Re: "panic: fxp_start_body: attempted use of a free mbuf!" in RELENG_5

2004-10-08 Thread Kris Kennaway
On Fri, Oct 08, 2004 at 11:56:52PM +0100, Brian Somers wrote: > On Thu, 7 Oct 2004 09:57:19 -0700, Kris Kennaway <[EMAIL PROTECTED]> wrote: > > Top of RELENG_5 on an smp machine: > > > > panic: fxp_start_body: attempted use of a free mbuf! > > I'm guessing you've got nothing weird saying options

Infinite loop in tcp_output on RELENG_5

2004-10-08 Thread Kris Kennaway
pointyhat (SMP machine running RELENG_5) has twice in the past 2 days gone into an infinite loop in the tcp_output() function (repeatedly breaking into DDB and continuing, I can see it at different points in the code). I made tcp_output keep a counter and increment when it hits the again: label.

Re: "panic: fxp_start_body: attempted use of a free mbuf!" in RELENG_5

2004-10-08 Thread Brian Somers
On Thu, 7 Oct 2004 09:57:19 -0700, Kris Kennaway <[EMAIL PROTECTED]> wrote: > Top of RELENG_5 on an smp machine: > > panic: fxp_start_body: attempted use of a free mbuf! I'm guessing you've got nothing weird saying options MSIZE= in your kernel config? I committed a fix to -current that moans if

Re: Calling m_pullup in ether_input

2004-10-08 Thread Julian Elischer
Giorgos Keramidas wrote: On 2004-10-09 00:34, Gleb Smirnoff <[EMAIL PROTECTED]> wrote: On Fri, Oct 08, 2004 at 06:15:15PM +0300, Giorgos Keramidas wrote: G> In ether_input() we unconditionaly discard the mbufs whose m_len is less than G> ETHER_HDR_LEN. A bit higher M_PKTHDR has been checked bu

Re: Calling m_pullup in ether_input

2004-10-08 Thread Giorgos Keramidas
On 2004-10-09 00:58, Gleb Smirnoff <[EMAIL PROTECTED]> wrote: > On Fri, Oct 08, 2004 at 11:51:17PM +0300, Giorgos Keramidas wrote: > G> Not really. Just trying to see if there _was_ any packet loss first. > G> After running for several hours, the counters are still zero. I think > G> this means t

Re: Calling m_pullup in ether_input

2004-10-08 Thread Gleb Smirnoff
On Fri, Oct 08, 2004 at 11:51:17PM +0300, Giorgos Keramidas wrote: G> Not really. Just trying to see if there _was_ any packet loss first. G> After running for several hours, the counters are still zero. I think G> this means the case of dropped packets would be so rare that there's no G> real re

Re: Calling m_pullup in ether_input

2004-10-08 Thread Giorgos Keramidas
On 2004-10-09 00:34, Gleb Smirnoff <[EMAIL PROTECTED]> wrote: > On Fri, Oct 08, 2004 at 06:15:15PM +0300, Giorgos Keramidas wrote: > G> In ether_input() we unconditionaly discard the mbufs whose m_len is less than > G> ETHER_HDR_LEN. A bit higher M_PKTHDR has been checked but the check made > G> b

Re: Calling m_pullup in ether_input

2004-10-08 Thread Gleb Smirnoff
On Fri, Oct 08, 2004 at 06:15:15PM +0300, Giorgos Keramidas wrote: G> In ether_input() we unconditionaly discard the mbufs whose m_len is less than G> ETHER_HDR_LEN. A bit higher M_PKTHDR has been checked but the check made G> before discarding the frame doesn't pay attention to the m->m_pkthdr.le

Re: modularization

2004-10-08 Thread Roman Kurakin
Waldemar Kornewald wrote: Roman Kurakin wrote: Of course, I will send progress reports on this list. We at Haiku would really appreciate working together with the FreeBSD team. By the way, what kind of hardware you are using? Currently, it is nothing special, just a Dual-Celeron 466 (since I ga

Re: modularization

2004-10-08 Thread Waldemar Kornewald
Roman Kurakin wrote: Of course, I will send progress reports on this list. We at Haiku would really appreciate working together with the FreeBSD team. By the way, what kind of hardware you are using? Currently, it is nothing special, just a Dual-Celeron 466 (since I gave my other computer to my

Re: modularization

2004-10-08 Thread Roman Kurakin
Waldemar Kornewald wrote: George V. Neville-Neil wrote: One other model to look at is The Click Modular Router, which is about modularizing the routing part of the code, as opposed to the end station code. Look at http://www.xorp.org because Click, and FreeBSD are in there. Thanks, this is really

Re: modularization

2004-10-08 Thread Waldemar Kornewald
George V. Neville-Neil wrote: One other model to look at is The Click Modular Router, which is about modularizing the routing part of the code, as opposed to the end station code. Look at http://www.xorp.org because Click, and FreeBSD are in there. Thanks, this is really interesting. I have always

Calling m_pullup in ether_input

2004-10-08 Thread Giorgos Keramidas
In ether_input() we unconditionaly discard the mbufs whose m_len is less than ETHER_HDR_LEN. A bit higher M_PKTHDR has been checked but the check made before discarding the frame doesn't pay attention to the m->m_pkthdr.len (the total packet length). I am trying to find out how often this happens

fxp(4) suspend/resume in 5.3

2004-10-08 Thread Bruce M Simpson
I was looking at the fxp suspend/resume routines and thought: hang on a minute, we already save PCI configuration space in the PCI bus suspend/resume routines, so do we need to do it in fxp(4)? Regards, BMS ___ [EMAIL PROTECTED] mailing list http://lists