(2013/05/15 11:43), Masanobu SAITOH wrote:
Hi, macallan. (2013/05/14 19:59), Michael wrote:Hello, On Tue, 14 May 2013 03:42:16 +0900 SAITOH Masanobu <msaitoh%execsw.org@localhost> wrote:(2012/10/18 2:47), Michael wrote:Hello, On Oct 17, 2012, at 2:08 AM, Erik E. Fair wrote:I took the latest distribution/builds of netbsd-6 and tried to netboot them this evening on a Sun Fire v215 and Sun Fire v240.Both got ofwboot.net fine, and loaded the GENERIC (MP) kernel fine (though the v215 complained about the aceride tc_count), but when it came time to mount NFS root, both failed: bge0 watchdog timeout -- resetting (resulting in "cannot mount root, error = 79").Have we screwed the pooch in interrupt routing, again?I see the same on a Blade 2500 - on some bges our driver doesn't work if OF touched the hardware first, it works fine if you boot from something else. have fun MichaelIs this problem still alive in -current?Can't tell, the kernel crashes with an illegal access to PCI config space right after bge establishes its interrupt handler. ( sorry, didn't have time to dig any further yet - the fault comes from schizo_conf_read() ) Same when booting from harddisk. have fun MichaelAlignment problem?
I tested with Netra T1 + BCM5705K NIC(non on-board). It worked.
bge0 at pci3 dev 15 function 0: Broadcom BCM5705K Gigabit Ethernet bge0: interrupting at ivec 3 bge0: HW config 00010054, 00000004, 00000000, 00000000 bge0: ASIC BCM5705 A3 (0x3003), Ethernet address 00:18:71:4d:3a:71 bge0: setting short Tx thresholds brgphy0 at bge0 phy 1: BCM5705 1000BASE-T media interface, rev. 2 brgphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
sparc64# ping6 ff02::2%bge0 PING6(56=40+8+8 bytes) fe80::218:71ff:fe4d:3a71%bge0 --> ff02::2%bge0 16 bytes from fe80::250:43ff:fe36:342d%bge0, icmp_seq=0 hlim=64 time=3.164 ms
Could you show me the full dmesg?
--
-----------------------------------------------
SAITOH Masanobu (msaitoh%execsw.org@localhost
msaitoh%netbsd.org@localhost)