Gahhh, I had a nice long thing typed up, but then I accidently hit the little X to close the tab in Safari. Nice going Apple, let's stick the X DIRECTLY next to another tab so that it's easy to accidently click when switching tabs!
Anyway, on with the retyping of this.
A few months back I setup a working Gentoo dual boot with Windows, running 2.4.20-gentoo-r7. For that install, I had to manually download and compile drivers off of the Highpoint website for my hpt374 card (onboard my IT7). It worked fine afterwards, showing up as a SCSI device.
This weekend, after going neglected for quite a few months due to other circumstances, I decided to get this Gentoo install completely up to date. Many, upon MANY days of compiling later, I had all software up to date, and after another day of configuring, I had it booting off of the Gentoo-Dev-Sources-2.6.5 fine. However, if I either build support for the hpt374 into the kernel or have it compiled as a module, I'm missing some partitions.
It loads my RAID0 array as /dev/hdg, with two partitions: /dev/hdg1 and /dev/hdg2. /dev/hdg1 mounts fine as an NTFS partition, as it's just my normal dedicated Windows swap partition. I can even see the 2gb pagefile fine.
/dev/hdg2, however, shouldn't even be there. fdisk /dev/hdg2 shows it as what it's supposed to be- the Extended partition that's supposed to hold another three partitions. However, fdisk doesn't see those partitions, just the giant extended one. Additionally, Windows shows the three partitions fine (two being FAT32, one being NTFS) and chkdsk /F /R /D on each drive produces no errors. Partition Magic doesn't even show any errors.
So I figured it must be an error with the hpt366 module (which contains the hpt374 module) and I proceeded to download updated source code from Highpoint's website that supports 2.6 kernels. However, after compiling without errors, I attempt to insmod it and I get a "-1 Unknown symbol in module" error. No matter what options I throw at it, I keep getting that error. I have v3.0 of modules-init-tools off of emerge, and that's supposed to support 2.6. But no matter what, I just don't know what I can do to get these partitions to show up. Does anyone have any ideas?
Gentoo-Dev-Sources-2.6.5 + hpt374 module = no love
Message
Having some added time, I took a bit of a closer look at what's going on. My RAID0 array on my hpt374 controller is 2x WD 40gb 7200rpm 2mb cache drives to form an 80gb RAID0 array. It seems as though the built in hpt374 module is reading each drive seperately and putting each drive in /dev as such. Using fdisk on both /dev/hd(e or g) or /dev/discs/disc(1 or 2)/disc produces the same output. For /dev/hde:
And for /dev/hdg, it outputs this when I first open the program:
And this when I hit p:
I'm completely at a loss for why the hpt374 module is doing this. I'm sure Highpoint's 374 module would work... but it's just not compiling successfully. Does anyone have any ideas for this?
I'm tempted to just hit "w" on fdisk to see if it fixes it, but I REALLY don't want to lose all of my data. I don't have enough space right now to back everything up that's on the RAID array (yes, I know RAID0 is fragile anyway... hasn't failed me yet in a year of service *knocks on wood) so I can't afford to just hit on the off chance it messes up my extended partition.
Does anyone know of a partition scanning/fixing tool I could use perhaps? Although I don't know if that would even work, considering it's showing as two drives instead of one.
Code: Select all
Disk /dev/hde: 40.0 GB, 40020664320 bytes
16 heads, 63 sectors/track, 77545 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes
Device Boot Start End Blocks Id System
Code: Select all
Warning: ignoring extra data in partition table 5
Warning: ignoring extra data in partition table 5
Warning: ignoring extra data in partition table 5
Warning: invalid flag 0x7f57 of partition table 5 will be corrected by w(rite)
Code: Select all
Disk /dev/hdg: 40.0 GB, 40020664320 bytes
255 heads, 63 sectors/track, 4865 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/hdg1 1 262 2104483+ 7 HPFS/NTFS
/dev/hdg2 263 9731 76059742+ f W95 Ext'd (LBA)
/dev/hdg5 ? 89379 89379 0+ 1 FAT12
I'm tempted to just hit "w" on fdisk to see if it fixes it, but I REALLY don't want to lose all of my data. I don't have enough space right now to back everything up that's on the RAID array (yes, I know RAID0 is fragile anyway... hasn't failed me yet in a year of service *knocks on wood) so I can't afford to just hit on the off chance it messes up my extended partition.
Does anyone know of a partition scanning/fixing tool I could use perhaps? Although I don't know if that would even work, considering it's showing as two drives instead of one.
PLEASE SOMEONE HELP ME!!!
Someone out there MUST know something to get this working!
No matter what kernel I try, dmesg always gives me this when I try to insmod the just built Highpoint module:
Please, for the love of god, SOMEONE must know how to fix this problem, this is aggrivating me to no end, especially when google searchs for save_flags and the like point to this being a problem that was supposed to be fixed closed to a year ago!
Someone out there MUST know something to get this working!
No matter what kernel I try, dmesg always gives me this when I try to insmod the just built Highpoint module:
Code: Select all
hpt374: module license 'Proprietary' taints kernel.
hpt374: Unknown symbol cli
hpt374: Unknown symbol restore_flags
hpt374: Unknown symbol save_flags
Well, it appears that no one could help but HighPoint themselves. I emailed them, and after a few exchanges, they said it was a bug in their software, and to use the attached patch for it. Unfortunately, they forgot to attach the file and it was Friday when they did it, so I can't get the patch until Monday.
