> -----Original Message-----
> From: Wu, Fengguang
> Sent: Tuesday, September 25, 2012 5:22 AM
> To: Vick, Matthew
> Cc: kernel-janitors@xxxxxxxxxxxxxxx; Kirsher, Jeffrey T; linux-
> can@xxxxxxxxxxxxxxx
> Subject: Re: [can:master 110/171]
> drivers/net/ethernet/intel/igb/igb.h:347:25: error: field
> 'hwtstamp_config' has incomplete type
>
> On Sat, Sep 22, 2012 at 07:38:12AM +0800, Vick, Matthew wrote:
> > > -----Original Message-----
> > > From: Wu, Fengguang
> > > Sent: Friday, September 21, 2012 5:17 AM
> > > To: Vick, Matthew
> > > Cc: kernel-janitors@xxxxxxxxxxxxxxx; Kirsher, Jeffrey T; linux-
> > > can@xxxxxxxxxxxxxxx
> > > Subject: [can:master 110/171]
> > > drivers/net/ethernet/intel/igb/igb.h:347:25: error: field
> > > 'hwtstamp_config' has incomplete type
> > >
> > > Hi Matthew,
> > >
> > > FYI, kernel build failed on
> > >
> > > tree: git://gitorious.org/linux-can/linux-can-next master
> > > head: a1161bc7bdf4eedccb1b91e7479f303cf95af533
> > > commit: 3c89f6d0d4aa92ff252c17f8f6b5c8983ae2b77e [110/171] igb:
> Tidy
> > > up wrapping for CONFIG_IGB_PTP.
> > > config: x86_64-allyesdebian (attached as .config)
> > >
> > > All related error/warning messages:
> > >
> > > In file included from
> > > drivers/net/ethernet/intel/igb/igb_ethtool.c:41:0:
> > > drivers/net/ethernet/intel/igb/igb.h:347:25: error: field
> > > 'hwtstamp_config' has incomplete type
> > > --
> > > In file included from
> drivers/net/ethernet/intel/igb/e1000_mac.c:36:0:
> > > drivers/net/ethernet/intel/igb/igb.h:347:25: error: field
> > > 'hwtstamp_config' has incomplete type
> > >
> > > vim +347 drivers/net/ethernet/intel/igb/igb.h
> > > 341 u8 tx_timeout_factor;
> > > 342 struct timer_list blink_timer;
> > > 343 unsigned long led_status;
> > > 344
> > > 345 /* OS defined structs */
> > > 346 struct pci_dev *pdev;
> > > > 347 struct hwtstamp_config hwtstamp_config;
> > > 348
> > > 349 spinlock_t stats64_lock;
> > > 350 struct rtnl_link_stats64 stats64;
> > >
> > > ---
> > > 0-DAY kernel build testing backend Open Source Technology
> > > Centre
> > > Fengguang Wu, Yuanhan Liu Intel
> > > Corporation
> >
> > Fengguang,
> >
> > I'm not certain I understand how this issue came back to me. The
> patch
> > in question doesn't touch that code, nor should it be changing
> > anything that would cause a build error here. The offending line is
> > also removed in the next patch
> > (a79f4f88261d7fd492121daf85beafff663f1f01). I'm also unable to
> > reproduce the error.
>
> Hi Matthew, this is interesting, but I do reproduced the issue by
> building 3c89f6d0d4aa92ff252c17f8f6b5c8983ae2b77e from scratch with the
> originally attached config file:
>
> CC drivers/net/ethernet/qlogic/qla3xxx.o
> CC drivers/net/ethernet/intel/ixgbevf/mbx.o
> CC drivers/net/ethernet/intel/ixgbe/ixgbe_main.o
> CC drivers/net/ethernet/qlogic/netxen/netxen_nic_hw.o
> CC drivers/net/ethernet/intel/ixgbevf/ethtool.o
> CC drivers/net/ethernet/intel/igb/e1000_i210.o
> CC drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.o
> CC drivers/net/ethernet/intel/ixgbe/ixgbe_common.o
> CC drivers/net/ethernet/realtek/r8169.o
> CC drivers/net/ethernet/qlogic/netxen/netxen_nic_main.o
> In file included from
> /c/wfg/linux/drivers/net/ethernet/intel/igb/igb_ethtool.c:41:0:
> /c/wfg/linux/drivers/net/ethernet/intel/igb/igb.h:347:25: error: field
> ‘hwtstamp_config’ has incomplete type
> CC drivers/net/ethernet/qlogic/netxen/netxen_nic_init.o
> In file included from
> /c/wfg/linux/drivers/net/ethernet/intel/igb/e1000_mac.c:36:0:
> /c/wfg/linux/drivers/net/ethernet/intel/igb/igb.h:347:25: error: field
> ‘hwtstamp_config’ has incomplete type
> make[6]: *** [drivers/net/ethernet/intel/igb/e1000_mac.o] Error 1
> make[6]: *** Waiting for unfinished jobs....
> CC drivers/net/ethernet/qlogic/netxen/netxen_nic_ethtool.o
> make[6]: *** [drivers/net/ethernet/intel/igb/igb_ethtool.o] Error 1
>
> Thanks,
> Fengguang
Ah, I see the issue now. Some combination of CONFIG_IGB_PTP not being supported alongside other kernel configuration strangeness I haven't fully dived into seems to be the culprit.
What would you like me to do with this information, though? The offending code is removed in the next patch, so I'm unable to clean it upstream at this point.
Cheers,
Matthew
��.n��������+%������w��{.n����z�ޗ�����n�r������&��z�ޗ�zf���h���~����������_��+v���)ߣ�
