On 02/19/2013 05:09 PM, Eric Dumazet wrote: > On Tue, Feb 19, 2013 at 2:30 PM, Allan, Bruce W <[email protected]> > wrote: >>> -----Original Message----- >>> From: Andrew Morton [mailto:[email protected]] >>> Sent: Tuesday, February 19, 2013 2:27 PM >>> To: Wu, Fengguang >>> Cc: Daniel Santos; Kirsher, Jeffrey T; Brandeburg, Jesse; Allan, Bruce W; >>> [email protected] >>> Subject: Re: [next:akpm 16/587] >>> drivers/net/ethernet/intel/igb/igb_main.c:6231:2: error: call to >>> '__compiletime_assert_6235' declared with attribute error: BUILD_BUG_ON >>> failed: SKB_WITH_OVERHEAD(IGB_RX_BUFSZ) < (NET_SKB_PAD + >>> NET_IP_ALIGN + IGB_TS_HDR_LEN + ETH... >>> >>> On Mon, 18 Feb 2013 19:27:22 +0800 >>> kbuild test robot <[email protected]> wrote: >>> >>>> tree: git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git >>> akpm >>>> head: ae9da2a24a80423b8dbda9c5a72674ce6e7861be >>>> commit: f7d7e38e6df399b2af7f04c177054abdd744a077 [16/587] bug.h, >>> compiler.h: Introduce compiletime_assert & BUILD_BUG_ON_MSG >>>> config: make ARCH=s390 allmodconfig >>>> >>>> All error/warnings: >>>> >>>> In function 'igb_build_rx_buffer', >>>> inlined from 'igb_clean_rx_irq' at >>> drivers/net/ethernet/intel/igb/igb_main.c:6705:8: >>>>>> drivers/net/ethernet/intel/igb/igb_main.c:6231:2: error: call to >>> '__compiletime_assert_6235' declared with attribute error: BUILD_BUG_ON >>> failed: SKB_WITH_OVERHEAD(IGB_RX_BUFSZ) < (NET_SKB_PAD + >>> NET_IP_ALIGN + IGB_TS_HDR_LEN + ETH_FRAME_LEN + ETH_FCS_LEN) >>>> vim +/__compiletime_assert_6235 +6231 >>> drivers/net/ethernet/intel/igb/igb_main.c >>>> 6225 #endif >>>> 6226 >>>> 6227 /* If we spanned a buffer we have a huge mess so test >>>> for it >>> */ >>>> 6228 BUG_ON(unlikely(!igb_test_staterr(rx_desc, >>> E1000_RXD_STAT_EOP))); >>>> 6229 >>>> 6230 /* Guarantee this function can be used by verifying >>>> buffer >>> sizes */ >>>>> 6231 >>> BUILD_BUG_ON(SKB_WITH_OVERHEAD(IGB_RX_BUFSZ) < >>> (NET_SKB_PAD + >>>> 6232 >>> NET_IP_ALIGN + >>>> 6233 >>> IGB_TS_HDR_LEN + >>>> 6234 >>> ETH_FRAME_LEN + >>> That BUILD_BUG_ON() triggers without Daniel's patches applied as well: >>> >>> CC [M] drivers/net/ethernet/intel/igb/igb_main.o >>> drivers/net/ethernet/intel/igb/igb_main.c: In function >>> 'igb_build_rx_buffer': >>> drivers/net/ethernet/intel/igb/igb_main.c:6231: error: size of array 'type >>> name' is negative >>> >>> I suspect that driver is busted on s390. Dunno why. Here's what cpp emits: >>> >>> >>> s390 allmodconfig: >>> >>> bool __cond = !(!(((2048) - (((sizeof(struct skb_shared_info)) + (256 - 1)) >>> & >>> ~(256 - 1))) < (32 + 2 + 16 + 1514 + 4))); >>> >>> >>> x86_64 allmodconfig: >>> >>> bool __cond = !(!(((2048) - (((sizeof(struct skb_shared_info)) + ((1 << >>> (6)) - 1)) & ~((1 << (6)) - 1))) < (({ typeof(32) _max1 = (32); typeof((1 >>> << (6))) _max2 = ((1 << (6))); (void) (&_max1 == &_max2); _max1 > _max2 ? >>> _max1 : _max2; }) + 0 + 16 + 1514 + 4))); >>> >>> >>> I doubt if anyone will be running this driver on s390, but I guess we >>> should work out why it broke. >> Adding Intel wired ethernet developers mailing list. >> > Oh well, NET_SKB_PAD is too large on s390 > > I guess we should use 64 bytes max. > > #define NET_SKB_PAD min(64, max(32, L1_CACHE_BYTES))
The problem is the 256 byte alignment for L1_CACHE_BYTES is increasing the size of the data and shared info significantly pushing us past the 2K limit. I'll look into this since it likely affects ixgbe as well. Thanks, Alex ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb _______________________________________________ E1000-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/e1000-devel To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired
