Mailing List Archive

> Is the sata_nv (based on SCSI) driver supposed to work with Xen?

I'd have hoped it would have worked, but it hasn't been tested.

Xen is less forgiving of technically buggy drivers than standard
x86. Drivers that have been made to work on other architectures
such as Alpha and PPC tend to have had such bugs beaten out of
them, but I guess sata_nv isn't in that category.

My guess would be that there's some code in the driver that is
taking liberties with treating physical and bus addresses as the
same. If you can add some tracing to the driver to narrow things
down we may be able to help you.

Ian




-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Fri, 05 Nov 2004 22:52:16 +0000
Ian Pratt <Ian.Pratt@cl.cam.ac.uk> wrote:

> > Is the sata_nv (based on SCSI) driver supposed to work with Xen?
>
> I'd have hoped it would have worked, but it hasn't been tested.
>
> Xen is less forgiving of technically buggy drivers than standard
> x86. Drivers that have been made to work on other architectures
> such as Alpha and PPC tend to have had such bugs beaten out of
> them, but I guess sata_nv isn't in that category.

Nope, since the nForce chipset is currently only used for x86/x86_64
hardware (afaik).

> My guess would be that there's some code in the driver that is
> taking liberties with treating physical and bus addresses as the
> same. If you can add some tracing to the driver to narrow things
> down we may be able to help you.

Well I have some basic knowledges of C, (at the very least enough for
printing/logging some numbers :) ) but I haven't done anything with the
Linux kernel or hardware related stuff so far, so can you give any
generally applicable pointers where to start with what? (or point me to
some pointers :) )

> Ian

thanks for your help
Milan

PS: currently I'm reading this list via gmane.org, so no real need to CC
me :)

- --

Milan Holzäpfel alias jagdfalke alias jag

Antworten direkt an mich Answers directly to me
gehen bitte an eine Addresse, go to an address one
die man hier finden kann: can find here, please:

Kontaktinfos sowie Contact infos as well as
Öff GnuPG-Schlüssel <URL:http://con.mjh.name/> GnuPG Public Key
GnuPG Fingerabdruck 4C8A 5FAF 5D32 6125 89D1 GnuPG Fingerprint
0CE5 DB0C AF4F 6583 7966



http://www.deppenleerzeichen.de/


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)

iD8DBQFBjKMe2wyvT2WDeWYRAgZBAJ4lDjPx9Q2LIGgLImsb2QNnQnPvLQCePbBt
omVmI3ozAHuNo3ShL+P09V8=
=xiMx
-----END PGP SIGNATURE-----



-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_idU88&alloc_id065&op=click
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

Ian Pratt <Ian.Pratt <at> cl.cam.ac.uk> writes:

> Xen is less forgiving of technically buggy drivers than standard
> x86. Drivers that have been made to work on other architectures
> such as Alpha and PPC tend to have had such bugs beaten out of
> them, but I guess sata_nv isn't in that category.
>
> My guess would be that there's some code in the driver that is
> taking liberties with treating physical and bus addresses as the
> same. If you can add some tracing to the driver to narrow things
> down we may be able to help you.


I have very similar problem with sata_via (SCSI) driver.
Can you give me pointers to track down the problem ?

Cédric





-------------------------------------------------------
This SF.Net email is sponsored by: InterSystems CACHE
FREE OODBMS DOWNLOAD - A multidimensional database that combines
robust object and relational technologies, making it a perfect match
for Java, C++,COM, XML, ODBC and JDBC. www.intersystems.com/match8
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

> > My guess would be that there's some code in the driver that
> is taking
> > liberties with treating physical and bus addresses as the
> same. If you
> > can add some tracing to the driver to narrow things down we may be
> > able to help you.
>
> I have very similar problem with sata_via (SCSI) driver.
> Can you give me pointers to track down the problem ?

Are you getting a hang or a kernel oops?

See if the driver has any debug tracing you can enable. If not, just sprinkle some printk's around. Pay particular attention to any 'virt_to_phy' or similar functions - they should probably be virt_to_bus.

I suspect the problem is localised to the driver, as the sata_promise driver works for us just fine.

Ian


-------------------------------------------------------
This SF.Net email is sponsored by: InterSystems CACHE
FREE OODBMS DOWNLOAD - A multidimensional database that combines
robust object and relational technologies, making it a perfect match
for Java, C++,COM, XML, ODBC and JDBC. www.intersystems.com/match8
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

> Ian Pratt <Ian.Pratt <at> cl.cam.ac.uk> writes:
>
> > Xen is less forgiving of technically buggy drivers than standard
> > x86. Drivers that have been made to work on other architectures
> > such as Alpha and PPC tend to have had such bugs beaten out of
> > them, but I guess sata_nv isn't in that category.
> >
> > My guess would be that there's some code in the driver that is
> > taking liberties with treating physical and bus addresses as the
> > same. If you can add some tracing to the driver to narrow things
> > down we may be able to help you.
>
>
> I have very similar problem with sata_via (SCSI) driver.
> Can you give me pointers to track down the problem ?
>
> Cédric

Perhaps a starting point would be to compare the start-of-day driver
messages from booting native Linux compared with booting
XenLinux. Also, after booting native Linux, take a look at the output
of 'cat /proc/interrupts' and see what IRQ# is allocated to teh SATA
controller.

