Fsys.sym8scsi Mar19, 1999 -> 4.24Q Changes??

Hi,

I posted this on 2000.09.12 and did not hear back from QSSL. I thought maybe
once the RTP was released someone would come back and examine this issue but
I haven’t seen anything yet. How’s about it??

Here’s the original message:
+++++++++++++++++++++++++++

I am a little late in installing 4.25D but better late then never.

We have a number of Compaq Proliant 1600 that have dual channel (875
chipset) embedded scsi. Using the pre-release version of the Fsys.sym8scsi
driver date Mar 19, 1999 we started the driver in the boot image with just
the “-L” option and it worked fine. It would find devices on both channels
and if they existed and would provide access to these devices (only 1 CPU
ever had a device on the 2 channel).

Now with the driver from the 4.25D patch this no longer works. What happens
is that the driver loads finds a device on channel 1 (good) and then
searches channel 2 for a device and never comes back if no devices are
present (bad). To get around the problem we now have to specify the PCI
index to ensure the unused channel is not used by the driver. This is not so
bad but it now makes my configurations less consistent because I have a
machine that will require different options for that driver. I am just
wondering why this changed?

A few problems I found when trying to use the PCI index are as follows:

  1. The use message generally indicates that -p is used to specify the PCI
    index but the example header specifies -I.
  2. I found that the -p option was useless without the -c option.

Cheers,

  • Richard

Are there any QSSL Fsys developers following this newsgroup?


“Brown, Richard” <brownr@aecl.ca> wrote in message
news:8qqq2j$nc8$1@inn.qnx.com

Hi,

I posted this on 2000.09.12 and did not hear back from QSSL. I thought
maybe
once the RTP was released someone would come back and examine this issue
but
I haven’t seen anything yet. How’s about it??

Here’s the original message:
+++++++++++++++++++++++++++

I am a little late in installing 4.25D but better late then never.

We have a number of Compaq Proliant 1600 that have dual channel (875
chipset) embedded scsi. Using the pre-release version of the Fsys.sym8scsi
driver date Mar 19, 1999 we started the driver in the boot image with just
the “-L” option and it worked fine. It would find devices on both channels
and if they existed and would provide access to these devices (only 1 CPU
ever had a device on the 2 channel).

Now with the driver from the 4.25D patch this no longer works. What
happens
is that the driver loads finds a device on channel 1 (good) and then
searches channel 2 for a device and never comes back if no devices are
present (bad). To get around the problem we now have to specify the PCI
index to ensure the unused channel is not used by the driver. This is not
so
bad but it now makes my configurations less consistent because I have a
machine that will require different options for that driver. I am just
wondering why this changed?

A few problems I found when trying to use the PCI index are as follows:

  1. The use message generally indicates that -p is used to specify the PCI
    index but the example header specifies -I.
  2. I found that the -p option was useless without the -c option.

Cheers,

  • Richard
    \

I got an e-mail from QSSL this morning that reports they have
reproduced my problem with this driver, which looks to me to be the same
as yours. In other words, there’s hope yet…

Richard

“Brown, Richard” wrote:

Are there any QSSL Fsys developers following this newsgroup?

“Brown, Richard” <> brownr@aecl.ca> > wrote in message
news:8qqq2j$nc8$> 1@inn.qnx.com> …
Hi,

I posted this on 2000.09.12 and did not hear back from QSSL. I thought
maybe
once the RTP was released someone would come back and examine this issue
but
I haven’t seen anything yet. How’s about it??

Here’s the original message:
+++++++++++++++++++++++++++

I am a little late in installing 4.25D but better late then never.

We have a number of Compaq Proliant 1600 that have dual channel (875
chipset) embedded scsi. Using the pre-release version of the Fsys.sym8scsi
driver date Mar 19, 1999 we started the driver in the boot image with just
the “-L” option and it worked fine. It would find devices on both channels
and if they existed and would provide access to these devices (only 1 CPU
ever had a device on the 2 channel).

Now with the driver from the 4.25D patch this no longer works. What
happens
is that the driver loads finds a device on channel 1 (good) and then
searches channel 2 for a device and never comes back if no devices are
present (bad). To get around the problem we now have to specify the PCI
index to ensure the unused channel is not used by the driver. This is not
so
bad but it now makes my configurations less consistent because I have a
machine that will require different options for that driver. I am just
wondering why this changed?

A few problems I found when trying to use the PCI index are as follows:

  1. The use message generally indicates that -p is used to specify the PCI
    index but the example header specifies -I.
  2. I found that the -p option was useless without the -c option.

Cheers,

  • Richard
    \