Re: /dev/hd[ijkl] only using udma (not udma 100)

Re: /dev/hd[ijkl] only using udma (not udma 100)

[Tomas Szepe]

> It would realy be nice to find a way to force transfer mode on boot up but
> i cant see to find any way.

I've been using "ide2=ata66" to force ATA66/ATA100 transfer modes
on a Promise Ultra100 TX2 controller.

T.

Re: /dev/hd[ijkl] only using udma (not udma 100)

[Simen Timian Thoresen]

> Hi,
> 
> Ive got a machine running debian test dist and 2.4.18. The machine has two
> promise ata100 tx2 controller cards. My question is why does the devices
> hde to hdh use udma100 but devices hdi to hdl only use udma. Note on this
> is that the devices hdi to hdl did i have to make myself (dont know if
> there is some other configure possibility). All drives are the same model.

I'm basically observing the same thing - 5 drives connected alone on their own 
ata100 channels (1 on a Via southbridge, 4 on HPT370 channels).


I'm seing speeds 33, 44, 66 and 69 as when examining /proc/ide/hdx/settings. 
There are 2 differing types of drives, but all are ATA100 capable.
 
This is also with 2.4.18, originally slackware 7.1

-Simen
--
Simen Thoresen
http://www.dolphinics.no

/dev/hd[ijkl] only using udma (not udma 100)

[me]

Hi,

Ive got a machine running debian test dist and 2.4.18. The machine has two
promise ata100 tx2 controller cards. My question is why does the devices
hde to hdh use udma100 but devices hdi to hdl only use udma. Note on this
is that the devices hdi to hdl did i have to make myself (dont know if
there is some other configure possibility). All drives are the same model.

--- dmesg ---
block: 128 slots per queue, batch=32
Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 33MHz system bus speed for PIO modes; override
with idebus=xx
PDC20268: IDE controller on PCI bus 00 dev 58
PDC20268: chipset revision 2
PDC20268: not 100% native mode: will probe irqs later
PDC20268: (U)DMA Burst Bit DISABLED Primary PCI Mode
Secondary PCI Mode.
PDC20268: FORCING BURST BIT 0x50 -> 0x51 INACTIVE
    ide2: BM-DMA at 0x2050-0x2057, BIOS settings: hde:pio, hdf:pio
    ide3: BM-DMA at 0x2058-0x205f, BIOS settings: hdg:pio, hdh:pio
PDC20268: IDE controller on PCI bus 00 dev 68
PDC20268: chipset revision 2
PDC20268: not 100% native mode: will probe irqs later
PDC20268: (U)DMA Burst Bit DISABLED Primary PCI Mode
Secondary PCI Mode.
PDC20268: FORCING BURST BIT 0x50 -> 0x51 INACTIVE
    ide4: BM-DMA at 0x2080-0x2087, BIOS settings: hdi:pio, hdj:pio
    ide5: BM-DMA at 0x2088-0x208f, BIOS settings: hdk:pio, hdl:pio
PIIX4: IDE controller on PCI bus 00 dev 91
PIIX4: chipset revision 1
PIIX4: not 100% native mode: will probe irqs later
    ide0: BM-DMA at 0x2040-0x2047, BIOS settings: hda:DMA, hdb:DMA
    ide1: BM-DMA at 0x2048-0x204f, BIOS settings: hdc:DMA, hdd:pio
hda: *****************, ATA DISK drive
hdc: ***********************, ATAPI CD/DVD-ROM drive
hde: ****************, ATA DISK drive
hdf: ****************, ATA DISK drive
hdh: ****************, ATA DISK drive
hdi: ****************, ATA DISK drive
hdj: ****************, ATA DISK drive
hdk: ****************, ATA DISK drive
hdl: ****************, ATA DISK drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
ide2 at 0x2070-0x2077,0x2066 on irq 18
ide3 at 0x2068-0x206f,0x2062 on irq 18
ide4 at 0x2098-0x209f,0x207e on irq 17
ide5 at 0x2090-0x2097,0x207a on irq 17
hda: ******** sectors (***** MB) w/512KiB Cache, CHS=****/***/**, UDMA(33)
hde: ********* sectors (****** MB) w/****KiB Cache, CHS=******/**/**,
UDMA(100)
hdf: ********* sectors (****** MB) w/****KiB Cache, CHS=******/**/**,
UDMA(100)
hdh: ********* sectors (****** MB) w/****KiB Cache, CHS=******/**/**,
UDMA(100)
hdi: ********* sectors (****** MB) w/****KiB Cache, CHS=******/**/**,
(U)DMA
hdj: ********* sectors (****** MB) w/****KiB Cache, CHS=******/**/**,
(U)DMA
hdk: ********* sectors (****** MB) w/****KiB Cache, CHS=******/**/**,
(U)DMA
hdl: ********* sectors (****** MB) w/****KiB Cache, CHS=******/**/**,
(U)DMA
hdc: ATAPI 48X CD-ROM drive, 128kB Cache, UDMA(33)
--- dmesg ---

--- hdparm test ---
# hdparm -t -T /dev/hd[efhijkl]

/dev/hde:
 Timing buffer-cache reads:   128 MB in  0.93 seconds =137.63 MB/sec
 Timing buffered disk reads:  64 MB in  1.82 seconds = 35.16 MB/sec

/dev/hdf:
 Timing buffer-cache reads:   128 MB in  0.94 seconds =136.17 MB/sec
 Timing buffered disk reads:  64 MB in  1.40 seconds = 45.71 MB/sec

/dev/hdh:
 Timing buffer-cache reads:   128 MB in  0.95 seconds =134.74 MB/sec
 Timing buffered disk reads:  64 MB in  1.82 seconds = 35.16 MB/sec

