On Mon, Jun 15, 2020 at 10:59 AM 'Eric Dumazet' via Clang Built Linux <clang-built-li...@googlegroups.com> wrote: > > On Mon, Jun 15, 2020 at 10:54 AM Eric Dumazet <eduma...@google.com> wrote: > > > > On Mon, Jun 15, 2020 at 10:44 AM Nick Desaulniers > > <ndesaulni...@google.com> wrote: > > > > > > On Mon, Jun 15, 2020 at 9:17 AM 'Eric Dumazet' via Clang Built Linux > > > <clang-built-li...@googlegroups.com> wrote: > > > > > > > > On Sun, Jun 14, 2020 at 11:26 PM kernel test robot <l...@intel.com> > > > > wrote: > > > > > > > > > > tree: > > > > > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git > > > > > master > > > > > head: 96144c58abe7ff767e754b5b80995f7b8846d49b > > > > > commit: 39d010504e6b4485d7ceee167743620dd33f4417 net_sched: sch_fq: > > > > > add horizon attribute > > > > > date: 6 weeks ago > > > > > :::::: branch date: 3 hours ago > > > > > :::::: commit date: 6 weeks ago > > > > > config: arm-randconfig-r006-20200614 (attached as .config) > > > > > compiler: clang version 11.0.0 (https://github.com/llvm/llvm-project > > > > > c669a1ed6386d57a75a602b53266466dae1e1d84) > > > > > reproduce (this is a W=1 build): > > > > > wget > > > > > https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross > > > > > -O ~/bin/make.cross > > > > > chmod +x ~/bin/make.cross > > > > > # install arm cross compiling tool for clang build > > > > > # apt-get install binutils-arm-linux-gnueabi > > > > > git checkout 39d010504e6b4485d7ceee167743620dd33f4417 > > > > > # save the attached .config to linux build tree > > > > > COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross > > > > > ARCH=arm > > > > > > > > > > If you fix the issue, kindly add following tag as appropriate > > > > > Reported-by: kernel test robot <l...@intel.com> > > > > > > > > > > All warnings (new ones prefixed by >>, old ones prefixed by <<): > > > > > > > > > > >> net/sched/sch_fq.c:966:12: warning: stack frame size of 1400 bytes > > > > > >> in function 'fq_dump' [-Wframe-larger-than=] > > > > > static int fq_dump(struct Qdisc *sch, struct sk_buff *skb) > > > > > ^ > > > > > > > > > > > > This looks like a bug in CLANG on ARM, there is no way fq_dump() could > > > > consume so much stack space. > > > > > > You can use > > > https://github.com/ClangBuiltLinux/frame-larger-than > > > to help debug these. You might be surprised who's doing larger stack > > > allocations than expected. > > > > What is wrong with scripts/checkstack.pl ? > > After using CLANG on x86 I get : > > $ objdump -d net/sched/sch_fq.o | scripts/checkstack.pl > 0x00001136 fq_change [sch_fq.o]: 192 > 0x00000016 __direct_call_Qdisc_enqueue1 [sch_fq.o]: 112 > 0x000019e0 fq_dump_stats [sch_fq.o]: 112 > 0x00001b43 fq_dump_stats [sch_fq.o]: 112 > > So CLANG on x86 seems fine. > > I said : " this looks like a bug in CLANG on ARM"
This is a randconfig, so some option is turning on something that's causing excessive stack usage with your patch on ARM. Building who knows what config with x86 and claiming it works doesn't really cut it. You looked at fq_dump and claimed it should be fine. Did you verify each called function that was inlined? And the issue with checkstack is it tells you the depth, but not *what variables* and from which inlined function may be the cause. > > > > > > > > > > > > > > > > > > > > > > > > > 1 warning generated. > > > > > > > > > > # > > > > > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=39d010504e6b4485d7ceee167743620dd33f4417 > > > > > git remote add linus > > > > > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git > > > > > git remote update linus > > > > > git checkout 39d010504e6b4485d7ceee167743620dd33f4417 > > > > > vim +/fq_dump +966 net/sched/sch_fq.c > > > > > > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 965 > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 @966 static int > > > > > fq_dump(struct Qdisc *sch, struct sk_buff *skb) > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 967 { > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 968 struct > > > > > fq_sched_data *q = qdisc_priv(sch); > > > > > 48872c11b77271 Eric Dumazet 2018-11-11 969 u64 > > > > > ce_threshold = q->ce_threshold; > > > > > 39d010504e6b44 Eric Dumazet 2020-05-01 970 u64 horizon = > > > > > q->horizon; > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 971 struct nlattr > > > > > *opts; > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 972 > > > > > ae0be8de9a53cd Michal Kubecek 2019-04-26 973 opts = > > > > > nla_nest_start_noflag(skb, TCA_OPTIONS); > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 974 if (opts == > > > > > NULL) > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 975 goto > > > > > nla_put_failure; > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 976 > > > > > 65c5189a2b57b9 Eric Dumazet 2013-11-15 977 /* > > > > > TCA_FQ_FLOW_DEFAULT_RATE is not used anymore */ > > > > > 65c5189a2b57b9 Eric Dumazet 2013-11-15 978 > > > > > 48872c11b77271 Eric Dumazet 2018-11-11 979 > > > > > do_div(ce_threshold, NSEC_PER_USEC); > > > > > 39d010504e6b44 Eric Dumazet 2020-05-01 980 > > > > > do_div(horizon, NSEC_PER_USEC); > > > > > 48872c11b77271 Eric Dumazet 2018-11-11 981 > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 982 if > > > > > (nla_put_u32(skb, TCA_FQ_PLIMIT, sch->limit) || > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 983 > > > > > nla_put_u32(skb, TCA_FQ_FLOW_PLIMIT, q->flow_plimit) || > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 984 > > > > > nla_put_u32(skb, TCA_FQ_QUANTUM, q->quantum) || > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 985 > > > > > nla_put_u32(skb, TCA_FQ_INITIAL_QUANTUM, q->initial_quantum) || > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 986 > > > > > nla_put_u32(skb, TCA_FQ_RATE_ENABLE, q->rate_enable) || > > > > > 76a9ebe811fb3d Eric Dumazet 2018-10-15 987 > > > > > nla_put_u32(skb, TCA_FQ_FLOW_MAX_RATE, > > > > > 76a9ebe811fb3d Eric Dumazet 2018-10-15 988 > > > > > min_t(unsigned long, q->flow_max_rate, ~0U)) || > > > > > f52ed89971adbe Eric Dumazet 2013-11-15 989 > > > > > nla_put_u32(skb, TCA_FQ_FLOW_REFILL_DELAY, > > > > > f52ed89971adbe Eric Dumazet 2013-11-15 990 > > > > > jiffies_to_usecs(q->flow_refill_delay)) || > > > > > 06eb395fa9856b Eric Dumazet 2015-02-04 991 > > > > > nla_put_u32(skb, TCA_FQ_ORPHAN_MASK, q->orphan_mask) || > > > > > 77879147a3481b Eric Dumazet 2016-09-19 992 > > > > > nla_put_u32(skb, TCA_FQ_LOW_RATE_THRESHOLD, > > > > > 77879147a3481b Eric Dumazet 2016-09-19 993 > > > > > q->low_rate_threshold) || > > > > > 48872c11b77271 Eric Dumazet 2018-11-11 994 > > > > > nla_put_u32(skb, TCA_FQ_CE_THRESHOLD, (u32)ce_threshold) || > > > > > 583396f4ca4d6e Eric Dumazet 2020-03-16 995 > > > > > nla_put_u32(skb, TCA_FQ_BUCKETS_LOG, q->fq_trees_log) || > > > > > 39d010504e6b44 Eric Dumazet 2020-05-01 996 > > > > > nla_put_u32(skb, TCA_FQ_TIMER_SLACK, q->timer_slack) || > > > > > 39d010504e6b44 Eric Dumazet 2020-05-01 997 > > > > > nla_put_u32(skb, TCA_FQ_HORIZON, (u32)horizon) || > > > > > 39d010504e6b44 Eric Dumazet 2020-05-01 998 > > > > > nla_put_u8(skb, TCA_FQ_HORIZON_DROP, q->horizon_drop)) > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 999 goto > > > > > nla_put_failure; > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 1000 > > > > > d59b7d8059ddc4 Yang Yingliang 2014-03-12 1001 return > > > > > nla_nest_end(skb, opts); > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 1002 > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 1003 nla_put_failure: > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 1004 return -1; > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 1005 } > > > > > afe4fd062416b1 Eric Dumazet 2013-08-29 1006 > > > > > > > > > > :::::: The code at line 966 was first introduced by commit > > > > > :::::: afe4fd062416b158a8a8538b23adc1930a9b88dc pkt_sched: fq: Fair > > > > > Queue packet scheduler > > > > > > > > > > :::::: TO: Eric Dumazet <eduma...@google.com> > > > > > :::::: CC: David S. Miller <da...@davemloft.net> > > > > > > > > > > --- > > > > > 0-DAY CI Kernel Test Service, Intel Corporation > > > > > https://lists.01.org/hyperkitty/list/kbuild-...@lists.01.org > > > > > _______________________________________________ > > > > > kbuild mailing list -- kbu...@lists.01.org > > > > > To unsubscribe send an email to kbuild-le...@lists.01.org > > > > > > > > -- > > > > You received this message because you are subscribed to the Google > > > > Groups "Clang Built Linux" group. > > > > To unsubscribe from this group and stop receiving emails from it, send > > > > an email to clang-built-linux+unsubscr...@googlegroups.com. > > > > To view this discussion on the web visit > > > > https://groups.google.com/d/msgid/clang-built-linux/CANn89i%2Bs%3DoFTq6KFhT0z1CReQPZCGoEPShEFHVvXAmC9sUDH7g%40mail.gmail.com. > > > > > > > > > > > > -- > > > Thanks, > > > ~Nick Desaulniers > > -- > You received this message because you are subscribed to the Google Groups > "Clang Built Linux" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to clang-built-linux+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/clang-built-linux/CANn89iLnmKaa30tPoyK_GrJr-_ALYBwjcJ0pBcvmckA9PAvrgA%40mail.gmail.com. -- Thanks, ~Nick Desaulniers
