From d8dd25a461e4eec7190cb9d66616aceacc5110ad Mon Sep 17 00:00:00 2001 From: Josh Poimboeuf Date: Sat, 25 Apr 2020 05:03:00 -0500 Subject: objtool: Fix stack offset tracking for indirect CFAs When the current frame address (CFA) is stored on the stack (i.e., cfa->base == CFI_SP_INDIRECT), objtool neglects to adjust the stack offset when there are subsequent pushes or pops. This results in bad ORC data at the end of the ENTER_IRQ_STACK macro, when it puts the previous stack pointer on the stack and does a subsequent push. This fixes the following unwinder warning: WARNING: can't dereference registers at 00000000f0a6bdba for ip interrupt_entry+0x9f/0xa0 Fixes: 627fce14809b ("objtool: Add ORC unwind table generation") Reported-by: Vince Weaver Reported-by: Dave Jones Reported-by: Steven Rostedt Reported-by: Vegard Nossum Reported-by: Joe Mario Reviewed-by: Miroslav Benes Signed-off-by: Josh Poimboeuf Signed-off-by: Ingo Molnar Cc: Andy Lutomirski Cc: Jann Horn Cc: Peter Zijlstra Cc: Thomas Gleixner Link: https://lore.kernel.org/r/853d5d691b29e250333332f09b8e27410b2d9924.1587808742.git.jpoimboe@redhat.com --- tools/objtool/check.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/objtool/check.c b/tools/objtool/check.c index 4b170fd08a28..e7184641a40c 100644 --- a/tools/objtool/check.c +++ b/tools/objtool/check.c @@ -1449,7 +1449,7 @@ static int update_insn_state_regs(struct instruction *insn, struct insn_state *s struct cfi_reg *cfa = &state->cfa; struct stack_op *op = &insn->stack_op; - if (cfa->base != CFI_SP) + if (cfa->base != CFI_SP && cfa->base != CFI_SP_INDIRECT) return 0; /* push */ -- cgit v1.2.3 From 53fb6e990d782ded62d7c76d566e107c03393b74 Mon Sep 17 00:00:00 2001 From: Josh Poimboeuf Date: Sat, 25 Apr 2020 14:19:01 -0500 Subject: objtool: Fix infinite loop in for_offset_range() Randy reported that objtool got stuck in an infinite loop when processing drivers/i2c/busses/i2c-parport.o. It was caused by the following code: 00000000000001fd : 1fd: 48 b8 00 00 00 00 00 movabs $0x0,%rax 204: 00 00 00 1ff: R_X86_64_64 .rodata-0x8 207: 41 55 push %r13 209: 41 89 f5 mov %esi,%r13d 20c: 41 54 push %r12 20e: 49 89 fc mov %rdi,%r12 211: 55 push %rbp 212: 48 89 d5 mov %rdx,%rbp 215: 53 push %rbx 216: 0f b6 5a 01 movzbl 0x1(%rdx),%ebx 21a: 48 8d 34 dd 00 00 00 lea 0x0(,%rbx,8),%rsi 221: 00 21e: R_X86_64_32S .rodata 222: 48 89 f1 mov %rsi,%rcx 225: 48 29 c1 sub %rax,%rcx find_jump_table() saw the .rodata reference and tried to find a jump table associated with it (though there wasn't one). The -0x8 rela addend is unusual. It caused find_jump_table() to send a negative table_offset (unsigned 0xfffffffffffffff8) to find_rela_by_dest(). The negative offset should have been harmless, but it actually threw for_offset_range() for a loop... literally. When the mask value got incremented past the end value, it also wrapped to zero, causing the loop exit condition to remain true forever. Prevent this scenario from happening by ensuring the incremented value is always >= the starting value. Fixes: 74b873e49d92 ("objtool: Optimize find_rela_by_dest_range()") Reported-by: Randy Dunlap Tested-by: Randy Dunlap Acked-by: Randy Dunlap Signed-off-by: Josh Poimboeuf Signed-off-by: Ingo Molnar