Kernel Recompiling

[hern_28]

I know this has got to be a repeat post but I can't find resources for kubuntu 8.10 on kernel compiling. Learning to work with kubuntu coming from gentoo and slackware. Have not gotten used to not having everything for building pre-installed yet lol.

What i need to do is recompile the kernel that i'm already using with a couple of minor changes so i can enable dma on my hard drive. Running version vmlinuz-2.6.27-9-generic and just need to move the general ide/sata drivers to modules and my main drivers to built in to do this. Currently running without dma enabled.

How do I get the source and what do I need to install to compile it? Once I have that I can do the rest.

[Snowhog]

Re: Kernel Recompiling

Slightly dated, but: How To Compile A Kernel - The Ubuntu Way
And for the kernel sources:
The Linux Kernel Archives

[dibl]

Re: Kernel Recompiling

You might want to install hdparm and check whether dma is already enabled, before you go kernel-compiling.

[hern_28]

Re: Kernel Recompiling

hdparm was installed by default.. no dice till kernel modification.

[dibl]

Re: Kernel Recompiling

Stock kernel supports max dma speed on my rig:

dibl@ibex:~$ sudo hdparm -i /dev/sdb
[sudo] password for dibl:

/dev/sdb:

 Model=WDC WD1500ADFD-00NLR1         , FwRev=20.07P20, SerialNo=  WD-WMAP41142727
 Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
 RawCHS=16383/16/63, TrkSize=57600, SectSize=600, ECCbytes=48
 BuffType=DualPortCache, BuffSize=16384kB, MaxMultSect=16, MultSect=?16?
 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=293046768
 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes: pio0 pio1 pio2 pio3 pio4
 DMA modes: mdma0 mdma1 mdma2
 UDMA modes: udma0 udma1 udma2 udma3 udma4 udma5 *udma6
 AdvancedPM=no WriteCache=enabled
 Drive conforms to: ATA/ATAPI-7 published, ANSI INCITS 397-2005: ATA/ATAPI-1,2,3,4,5,6,7

 * signifies the current active mode

dibl@ibex:~$ uname -a
Linux ibex 2.6.27-7-generic #1 SMP Tue Nov 4 19:33:06 UTC 2008 x86_64 GNU/Linux

[hern_28]

Re: Kernel Recompiling

try hdparm -d1 (drive) to see if dma was enabled on the drive.

hdparm -i gives :

Model=WDC WD2000JS-60NCB1 , FwRev=10.02E02, SerialNo= WD-WCANL2066908
Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=50
BuffType=unknown, BuffSize=8192kB, MaxMultSect=16, MultSect=?1?
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=390721968
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5
AdvancedPM=no WriteCache=enabled
Drive conforms to: Unspecified: ATA/ATAPI-1,2,3,4,5,6,7

* signifies the current active mode

But... the kernel builders had the built in modules loading in the wrong order so you get :

hdparm -d1 /dev/sda :

/dev/sda:
setting using_dma to 1 (on)
HDIO_SET_DMA failed: Inappropriate ioctl for device
HDIO_GET_DMA failed: Inappropriate ioctl for device


So even though the drive is using udma5... dma is not set on the drive. So at least with the 'sata_nv' and 'nv' drivers, you have to manually compile the generic ide and sata drivers as 'modules' and the nv drivers as 'built' in so the nv drivers load first. Otherwise my video and audio run a little funny, erratic and irregular pauses in video and audio in movies and music. Done this a few times.. it works i promise .


Before recompiling the kernel hdparm :

hdparm -tT /dev/sda

/dev/sda:
Timing cached reads: 2332 MB in 2.00 seconds = 1166.49 MB/sec


And after

/dev/sda:
Timing cached reads: 5527 MB in 2.00 seconds = 2763.5 MB/sec

K now slimming down the new KDE . Thanks for the links.

[dibl]

Re: Kernel Recompiling

That's interesting! Having never been a student of DMA and Ultra DMA, but knowing enough to use Google for "dma vs udma", "udma libata", I made a quick study here and here.

These are important (from the second article):

Why does HDIO_SET_DMA fail? I want to use DMA!
Why does HDIO_SET_UNMASKINTR fail?

libata intentionally does not support all the HDIO_xxx ioctls that were supported by the older IDE driver. It is now preferred to use SG_IO as a generalized ATA command submission method, rather than creating a myriad of ioctls for each specific purpose.

The design decision was made only to support the HDIO_xxx ioctls that were heavily used by other programs. Generally the driver always programs the hardware to its maximum capability automatically, completely without user intervention. Therefore, for example, HDIO_SET_DMA is not needed for the vast majority of users because DMA is automatically enabled and used where available.

and

Linux NCQ (SATA native command queueing) support is enabled automatically, if your SATA drive supports it. At boot, you will see a line in dmesg (or kernel log) like

ata2.00: ATA-7, max UDMA/133, 321672960 sectors: LBA48 NCQ (depth 31/32)

So, I learned it depends a lot on your motherboard and hard disk drive. On mine (Intel D975XBX, Western Digital WD1500s), dmesg shows this:

ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata2.00: ATA-7: WDC WD1500ADFD-00NLR1, 20.07P20, max UDMA/133
ata2.00: 293046768 sectors, multi 16: LBA48 NCQ (depth 31/32)
ata2.00: configured for UDMA/133

So, the stock kernel with libata driver gives me this, for that drive:

dibl@ibex:/home/dibl$ sudo hdparm -tT /dev/sdb

/dev/sdb:
Timing cached reads: 7574 MB in 2.00 seconds = 3789.49 MB/sec
Timing buffered disk reads: 244 MB in 3.01 seconds = 81.00 MB/sec

Which is not exactly 100MB/sec as per the marketing claim, but it's close enough for practical purposes.

[hern_28]

Re: Kernel Recompiling

LoL.. Booted to the old kernel and here is relevant dmesg:

[ 3.882448] ata6: SATA max UDMA/133 cmd 0x960 ctl 0xb60 bmdma 0xc008 irq 23
[ 4.676014] ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 4.684093] ata6.00: ATA-7: WDC WD2000JS-60NCB1, 10.02E02, max UDMA/100
[ 4.684095] ata6.00: 390721968 sectors, multi 1: LBA48 NCQ (depth 0/32)
[ 4.692093] ata6.00: configured for UDMA/100

Note multi 1, depth 0, 16 bit I/O instead of 32. Obviously with these setting no wonder why I'm loosing performance. Guess I have one of those odd chipsets thanks for the links.