Re: HEADS-UP: PCI ID driver conflict with ahd(4) and pms(4)

2015-07-31 Thread Kurt Jaeger
Hi! > prior to the releng/10.2 branch), please exercise caution when upgrading > to the latest stable/10 or releng/10.2 if your storage controller > matches any of the following PCI device IDs: Which field of pciconf -lvb or sysctl output needs to be compared ? Is it the device= field ? dev.ata

Re: HEADS-UP: PCI ID driver conflict with ahd(4) and pms(4)

2015-07-31 Thread Glen Barber
On Fri, Jul 31, 2015 at 09:20:49AM +0200, Kurt Jaeger wrote: > Hi! > > > prior to the releng/10.2 branch), please exercise caution when upgrading > > to the latest stable/10 or releng/10.2 if your storage controller > > matches any of the following PCI device IDs: > > Which field of pciconf -lvb

Re: HEADS-UP: PCI ID driver conflict with ahd(4) and pms(4)

2015-07-31 Thread Glen Barber
On Fri, Jul 31, 2015 at 01:48:32AM +, Glen Barber wrote: > The new driver for PMC Sierra SAS/SATA HBAs does not properly validate > that both the vendor ID and device ID of some storage device drivers. > At present, we know there is a conflict with ahd(4), it is possible > there are more. > T

KSTACK_PAGES stuck at 2 despite options KSTACK_PAGES=4

2015-07-31 Thread Trond Endrestøl
stable/10, i386, r286139, 4 GiB RAM, custom kernel loudly claims: ZFS NOTICE: KSTACK_PAGES is 2 which could result in stack overflow panic! Please consider adding 'options KSTACK_PAGES=4' to your kernel config Well, my custom kernel config does contain: options KSTACK_PAGES=4 and options ZFS