Cc: Julien Thierry Cc: Miroslav Benes Cc: Peter Zijlstra Link: https://lore.kernel.org/r/02b719674b031800b61e33c30b2e823183627c19.1587842122.git.jpoimboe@redhat.com --- tools/objtool/elf.h | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/objtool/elf.h b/tools/objtool/elf.h index ebbb10c61e24..c227a2e55751 100644 --- a/tools/objtool/elf.h +++ b/tools/objtool/elf.h @@ -87,9 +87,10 @@ struct elf { #define OFFSET_STRIDE (1UL << OFFSET_STRIDE_BITS) #define OFFSET_STRIDE_MASK (~(OFFSET_STRIDE - 1)) -#define for_offset_range(_offset, _start, _end) \ - for (_offset = ((_start) & OFFSET_STRIDE_MASK); \ - _offset <= ((_end) & OFFSET_STRIDE_MASK); \ +#define for_offset_range(_offset, _start, _end) \ + for (_offset = ((_start) & OFFSET_STRIDE_MASK); \ + _offset >= ((_start) & OFFSET_STRIDE_MASK) && \ + _offset <= ((_end) & OFFSET_STRIDE_MASK); \ _offset += OFFSET_STRIDE) static inline u32 sec_offset_hash(struct section *sec, unsigned long offset) -- cgit v1.2.3 From 57c4cfd4a2eef8f94052bd7c0fce0981f74fb213 Mon Sep 17 00:00:00 2001 From: Alan Maguire Date: Wed, 19 Feb 2020 09:33:29 +0000 Subject: ftrace/selftests: workaround cgroup RT scheduling issues wakeup_rt.tc and wakeup.tc tests in tracers/ subdirectory fail due to the chrt command returning: chrt: failed to set pid 0's policy: Operation not permitted. To work around this, temporarily disable grout RT scheduling during ftracetest execution. Restore original value on test run completion. With these changes in place, both tests consistently pass. Fixes: c575dea2c1a5 ("selftests/ftrace: Add wakeup_rt tracer testcase") Fixes: c1edd060b413 ("selftests/ftrace: Add wakeup tracer testcase") Signed-off-by: Alan Maguire Acked-by: Steven Rostedt (VMware) Signed-off-by: Shuah Khan --- tools/testing/selftests/ftrace/ftracetest | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) (limited to 'tools') diff --git a/tools/testing/selftests/ftrace/ftracetest b/tools/testing/selftests/ftrace/ftracetest index 063ecb290a5a..144308a757b7 100755 --- a/tools/testing/selftests/ftrace/ftracetest +++ b/tools/testing/selftests/ftrace/ftracetest @@ -29,8 +29,25 @@ err_ret=1 # kselftest skip code is 4 err_skip=4 +# cgroup RT scheduling prevents chrt commands from succeeding, which +# induces failures in test wakeup tests. Disable for the duration of +# the tests. + +readonly sched_rt_runtime=/proc/sys/kernel/sched_rt_runtime_us + +sched_rt_runtime_orig=$(cat $sched_rt_runtime) + +setup() { + echo -1 > $sched_rt_runtime +} + +cleanup() { + echo $sched_rt_runtime_orig > $sched_rt_runtime +} + errexit() { # message echo "Error: $1" 1>&2 + cleanup exit $err_ret } @@ -39,6 +56,8 @@ if [ `id -u` -ne 0 ]; then errexit "this must be run by root user" fi +setup + # Utilities absdir() { # file_path (cd `dirname $1`; pwd) @@ -235,6 +254,7 @@ TOTAL_RESULT=0 INSTANCE= CASENO=0 + testcase() { # testfile CASENO=$((CASENO+1)) desc=`grep "^#[ \t]*description:" $1 | cut -f2 -d:` @@ -406,5 +426,7 @@ prlog "# of unsupported: " `echo $UNSUPPORTED_CASES | wc -w` prlog "# of xfailed: " `echo $XFAILED_CASES | wc -w` prlog "# of undefined(test bug): " `echo $UNDEFINED_CASES | wc -w` +cleanup + # if no error, return 0 exit $TOTAL_RESULT -- cgit v1.2.3 From b730d668138cb3dd9ce78f8003986d1adae5523a