-- Keir


-------------------------------------------------------
This SF.Net email is sponsored by: InterSystems CACHE
FREE OODBMS DOWNLOAD - A multidimensional database that combines
robust object and relational technologies, making it a perfect match
for Java, C++,COM, XML, ODBC and JDBC. www.intersystems.com/match8
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sat, 13 Nov 2004 20:03:58 -0000
"Ian Pratt" <m+Ian.Pratt@cl.cam.ac.uk> wrote:

>
> > > My guess would be that there's some code in the driver that
> > > is taking
> > > liberties with treating physical and bus addresses as the
> > > same. If you
> > > can add some tracing to the driver to narrow things down we may be
> > > able to help you.
> >
> > I have very similar problem with sata_via (SCSI) driver.
> > Can you give me pointers to track down the problem ?
>
> Are you getting a hang or a kernel oops?

hang, no oops message or anything, it just stops

> See if the driver has any debug tracing you can enable. If not, just sprinkle some printk's around. Pay particular attention to any 'virt_to_phy' or similar functions - they should probably be virt_to_bus.

Well just after writing my last mail I had a short look at the driver,
and I think I have found some path from the driver initialization part
down to where the last messages are printed, but I hadn't time to add
printks or whatever so far. But thanks for the funtion name, I will pay
attention to it when I get to having a look at it.

A word to Keir's post: From my experimentation a week ago, I do
remember that some different IRQ was used for the S-ATA controller with
native linux vs Xen linux.

On another point, I don't know how much output I will produce (for my
two disks at two ports), but I doubt that it will fit my screen if I am
finally diligent in the driver :) So either I capture the log via some
serial port (if that is possible to do with just another PC next to it
and some cable) or via the kernel's net-log-facility, but that would
require it working before the initialization of the S-ATA controller...
ok I won't look at it right now anyway but if you have some links lying
around, feel free to post them ;)

(in case you should be interested, I just spend this weeks free time
participating in bwinf, "Bundeswettbewerb für Informatik", some German
informatics competition, and I have 3 class tests to do this week, so
maybe I get to it next weekend or so :) )

> I suspect the problem is localised to the driver, as the sata_promise driver works for us just fine.

that sounds reasonable

> Ian

Regards,
Milan


- --

Milan Holzäpfel alias jagdfalke alias jag

Antworten direkt an mich Answers directly to me
gehen bitte an eine Addresse, go to an address one
die man hier finden kann: can find here, please:

Kontaktinfos sowie Contact infos as well as
Öff GnuPG-Schlüssel <URL:http://con.mjh.name/> GnuPG Public Key
GnuPG Fingerabdruck 4C8A 5FAF 5D32 6125 89D1 GnuPG Fingerprint
0CE5 DB0C AF4F 6583 7966



http://www.deppenleerzeichen.de/


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)

iD8DBQFBlzU62wyvT2WDeWYRArpZAKC4v5oDAcgN5DrMHKN71HC8tZ85zwCgki7q
RmH4ugavTiAQ1jGNx3pzRV8=
=8Ft3
-----END PGP SIGNATURE-----



-------------------------------------------------------
This SF.Net email is sponsored by: InterSystems CACHE
FREE OODBMS DOWNLOAD - A multidimensional database that combines
robust object and relational technologies, making it a perfect match
for Java, C++,COM, XML, ODBC and JDBC. www.intersystems.com/match8
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

> On another point, I don't know how much output I will produce (for my
> two disks at two ports), but I doubt that it will fit my screen if I am
> finally diligent in the driver :) So either I capture the log via some
> serial port (if that is possible to do with just another PC next to it
> and some cable) or via the kernel's net-log-facility, but that would
> require it working before the initialization of the S-ATA controller...
> ok I won't look at it right now anyway but if you have some links lying
> around, feel free to post them ;)

Serial port is your best bet. If you have 115.2kbaud connection then:
1. Add 'com1=115200,8n1' to Xen's command line.
2. 'console=tty0' becomes 'console=ttyS0' on XenLinux's command line.

-- Keir


-------------------------------------------------------
This SF.Net email is sponsored by: InterSystems CACHE
FREE OODBMS DOWNLOAD - A multidimensional database that combines
robust object and relational technologies, making it a perfect match
for Java, C++,COM, XML, ODBC and JDBC. www.intersystems.com/match8
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel