* DMA blues...System lockup on setting DMA mode using hdparam
@ 2001-02-23 20:48 Jasmeet Sidhu
2001-02-23 21:59 ` Joel Jaeggli
2001-02-24 9:21 ` Vojtech Pavlik
0 siblings, 2 replies; 3+ messages in thread
From: Jasmeet Sidhu @ 2001-02-23 20:48 UTC (permalink / raw)
To: linux-kernel
Hey guys,
I have five Promise ATA100 controllers configured using kernel version
2.4.2-ac1 (using pdc202xx drivers of course) on ASUS A7V with a AMD Tbird
1GHz processor. Now for the most part this kernel is very stable. I have
premium cables connected to the hard drives and all drives in the system
are masters as you can probably tell by the drive device letters
assigned. The cables are 80pin UDMA (100% Data Integrity). I have not
seen any CRC errors, in fact the system has been up overnight and has
almost transferred about 105GB of data in various file sizes.
The problem:
when trying to set the DMA mode on the drives, using "hdparm -X69
/dev/hda", it works fine. As a matter of fact, this command succeeds for
the following devices:
/dev/hda, /dev/hdc, /dev/hdm, /dev/hdo, /dev/hdq, /dev/hds
However, the system locks up completely when I try the same exact command
for *any* of the following devices: /dev/hde, /dev/hdg, /dev/hdi, /dev/hdk.
*NOTE* the Raid5 array /dev/md0 is not running when I am trying to set the
DMA modes. The raid is not mounted and is in stopped mode using raidstop
/dev/md0.
Also, when I try and use the -k and the -K switches (keep settings after
reset), the programs says that it worked. However, after I restart the
system, these "flags" are set to 0 again. Is this normal? In other words:
hdparam -k /dev/hda
keepsettings = 0 (off)
# now lets set the -k option (keep settings after refresh).
hdparam -k1 /dev/hda
setting keep_settings to 1 (on)
keepsettings = 1 (on)
# noe lets restart the system and query again
hdparam -k /dev/hda
keepsettings = 0 (off)
Is this normal?
Also another question related to IDE:
Is there anyway we can see how good/bad the system performance is while
the system is working? I am not talking about a benchmarking tool like
bonnie that simply tries to figure out how good a system can perform. I
would like to see something, maybe in /proc/ide/, that shows me the current
throughput of the ide subsystem. For example how many kb of data is going
in, how much coming out of each device. Any ideas on how to go about maybe
adding this? Where would be an ideal place to add such functionality? In
the ide code or maybe in the raid section? Or maybe these two should be
kept separate. Any thoughts guys?
Any additional required information can be posted, let me know.
Anybody else out there with a similar situation? Your thoughts on this
would be really appreciated.
Here's the setup:
ide0 at 0x3800-0x3807,0x3402 on irq 11 PDC20265
ide1 at 0x3000-0x3007,0x2802 on irq 11
ide2 at 0x5400-0x5407,0x5002 on irq 15 PDC20267
ide3 at 0x4800-0x4807,0x4402 on irq 15
ide4 at 0x7000-0x7007,0x6802 on irq 11 PDC20267
ide5 at 0x6400-0x6407,0x6002 on irq 11
ide6 at 0x8800-0x8807,0x8402 on irq 14 PDC20267
ide7 at 0x8000-0x8007,0x7802 on irq 14
ide8 at 0xa400-0xa407,0xa002 on irq 10 PDC20267
ide9 at 0x9800-0x9807,0x9402 on irq 10
hda: 40188960 sectors (20577 MB) w/1916KiB Cache, CHS=39870/16/63, UDMA(100)
hdc: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hde: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdg: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdi: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdk: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdm: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdo: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdq: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hds: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
# Raid-5 configuration
#
raiddev /dev/md0
raid-level 5
chunk-size 4
parity-algorithm left-symmetric
persistent-superblock 1
nr-raid-disks 8
nr-spare-disks 1
device /dev/hde1
raid-disk 0
device /dev/hdg1
raid-disk 1
device /dev/hdi1
raid-disk 2
device /dev/hdk1
raid-disk 3
device /dev/hdm1
raid-disk 4
device /dev/hdo1
raid-disk 5
device /dev/hdq1
raid-disk 6
device /dev/hds1
raid-disk 7
device /dev/hdc1
spare-disk 0
[root@bertha hdparm-3.9]# df -k
Filesystem 1k-blocks Used Available Use% Mounted on
/dev/hda3 19072868 3589612 14514392 20% /
/dev/hda1 198313 11667 176392 7% /boot
/dev/md0 525461076 108657156 416803920 21% /raid
sj-f760-1:/vol/vol03/data01
142993408 140675648 2317760 99% /mnt/netapps
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: DMA blues...System lockup on setting DMA mode using hdparam
2001-02-23 20:48 DMA blues...System lockup on setting DMA mode using hdparam Jasmeet Sidhu
@ 2001-02-23 21:59 ` Joel Jaeggli
2001-02-24 9:21 ` Vojtech Pavlik
1 sibling, 0 replies; 3+ messages in thread
From: Joel Jaeggli @ 2001-02-23 21:59 UTC (permalink / raw)
To: Jasmeet Sidhu; +Cc: linux-kernel
On Fri, 23 Feb 2001, Jasmeet Sidhu wrote:
> Also another question related to IDE:
> Is there anyway we can see how good/bad the system performance is while
> the system is working?
that information (what the disk subsytem as a whole is doing) is collected
in /proc/stat something like xosview can help you visualize it.