Mon Sep 17 00:00:00 2001 From: Alan Maguire Date: Wed, 19 Feb 2020 09:33:30 +0000 Subject: ftrace/selftest: make unresolved cases cause failure if --fail-unresolved set Currently, ftracetest will return 1 (failure) if any unresolved cases are encountered. The unresolved status results from modules and programs not being available, and as such does not indicate any issues with ftrace itself. As such, change the behaviour of ftracetest in line with unsupported cases; if unsupported cases happen, ftracetest still returns 0 unless --fail-unsupported. Here --fail-unresolved is added and the default is to return 0 if unresolved results occur. Signed-off-by: Alan Maguire Acked-by: Masami Hiramatsu Acked-by: Steven Rostedt (VMware) Signed-off-by: Shuah Khan --- tools/testing/selftests/ftrace/ftracetest | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/ftrace/ftracetest b/tools/testing/selftests/ftrace/ftracetest index 144308a757b7..19e9236dec5e 100755 --- a/tools/testing/selftests/ftrace/ftracetest +++ b/tools/testing/selftests/ftrace/ftracetest @@ -17,6 +17,7 @@ echo " -v|--verbose Increase verbosity of test messages" echo " -vv Alias of -v -v (Show all results in stdout)" echo " -vvv Alias of -v -v -v (Show all commands immediately)" echo " --fail-unsupported Treat UNSUPPORTED as a failure" +echo " --fail-unresolved Treat UNRESOLVED as a failure" echo " -d|--debug Debug mode (trace all shell commands)" echo " -l|--logdir Save logs on the " echo " If is -, all logs output in console only" @@ -112,6 +113,10 @@ parse_opts() { # opts UNSUPPORTED_RESULT=1 shift 1 ;; + --fail-unresolved) + UNRESOLVED_RESULT=1 + shift 1 + ;; --logdir|-l) LOG_DIR=$2 shift 2 @@ -176,6 +181,7 @@ KEEP_LOG=0 DEBUG=0 VERBOSE=0 UNSUPPORTED_RESULT=0 +UNRESOLVED_RESULT=0 STOP_FAILURE=0 # Parse command-line options parse_opts $* @@ -280,7 +286,7 @@ eval_result() { # sigval $UNRESOLVED) prlog " [${color_blue}UNRESOLVED${color_reset}]" UNRESOLVED_CASES="$UNRESOLVED_CASES $CASENO" - return 1 # this is a kind of bug.. something happened. + return $UNRESOLVED_RESULT # depends on use case ;; $UNTESTED) prlog " [${color_blue}UNTESTED${color_reset}]" -- cgit v1.2.3 From 6734d211feaec0c24f90da77313dbe704437d8a8 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Fri, 1 May 2020 22:37:41 +0900 Subject: selftests/ftrace: Make XFAIL green color Since XFAIL (Expected Failure) is expected to fail the test, which means that test case works as we expected. IOW, XFAIL is same as PASS. So make it green. Signed-off-by: Masami Hiramatsu Acked-by: Steven Rostedt (VMware) Signed-off-by: Shuah Khan --- tools/testing/selftests/ftrace/ftracetest | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/ftrace/ftracetest b/tools/testing/selftests/ftrace/ftracetest index 19e9236dec5e..a4605b5ee66d 100755 --- a/tools/testing/selftests/ftrace/ftracetest +++ b/tools/testing/selftests/ftrace/ftracetest @@ -299,7 +299,7 @@ eval_result() { # sigval return $UNSUPPORTED_RESULT # depends on use case ;; $XFAIL) - prlog " [${color_red}XFAIL${color_reset}]" + prlog " [${color_green}XFAIL${color_reset}]" XFAILED_CASES="$XFAILED_CASES $CASENO" return 0 ;; -- cgit v1.2.3 From 66d69e081b526b6a6031f0d3ca8ddff71e5707a5 Mon Sep 17 