Re: CVS commit: src/sys/dev/usb

[Michael van Elst <mlelstv%serpens.de@localhost>]

On Wed, Jun 11, 2014 at 01:23:37PM +0900, Masanobu SAITOH wrote:
> (2014/06/11 13:20), Jun Ebihara wrote:
> >From: Michael van Elst <mlelstv%serpens.de@localhost>
> >Subject: Re: CVS commit: src/sys/dev/usb
> >Date: Wed, 11 Jun 2014 05:37:23 +0200
> >
> >>>After this usb improve, I've got some uvm_fault on RPI
> >>>when I try to pkg_add packages, such as pkgsrc/fonts/vlgothic-ttf.
> >>Do you use bpf at the same time?
> >
> >Not use any bpf.
> >At pkg setup time, I use dhcpd and ld0 and serial console.
> 
> Offcourse dhcpd use bpf.

Maybe dhcpcd, which also uses bpf. But I also run dhcpcd.


> FSR=00000005, FAR=10009992, spsr=80000113
> r0 =c884e800, r1 =10009992, r2 =cbf49dbc, r3 =00000000
> r4 =0000000e, r5 =00000002, r6 =00000000, r7 =cbf49e08
> r8 =c884e800, r9 =c0549db8, r10=c0549ca8, r11=cbf49de4
> r12=10009986, ssp=cbf49da8, slr=c00393e8, pc =c0038d2c

r12=10009986 is supposed to be a mbuf data pointer but is already
invalid.

r0=c884e800 should be the mbuf pointer and looks ok. Can you see
wether the data pointer is already invalid in the mbuf header?

                                                                                
> Stopped in pid 0.3 (system) at  netbsd:m_xhalf+0x94:    ldrh    r3, [r1]
> db> bt
> 0xcbf49de4: netbsd:bpf_filter_ext+0xc

And this is strange. There is no call to m_xhalf() at bpf_filter_ext+0xc
which is still in the function prologue. Did some characters get lost
on the console?





Greetings,
-- 
                                Michael van Elst
Internet: mlelstv%serpens.de@localhost
                                "A potential Snark may lurk in every tree."