/dev/hdi:
 Timing buffer-cache reads:   128 MB in  0.94 seconds =136.17 MB/sec
 Timing buffered disk reads:  64 MB in  4.16 seconds = 15.38 MB/sec

/dev/hdj:
 Timing buffer-cache reads:   128 MB in  0.94 seconds =136.17 MB/sec
 Timing buffered disk reads:  64 MB in  4.17 seconds = 15.35 MB/sec

/dev/hdk:
 Timing buffer-cache reads:   128 MB in  0.93 seconds =137.63 MB/sec
 Timing buffered disk reads:  64 MB in  4.16 seconds = 15.38 MB/sec

/dev/hdl:
 Timing buffer-cache reads:   128 MB in  0.94 seconds =136.17 MB/sec
 Timing buffered disk reads:  64 MB in  4.17 seconds = 15.35 MB/sec
--- hdparm test ---

As you see the speed on the other devices is less then half.

Please include me in the response mails because im not on the list (not
from this account anyway).

Thanks to anyone that can help me! =) If you need more info, just drop me
a mail.

Re: /dev/hd[ijkl] only using udma (not udma 100)

[me]

On Mon, 27 May 2002, Simen Timian Thoresen wrote:

> > Hi,
> > 
> > Ive got a machine running debian test dist and 2.4.18. The machine has two
> > promise ata100 tx2 controller cards. My question is why does the devices
> > hde to hdh use udma100 but devices hdi to hdl only use udma. Note on this
> > is that the devices hdi to hdl did i have to make myself (dont know if
> > there is some other configure possibility). All drives are the same model.
> 
> I'm basically observing the same thing - 5 drives connected alone on their own 
> ata100 channels (1 on a Via southbridge, 4 on HPT370 channels).
> 
> 
> I'm seing speeds 33, 44, 66 and 69 as when examining /proc/ide/hdx/settings. 
> There are 2 differing types of drives, but all are ATA100 capable.
>  
> This is also with 2.4.18, originally slackware 7.1
> 

I booted with:
linux ide0=autotune ide1=autotune ide2=autotune ide3=autotune

and now all drives are running just (U)DMA and speeds on the 3 first
droped to 15m/s.

The settings file you gave here, is there some possibility to configure
settings there? I see things like current_speed and init_speed that would
be nice to try and tweak.

Re: /dev/hd[ijkl] only using udma (not udma 100)

[me]

On Mon, 27 May 2002, Simen Timian Thoresen wrote:

> > I booted with:
> > linux ide0=autotune ide1=autotune ide2=autotune ide3=autotune
> > 
> > and now all drives are running just (U)DMA and speeds on the 3 first
> > droped to 15m/s.
> > 
> > The settings file you gave here, is there some possibility to configure
> > settings there? I see things like current_speed and init_speed that would
> > be nice to try and tweak.
> > 
> I believe that file is read only, and that the changes are made with hdparm or 
> autotune.
> 
> I have not tried to change anything yet (I first noticed this yesterday while 
> replacing disks in a raid), but I'll have a go with hdparm when I get home and 
> see what I can shake up.
> 
> I believe you can set trasfer-mode directly using hdparm -x<mode>, but I 
> don't see why the kernel should not set this by default.
> 

Well, as i read the ide .c i see that preconfigureation (if you can call
it that, more like what drives/interfaces do we assume is there), i only
see that for ide0-3. Now ide0-1 is onboard and ide1-2 is the first ide
card but nothing is made with ide4-5 which is the second ide card and i
think that the problem.

It would realy be nice to find a way to force transfer mode on boot up but
i cant see to find any way.

Re: /dev/hd[ijkl] only using udma (not udma 100)

[me]

On Mon, 27 May 2002, Tomas Szepe wrote:

> > It would realy be nice to find a way to force transfer mode on boot up but
> > i cant see to find any way.
> 
> I've been using "ide2=ata66" to force ATA66/ATA100 transfer modes
> on a Promise Ultra100 TX2 controller.
> 

I might have worked for ide2 and 3 but it didnt work for ide4 and 5.

But hdparm -X69 /dev works just fine, just bad that it isnt at boot up but
I will set it as soon as possible.

-X69 is UDMA 5 mode, funny thing is i can set it to UDMA 6 =) but there is
no speed change, just funny that the ide card takes that setting.

Thanks for all your help, this solution will cover my needs for the
moment.

Re: /dev/hd[ijkl] only using udma (not udma 100)

[me]

On Mon, 27 May 2002, Thunder from the hill wrote:

> Hi,
> 
> On Mon, 27 May 2002 me@vger.org wrote:
> > Ive got a machine running debian test dist and 2.4.18. The machine has
> > two promise ata100 tx2 controller cards. My question is why does the
> > devices hde to hdh use udma100 but devices hdi to hdl only use udma.
> > Note on this is that the devices hdi to hdl did i have to make myself
> > (dont know if there is some other configure possibility). All drives are
> > the same model.
> 
> Well, you found the /sbin/hdparm solution, which isn't too bad (And 
> hopefully you added it to your boot scripts). However, watch that! If the 
> mode switches back to lower UDMA modes, it doesn't support the mode 
> properly and thus was downgraded.
> 

Only thing i get, while i was stresstesting the hardware, is that DMA gets
timeouts time from time and then the DMA gets disabled (logical
realy). But nothing happens with the transfer mode and thats good.

Re: /dev/hd[ijkl] only using udma (not udma 100)

[me]

Upgraded to 2.4.19-pre9 and now everything work with the Promise ATA100
TX2 card when setting idex=ata66 on boot up... no need to run hdparm =)

Thanks all for the help.