Hi
I configuring the kernel, there is a option "Build adapter firmware with
Kernel build". There is no help for this. It's obvious that it build
firmware but is it installed in the adapter automagically ???
I also wonder why the reset delay is 15000 Msec. It used to be 5000
Msec. I've usually set it to that without nasty results. I just wonder
what the reasoning is behind such a long delay.
TIA
John
On Thu, Nov 08, 2001 at 12:52:57PM -0500, John Gluck wrote:
> Hi
>
> I configuring the kernel, there is a option "Build adapter firmware with
> Kernel build". There is no help for this. It's obvious that it build
> firmware but is it installed in the adapter automagically ???
IIRC the kernel comes with an already built firmware. This option allows
you to rebuild it from source. You certainly not need it.
>
> I also wonder why the reset delay is 15000 Msec. It used to be 5000
> Msec. I've usually set it to that without nasty results. I just wonder
> what the reasoning is behind such a long delay.
This is a drawback of single driver for multiple cards. Good cards
suffer to enable the driver to support bad ones.
> TIA
>
> John
cab.
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
--
Christophe Barb? <[email protected]>
GnuPG FingerPrint: E0F6 FADF 2A5C F072 6AF8 F67A 8F45 2F1E D72C B41E
>Hi
>
>I configuring the kernel, there is a option "Build adapter firmware with
>Kernel build".
There should be. This comment has been in my distributed patches for
Configure.help for some time:
Build Adapter Firmware with Kernel Build
CONFIG_AIC7XXX_BUILD_FIRMWARE
This option should only be enabled if you are modifying the
firmware source to the aic7xxx driver and wish to have the
generated firmware include files updated during a normal
kernel build. The assembler for the firmware requires
lex and yacc or their equivalents, as well as the db v1
library. You may have to install additional packages or
modify the assembler make file or the files it includes
if your build environment is different than that of the
author.
I guess it didn't make it in with the rest of the 6.2.4 driver.
>There is no help for this. It's obvious that it build
>firmware but is it installed in the adapter automagically ???
Firmware is always downloaded by the driver during card initialization.
As noted above, only someone working on the firmware should need to
recompile it.
>I also wonder why the reset delay is 15000 Msec. It used to be 5000
>Msec. I've usually set it to that without nasty results. I just wonder
>what the reasoning is behind such a long delay.
Some devices require it. The default should be long enough to accomodate
all configurations. If you can use something shorter, feel free to
reconfig your kernel that way.
--
Justin
>> I also wonder why the reset delay is 15000 Msec. It used to be 5000
>> Msec. I've usually set it to that without nasty results. I just wonder
>> what the reasoning is behind such a long delay.
>
>This is a drawback of single driver for multiple cards. Good cards
>suffer to enable the driver to support bad ones.
This has nothing to do with the card the aic7xxx driver is accessing.
--
Justin
Hi Justin
Thanks much that really cleared it up for me.
John
"Justin T. Gibbs" wrote:
> >Hi
> >
> >I configuring the kernel, there is a option "Build adapter firmware with
> >Kernel build".
>
> There should be. This comment has been in my distributed patches for
> Configure.help for some time:
>
> Build Adapter Firmware with Kernel Build
> CONFIG_AIC7XXX_BUILD_FIRMWARE
> This option should only be enabled if you are modifying the
> firmware source to the aic7xxx driver and wish to have the
> generated firmware include files updated during a normal
> kernel build. The assembler for the firmware requires
> lex and yacc or their equivalents, as well as the db v1
> library. You may have to install additional packages or
> modify the assembler make file or the files it includes
> if your build environment is different than that of the
> author.
>
> I guess it didn't make it in with the rest of the 6.2.4 driver.
>
> >There is no help for this. It's obvious that it build
> >firmware but is it installed in the adapter automagically ???
>
> Firmware is always downloaded by the driver during card initialization.
> As noted above, only someone working on the firmware should need to
> recompile it.
>
> >I also wonder why the reset delay is 15000 Msec. It used to be 5000
> >Msec. I've usually set it to that without nasty results. I just wonder
> >what the reasoning is behind such a long delay.
>
> Some devices require it. The default should be long enough to accomodate
> all configurations. If you can use something shorter, feel free to
> reconfig your kernel that way.
>
> --
> Justin
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
On Thu, Nov 08, 2001 at 01:06:11PM -0700, Justin T. Gibbs wrote:
> >> I also wonder why the reset delay is 15000 Msec. It used to be 5000
> >> Msec. I've usually set it to that without nasty results. I just wonder
> >> what the reasoning is behind such a long delay.
> >
> >This is a drawback of single driver for multiple cards. Good cards
> >suffer to enable the driver to support bad ones.
>
> This has nothing to do with the card the aic7xxx driver is accessing.
Sorry if I upset you. My point was not to criticize your driver, I'm a
happy aic7xxx user.
I've used your drivers with three cards, all recent, and for each the
timeout seems oversized. I was convinced it was to support old hardware.
You seems to indicate that it depends more on the attached scsi targets,
and I believe it but I have never seen this kind of configuration where
the timeout needs to be 15000 Msec. Is this a so common config to set
this value as default ?
Christophe
>
> --
> Justin
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
--
Christophe Barb? <[email protected]>
GnuPG FingerPrint: E0F6 FADF 2A5C F072 6AF8 F67A 8F45 2F1E D72C B41E
>> >This is a drawback of single driver for multiple cards. Good cards
>> >suffer to enable the driver to support bad ones.
>>=20
>> This has nothing to do with the card the aic7xxx driver is accessing.
>
>Sorry if I upset you.
You didn't upset me. I just wanted to ensure that an inacurate statement
wasn't propogated.
>I was convinced it was to support old hardware.
Yes and no. Although more common in older hardware, there are still
companies that bring new products to market that take *forever* to
wakeup after a bus reset.
>You seems to indicate that it depends more on the attached scsi targets,
>and I believe it but I have never seen this kind of configuration where
>the timeout needs to be 15000 Msec. Is this a so common config to set
>this value as default ?
There are still lots of devices out there that don't work with shorter
timeouts. I don't know that it matters what percentage of people have
these devices so much as allowing them to install/use the OS without
the frustration of trying to figure out how to make their system work.
Other than a bit of wall clock time, the extra delay does not "hurt"
setups that can use a lower timeout.
--
Justin