> I am not talking about a benchmarking tool like
> bonnie that simply tries to figure out how good a system can perform. I
> would like to see something, maybe in /proc/ide/, that shows me the current
> throughput of the ide subsystem. For example how many kb of data is going
> in, how much coming out of each device. Any ideas on how to go about maybe
> adding this? Where would be an ideal place to add such functionality? In
> the ide code or maybe in the raid section? Or maybe these two should be
> kept separate. Any thoughts guys?
>
> Any additional required information can be posted, let me know.
>
> Anybody else out there with a similar situation? Your thoughts on this
> would be really appreciated.
I actually have a similar situation although without hanging, with three
controllers. the kernels detects the two devices on the third controller
as udma rather than udma(100) with 2.4.1.
hda: 30003120 sectors (15362 MB) w/1916KiB Cache, CHS=1867/255/63, UDMA(33)
hdc: 30003120 sectors (15362 MB) w/1916KiB Cache, CHS=29765/16/63, UDMA(33)
first two are on the server-works chipset controller
hde: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdg: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdi: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdk: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
hdm: 30003120 sectors (15362 MB) w/1916KiB Cache, CHS=29765/16/63, (U)DMA
hdo: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, (U)DMA
these 6 are on three promise contollers... all with 18" 80pin cable
hdd: ATAPI 40X CD-ROM drive, 128kB Cache, UDMA(33)
> Here's the setup:
>
> ide0 at 0x3800-0x3807,0x3402 on irq 11 PDC20265
> ide1 at 0x3000-0x3007,0x2802 on irq 11
> ide2 at 0x5400-0x5407,0x5002 on irq 15 PDC20267
> ide3 at 0x4800-0x4807,0x4402 on irq 15
> ide4 at 0x7000-0x7007,0x6802 on irq 11 PDC20267
> ide5 at 0x6400-0x6407,0x6002 on irq 11
> ide6 at 0x8800-0x8807,0x8402 on irq 14 PDC20267
> ide7 at 0x8000-0x8007,0x7802 on irq 14
> ide8 at 0xa400-0xa407,0xa002 on irq 10 PDC20267
> ide9 at 0x9800-0x9807,0x9402 on irq 10
>
> hda: 40188960 sectors (20577 MB) w/1916KiB Cache, CHS=39870/16/63, UDMA(100)
> hdc: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
> hde: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
> hdg: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
> hdi: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
> hdk: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
> hdm: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
> hdo: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
> hdq: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
> hds: 150136560 sectors (76870 MB) w/1916KiB Cache, CHS=148945/16/63, UDMA(100)
>
> # Raid-5 configuration
> #
> raiddev /dev/md0
> raid-level 5
> chunk-size 4
> parity-algorithm left-symmetric
> persistent-superblock 1
> nr-raid-disks 8
> nr-spare-disks 1
> device /dev/hde1
> raid-disk 0
> device /dev/hdg1
> raid-disk 1
> device /dev/hdi1
> raid-disk 2
> device /dev/hdk1
> raid-disk 3
> device /dev/hdm1
> raid-disk 4
> device /dev/hdo1
> raid-disk 5
> device /dev/hdq1
> raid-disk 6
> device /dev/hds1
> raid-disk 7
> device /dev/hdc1
> spare-disk 0
>
> [root@bertha hdparm-3.9]# df -k
> Filesystem 1k-blocks Used Available Use% Mounted on
> /dev/hda3 19072868 3589612 14514392 20% /
> /dev/hda1 198313 11667 176392 7% /boot
> /dev/md0 525461076 108657156 416803920 21% /raid
> sj-f760-1:/vol/vol03/data01
> 142993408 140675648 2317760 99% /mnt/netapps
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
--
--------------------------------------------------------------------------
Joel Jaeggli joelja@darkwing.uoregon.edu
Academic User Services consult@gladstone.uoregon.edu
PGP Key Fingerprint: 1DE9 8FCA 51FB 4195 B42A 9C32 A30D 121E
--------------------------------------------------------------------------
It is clear that the arm of criticism cannot replace the criticism of
arms. Karl Marx -- Introduction to the critique of Hegel's Philosophy of
the right, 1843.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: DMA blues...System lockup on setting DMA mode using hdparam
2001-02-23 20:48 DMA blues...System lockup on setting DMA mode using hdparam Jasmeet Sidhu
2001-02-23 21:59 ` Joel Jaeggli
@ 2001-02-24 9:21 ` Vojtech Pavlik
1 sibling, 0 replies; 3+ messages in thread
From: Vojtech Pavlik @ 2001-02-24 9:21 UTC (permalink / raw)
To: Jasmeet Sidhu; +Cc: linux-kernel
On Fri, Feb 23, 2001 at 12:48:29PM -0800, Jasmeet Sidhu wrote:
> Also, when I try and use the -k and the -K switches (keep settings after
> reset), the programs says that it worked. However, after I restart the
> system, these "flags" are set to 0 again. Is this normal? In other words:
> hdparam -k /dev/hda
> keepsettings = 0 (off)
> # now lets set the -k option (keep settings after refresh).
> hdparam -k1 /dev/hda
> setting keep_settings to 1 (on)
> keepsettings = 1 (on)
> # noe lets restart the system and query again
> hdparam -k /dev/hda
> keepsettings = 0 (off)
>
> Is this normal?
This only relates to a ide bus reset in case of a failure, not system reset.
--
Vojtech Pavlik
SuSE Labs
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2001-02-24 9:38 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-02-23 20:48 DMA blues...System lockup on setting DMA mode using hdparam Jasmeet Sidhu
2001-02-23 21:59 ` Joel Jaeggli
2001-02-24 9:21 ` Vojtech Pavlik
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).