Driver DMA/ATA bug on B72 during installation
Hello,
I'm trying to jumpstart install Solaris 10 x86 b72.
My hw: Athlon XP 800mhz, 120gb Hitachi deskstar, etc.
The following message appears when creating the '/' filesystem:
"Creating / (c0d0s0)"
"Please file a ***BUG*** for driver ata, dma_*_bind() will return a failure.
sgllen 17, cookie count=68, and DDI_DMA_PARTIAL is *not* set"
My disk sounds also weird in this phase (copying packages)... like it's parking the heads now and then. Not sure maybe its a hardware problem then... or it could be a bad driver causing this behaviour.
If anyone has a clue, let me know. The install seems to go through though, but I'm afraid it might crash my disk.
// Magnus
[736 byte] By [
snejk] at [2007-11-26 0:43:03]
# 1
Also got this error on AMD Duron 900 Mhz, 120gb Hitachi deskstar.
btw: only got the error during a text mode installation. No error alike during a DT Install, but also a blanc window.
Furtermore, screen is blanc during install.
Indeed cd makes hovering noise, normal for the installation. Also leds for HDD & CD indicate install is going on. The time span alike.
But, after reboot I got an older build reported and a lot of errors, so install is not working fine.
Magnus, what are your results today ?
I think I will report this issue as a Beta 8 bug to the BetaTeam ! Provided the B72 = Beta 8
# 2
Hi,
This popped up in the dmesg output when stresstesting an IDE disk today:
Dec 16 22:08:29 sara rootnex: [ID 137135 kern.warning] WARNING: Please file a ***BUG*** for driver ata, dma_*_bind() will return a failure to ata. sgllen = 17, cookie count = 68, and DDI_DMA_PARTIAL is *not* set.
Some info on the hardware:
Motherboard: AS ROCK K7VT2
Disk: Seagte 120GB IDE 7200RPM 8MB CACHE
uname -a: SunOS sara 5.10 s10_72 i86pc i386 i86pc
# 4
I received the same error....
rootnex: WARNING: Please file a ***BUG*** for driver ata, dma_*_bind() will return a failure to ata. sgllen = 17, cookie count = 68, and DDI_DMA_PARTIAL is *not* set.
I noticed it every time I tried to run the metadb command... which would fail everytime with an "Invalid Argument" message. I knew my syntax was right, but this kept me stumped all night. So -- first of all, thanks to you all for the pointers.
Also note that the Enhanced Storage function in the smc utility would just hang as well.
Now, I added the following line to my /etc/system file:
set rootnex:rootnex_bind_fail = 0
Now, once I rebooted, the metadb command worked and so did smc. That was great, except then I got the following messages on the console:
inetd[231]: Failed to lookup netid 'tcp6' for instance svc:/network/rpc/meta:default: netid not found in /etc/netconfig
inetd[231]: Failed to bind on all protocols for instance
svc:/network/rpc/meta:default, transitioning to degraded
What do these mean? They came up when I first ran the metadb command after the reboot after I added the rootnex line to /etc/system.
Everything seems to be working, but I'm still a little worried. ..
Why didnt I see this problem in Solaris 8? This is the same hardware. The article on yahoo provided in the link in a previous post mentions "silent data corruption" -- how will I know if I am being bit?
Also - the error message says to report a bug -- ok, to who? How?
Thanks in advance to anyone to can help.
