 2.4.22-ck2. The HPT374 chip is on the motherboard. I don't have any ACPI stuff enabled, and I'm not sure what apic is.

Here's my /proc/interrupts:

CPU0

  0:  490407020          XT-PIC  timer
  1:     160679          XT-PIC  keyboard
  2:          0          XT-PIC  cascade
  4:          0          XT-PIC  usb-uhci
  5:   31664356          XT-PIC  eth1, Intel 82801DB-ICH4
  7:   35088687          XT-PIC  usb-uhci, nvidia
  9:          0          XT-PIC  usb-uhci
 11:    1895235          XT-PIC  ide2, ide3
 12:    4340909          XT-PIC  PS/2 Mouse
 15:     117429          XT-PIC  ide1
NMI:          0
ERR:          0

HTH,

Joel Smith

I get a "no such device" when I try to load the driver, what card exactly do you have and how do you load the driver? Any special parameter?

----- Original Message -----
From: "Joel Smith" <[email protected]>
To: <[email protected]>
Sent: Sunday, October 19, 2003 17:59
Subject: Re: HighPoint 374

kernel

I don't use modules for stuff like this. Have you tried compiling it into the kernel?

Joel Smith