Panic after using bonding driver

From: Jeff Golds (jgolds@resilience.com)
Date: Thu Mar 29 2001 - 12:41:18 EST

Next message: Pavel Roskin: "[PATCH] Support for "mode" parameter in ramfs"
Previous message: Shawn Starr: "Re: [WISHLIST] Addition of suspend patch into 2.5?"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

I have been working on a driver similar to the bonding driver and have
come across a bug in the bonding driver code. When the bonding driver
enslaves a device, it modifies the slave's multicast list to be the
master's multicast list. Later, after the master is downed, the kernel
gets a panic if you try to down the slave device.

To get around this problem, there are two solutions that I see:

1) Don't do multicasting for bonding devices
        While this works (I've tested) some peple might call this a serious
limitation.
2) Keep track of the slave's multicast list
        This would require keeping a copy of the slave's pointer and restoring
it when the bonding
        device is downed. Not sure if this would even work since the slave's
multicast list might
        be stale by the time it is restored.

I'd like to get this fixed in the best possible way. What are your
folks comments in regard to the matter?

-Jeff

-- 
Jeff Golds
jgolds@resilience.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Next message: Pavel Roskin: "[PATCH] Support for "mode" parameter in ramfs"
Previous message: Shawn Starr: "Re: [WISHLIST] Addition of suspend patch into 2.5?"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

This archive was generated by hypermail 2b29 : Sat Mar 31 2001 - 21:00:21 EST