[SOLVED] hdparm - How to handle hardware 'write cache' properly?

BeachHead · 07-31-2013, 02:06 PM

Hallo,
i'm running a debian based system on a CompactFlash card. I've enabled hardware write cache with 'hdparm -W1 /dev/sda' in 'rc.local'.

Do i have to manually handle the device write cache at shutdown as well? Something like 'hdparm -f -F -W0 /dev/sda' in shutdown script or is it already covered by something else ('ATA reset' or something...).

I'm not 100% sure but i don't think it's covered by 'sync'.

hdparm -iI /dev/sda

Code:

/dev/sda:

 Model=TS16GCF133, FwRev=20110407, SerialNo=2011xxxx    CF13xxxx
 Config={ HardSect NotMFM Removeable DTR>10Mbs nonMagnetic }
 RawCHS=16383/15/63, TrkSize=0, SectSize=576, ECCbytes=4
 BuffType=DualPort, BuffSize=1kB, MaxMultSect=1, MultSect=off
 CurCHS=16383/15/63, CurSects=15481935, LBA=yes, LBAsects=31326208
 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 
 AdvancedPM=yes: disabled (255) WriteCache=enabled
 Drive conforms to: Unspecified:  ATA/ATAPI-7

 * signifies the current active mode


CompactFlash ATA device
	Model Number:       TS16GCF133                              
	Serial Number:      2011xxxx    CF13xxxx
	Firmware Revision:  20110407
Standards:
	Supported: 7 
	Likely used: 7
Configuration:
	Logical		max	current
	cylinders	16383	16383
	heads		15	15
	sectors/track	63	63
	--
	CHS current addressable sectors:   15481935
	LBA    user addressable sectors:   31326208
	Logical/Physical Sector size:           512 bytes
	device size with M = 1024*1024:       15296 MBytes
	device size with M = 1000*1000:       16039 MBytes (16 GB)
	cache/buffer size  = 1 KBytes (type=DualPort)
Capabilities:
	LBA, IORDY(can be disabled)
	Standby timer values: spec'd by Vendor
	R/W multiple sector transfer: Max = 1	Current = 0
	Advanced power management level: disabled
	DMA: mdma0 mdma1 mdma2 udma0 udma1 *udma2 udma3 udma4 
	     Cycle time: min=120ns recommended=120ns
	PIO: pio0 pio1 pio2 pio3 pio4 
	     Cycle time: no flow control=120ns  IORDY flow control=120ns
Commands/features:
	Enabled	Supported:
	   *	SMART feature set
	    	Security Mode feature set
	    	Power Management feature set
	   *	Write cache
	    	WRITE_BUFFER command
	    	READ_BUFFER command
	    	NOP cmd
	   *	CFA feature set
	    	Advanced Power Management feature set
	    	Mandatory FLUSH_CACHE
	   *	SMART self-test
	   *	CFA advanced modes: pio5 pio6 mdma3 mdma4 
	   *	CFA Power Level 1  (max 500mA)
Security: 
	Master password revision code = 65534
		supported
	not	enabled
	not	locked
	not	frozen
	not	expired: security count
	not	supported: enhanced erase
	2min for SECURITY ERASE UNIT. 
HW reset results:
	CBLID- above Vih
	Device num = 0
Integrity word not set (found 0x0000, expected 0xbba5)

mlslk31 · 07-31-2013, 02:44 PM

Sure. Go ahead. While debugging, get a sense of whether the drive is actually sync'ed before the PC shuts off or reboots. On Slackware, there's a section in /etc/rc.d/rc.6 where there's a sync followed by a sleep command. I tuck an `/sbin/hdparm -F /dev/sda` between those two statements. It's worked well for years, but that's for a spinning drive where it's 100% obvious when the `hdparm -F /dev/sda` is taking effect (rattling sound, drive light on, etc.). You'll be looking for similar statements in your shutdown scripts.

BeachHead · 07-31-2013, 03:34 PM

Yep, i'm on identical environment (rc0/shutdown, rc6/restart). I will place the command there. Shouldn't harm.