Skip to contentSkip to conversation

3-Ware Raid driver fails to update GenDisk head

DISCUSSION Jeff V. Merkey 3 participants 7 messages Apr 23, 2001

lore
Share & tools
lore
#0 lore raw

Andre/LKML,

I am still working on this, but would appreciate some help from whomever owns this driver proper. I have discovered that the 3Ware drivers are not updating the gendisk_head with devices reported and exposed to user space as /dev/sda, sdb, etc. The adapter driver does correctly report devices to user space, but is not properly reflecting these devices in the gendisk_head, which makes it rather hard for kernel level drivers to scan and determine which devices are attached to the controller.

I attempted to correct this in the code, but found that these changes are a very significant effort to get working.

Please advise.

Jeff

#1 lore raw
Quoted 5 lines · Jeff V. Merkey

I am still working on this, but would appreciate some help from whomever owns this driver proper. I have discovered that the 3Ware drivers are not updating the gendisk_head with devices reported and exposed to user space as /dev/sda, sdb, etc.

But that is the job of sd.c, not of a driver.

#2 lore raw
In reply to Jeff V. Merkey
Quoted 8 lines · Guest section DW
Quoted 5 lines · Jeff V. Merkey

I am still working on this, but would appreciate some help from whomever owns this driver proper. I have discovered that the 3Ware drivers are not updating the gendisk_head with devices reported and exposed to user space as /dev/sda, sdb, etc.

But that is the job of sd.c, not of a driver.

These drivers are an IDE driver that simulates a SCSI interface. It reported IDE devices as SCSI handles, so there's some holes. I guess you were not aware of this, or you would have known that standard sd.c is not working.

:-)

Jeff

#3 lore raw
In reply to Jeff V. Merkey
Quoted 8 lines · Jeff V. Merkey
Quoted 7 lines · Guest section DW
Quoted 5 lines · Jeff V. Merkey

I am still working on this, but would appreciate some help from whomever owns this driver proper. I have discovered that the 3Ware drivers are not updating the gendisk_head with devices reported and exposed to user space as /dev/sda, sdb, etc.

Quoted 6 lines
Quoted 1 line

But that is the job of sd.c, not of a driver.

These drivers are an IDE driver that simulates a SCSI interface. It reported IDE devices as SCSI handles, so there's some holes. I guess you were not aware of this, or you would have known that standard sd.c is not working.

Just like ide-scsi.c you mean?

#4 lore raw
In reply to Jeff V. Merkey
Quoted 9 lines · Guest section DW
Quoted 8 lines · Jeff V. Merkey
Quoted 7 lines · Guest section DW
Quoted 5 lines · Jeff V. Merkey

I am still working on this, but would appreciate some help from whomever owns this driver proper. I have discovered that the 3Ware drivers are not updating the gendisk_head with devices reported and exposed to user space as /dev/sda, sdb, etc.

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

But that is the job of sd.c, not of a driver.

These drivers are an IDE driver that simulates a SCSI interface. It reported IDE devices as SCSI handles, so there's some holes. I guess you were not aware of this, or you would have known that standard sd.c is not working.

Just like ide-scsi.c you mean?

No. They're not that clean and well organized, though they are rather clever adapters and are pretty cool.

Jeff

#5 lore raw
In reply to Jeff V. Merkey
Quoted 17 lines · Jeff V. Merkey
Quoted 16 lines · Guest section DW
Quoted 14 lines · Jeff V. Merkey
Quoted 8 lines · Guest section DW
Quoted 5 lines · Jeff V. Merkey

I am still working on this, but would appreciate some help from whomever owns this driver proper. I have discovered that the 3Ware drivers are not updating the gendisk_head with devices reported and exposed to user space as /dev/sda, sdb, etc.

But that is the job of sd.c, not of a driver.

These drivers are an IDE driver that simulates a SCSI interface. It reported IDE devices as SCSI handles, so there's some holes. I guess you were not aware of this, or you would have known that standard sd.c is not working.

Quoted 4 lines
Quoted 1 line

Just like ide-scsi.c you mean?

No. They're not that clean and well organized, though they are rather clever adapters and are pretty cool.

Come on Jeff, what Andries is saying is that it _works_ like ide-scsi. It registers itself as a SCSI hba like a "regular" SCSI card, hence it uses the SCSI infrastructure -- alas, disks driven by sd.c

Signature · 1 line
Jens Axboe
#6 lore raw
In reply to Jeff V. Merkey
Quoted 18 lines · Jens Axboe
Quoted 17 lines · Jeff V. Merkey
Quoted 16 lines · Guest section DW
Quoted 14 lines · Jeff V. Merkey
Quoted 8 lines · Guest section DW
Quoted 5 lines · Jeff V. Merkey

I am still working on this, but would appreciate some help from whomever owns this driver proper. I have discovered that the 3Ware drivers are not updating the gendisk_head with devices reported and exposed to user space as /dev/sda, sdb, etc.

But that is the job of sd.c, not of a driver.

These drivers are an IDE driver that simulates a SCSI interface. It reported IDE devices as SCSI handles, so there's some holes. I guess you were not aware of this, or you would have known that standard sd.c is not working.

Quoted 8 lines
Quoted 4 lines
Quoted 1 line

Just like ide-scsi.c you mean?

No. They're not that clean and well organized, though they are rather clever adapters and are pretty cool.

Come on Jeff, what Andries is saying is that it _works_ like ide-scsi. It registers itself as a SCSI hba like a "regular" SCSI card, hence it uses the SCSI infrastructure -- alas, disks driven by sd.c

OK, OK. I get it. Sorry for the digs ....

:-)

Jeff

Quoted 3 lines
Signature · 2 lines
-- 
Jens Axboe