G
Guest
Guest
Archived from groups: microsoft.public.windowsxp.hardware (More info?)
I have noticed a problem on the newer Intel chipsets. They all seem to
give ATAPI timeout errors in the event viewer (see Problem #1 below).
This could be an Intel chipset problem, MS Win XP Driver problem, or
something else. As I see it on different systems with different hard
drives, I think it is one of the first two. It also causes serious
problems when using camera grab boards (Problems #2 and #3 below).
Has anyone seen this problem elsewhere and does anyone know what is
causing it?
Problem Definition
It appears that there is a problem on the Intel 875, 865 and 915
chipsets running P4 processors. Other similar chipsets may have the same
problems. The problem seems to be missing DMA done interrupts. The
failure rate seems to be quite low, only several times a day. Different
motherboard and PCI card combinations seem to fail at different rates.
Hyper-threading doesn't appear to be part of the problem because
disabling it in the BIOS doesn't help. An older 845 chipset doesn't show
Problem #1 and has not been tested for the other problems.
Problem #1, Hard Disk or CD-ROM Timeouts
Windows XP Event Viewer, System Folder
Type = Error
Source = atapi
Category = None
Event = 9
Message = The device, \Device\Ide\IdePort0, did not respond within the
timeout period.
Frequency of errors on 915 chipset, Windows Server 2003:
12/16 at 4:21am, 6:37am
12/15 at 2:42am, 5:18am, 9:36am, 6:46pm, 7:09pm
12/14 at 1:57pm, 4:59pm, 6:27pm
12/13 at 12:56pm
12/12 at 2:48am, 10:20am, 6:21pm
Different brands of hard drive don't seem to change the problem (Western
Digital on PATA or Seagate on SATA).
Problem #2, Matrox Meteor-II Grab Board Hangs
This grab board allows the program to queue up image buffers for DMA
input into memory. Once one is input, a callback is issued to the
program which then processes it and queues it again. Thus, a circular
queue of buffers are continuously queued and being processed. For some
reason, the callback fails and the whole process hangs. This can happen
several times a day. Several different ways of operating the board were
tried and they all show the failure mode.
Problem #3, BitFlow R3 Grab Board Missing Events
This grab board allows a set of image buffers to be chained into its DMA
engine which then inputs them into memory in sequence when the start
scanning command is issued. At the end of each buffer input, the board
signals the buffer is done using a callback. Every so often these
callbacks seem to not occur and the buffer processing gets progressively
behind the input operation.
I have noticed a problem on the newer Intel chipsets. They all seem to
give ATAPI timeout errors in the event viewer (see Problem #1 below).
This could be an Intel chipset problem, MS Win XP Driver problem, or
something else. As I see it on different systems with different hard
drives, I think it is one of the first two. It also causes serious
problems when using camera grab boards (Problems #2 and #3 below).
Has anyone seen this problem elsewhere and does anyone know what is
causing it?
Problem Definition
It appears that there is a problem on the Intel 875, 865 and 915
chipsets running P4 processors. Other similar chipsets may have the same
problems. The problem seems to be missing DMA done interrupts. The
failure rate seems to be quite low, only several times a day. Different
motherboard and PCI card combinations seem to fail at different rates.
Hyper-threading doesn't appear to be part of the problem because
disabling it in the BIOS doesn't help. An older 845 chipset doesn't show
Problem #1 and has not been tested for the other problems.
Problem #1, Hard Disk or CD-ROM Timeouts
Windows XP Event Viewer, System Folder
Type = Error
Source = atapi
Category = None
Event = 9
Message = The device, \Device\Ide\IdePort0, did not respond within the
timeout period.
Frequency of errors on 915 chipset, Windows Server 2003:
12/16 at 4:21am, 6:37am
12/15 at 2:42am, 5:18am, 9:36am, 6:46pm, 7:09pm
12/14 at 1:57pm, 4:59pm, 6:27pm
12/13 at 12:56pm
12/12 at 2:48am, 10:20am, 6:21pm
Different brands of hard drive don't seem to change the problem (Western
Digital on PATA or Seagate on SATA).
Problem #2, Matrox Meteor-II Grab Board Hangs
This grab board allows the program to queue up image buffers for DMA
input into memory. Once one is input, a callback is issued to the
program which then processes it and queues it again. Thus, a circular
queue of buffers are continuously queued and being processed. For some
reason, the callback fails and the whole process hangs. This can happen
several times a day. Several different ways of operating the board were
tried and they all show the failure mode.
Problem #3, BitFlow R3 Grab Board Missing Events
This grab board allows a set of image buffers to be chained into its DMA
engine which then inputs them into memory in sequence when the start
scanning command is issued. At the end of each buffer input, the board
signals the buffer is done using a callback. Every so often these
callbacks seem to not occur and the buffer processing gets progressively
behind the input operation.