00:00:00 2001 From: Shuah Khan Date: Mon, 27 Apr 2020 18:11:07 -0600 Subject: selftests: fix kvm relocatable native/cross builds and installs kvm test Makefile doesn't fully support cross-builds and installs. UNAME_M = $(shell uname -m) variable is used to define the target programs and libraries to be built from arch specific sources in sub-directories. For cross-builds to work, UNAME_M has to map to ARCH and arch specific directories and targets in this Makefile. UNAME_M variable to used to run the compiles pointing to the right arch directories and build the right targets for these supported architectures. TEST_GEN_PROGS and LIBKVM are set using UNAME_M variable. LINUX_TOOL_ARCH_INCLUDE is set using ARCH variable. x86_64 targets are named to include x86_64 as a suffix and directories for includes are in x86_64 sub-directory. s390x and aarch64 follow the same convention. "uname -m" doesn't result in the correct mapping for s390x and aarch64. Fix it to set UNAME_M correctly for s390x and aarch64 cross-builds. In addition, Makefile doesn't create arch sub-directories in the case of relocatable builds and test programs under s390x and x86_64 directories fail to build. This is a problem for native and cross-builds. Fix it to create all necessary directories keying off of TEST_GEN_PROGS. The following use-cases work with this change: Native x86_64: make O=/tmp/kselftest -C tools/testing/selftests TARGETS=kvm install \ INSTALL_PATH=$HOME/x86_64 arm64 cross-build: make O=$HOME/arm64_build/ ARCH=arm64 HOSTCC=gcc \ CROSS_COMPILE=aarch64-linux-gnu- defconfig make O=$HOME/arm64_build/ ARCH=arm64 HOSTCC=gcc \ CROSS_COMPILE=aarch64-linux-gnu- all make kselftest-install TARGETS=kvm O=$HOME/arm64_build ARCH=arm64 \ HOSTCC=gcc CROSS_COMPILE=aarch64-linux-gnu- s390x cross-build: make O=$HOME/s390x_build/ ARCH=s390 HOSTCC=gcc \ CROSS_COMPILE=s390x-linux-gnu- defconfig make O=$HOME/s390x_build/ ARCH=s390 HOSTCC=gcc \ CROSS_COMPILE=s390x-linux-gnu- all make kselftest-install TARGETS=kvm O=$HOME/s390x_build/ ARCH=s390 \ HOSTCC=gcc CROSS_COMPILE=s390x-linux-gnu- all No regressions in the following use-cases: make -C tools/testing/selftests TARGETS=kvm make kselftest-all TARGETS=kvm Signed-off-by: Shuah Khan --- tools/testing/selftests/kvm/Makefile | 29 ++++++++++++++++++++++++++++- 1 file changed, 28 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/kvm/Makefile b/tools/testing/selftests/kvm/Makefile index 712a2ddd2a27..b728c0a0f9b2 100644 --- a/tools/testing/selftests/kvm/Makefile +++ b/tools/testing/selftests/kvm/Makefile @@ -5,8 +5,34 @@ all: top_srcdir = ../../../.. KSFT_KHDR_INSTALL := 1 + +# For cross-builds to work, UNAME_M has to map to ARCH and arch specific +# directories and targets in this Makefile. "uname -m" doesn't map to +# arch specific sub-directory names. +# +# UNAME_M variable to used to run the compiles pointing to the right arch +# directories and build the right targets for these supported architectures. +# +# TEST_GEN_PROGS and LIBKVM are set using UNAME_M variable. +# LINUX_TOOL_ARCH_INCLUDE is set using ARCH variable. +# +# x86_64 targets are named to include x86_64 as a suffix and directories +# for includes are in x86_64 sub-directory. s390x and aarch64 follow the +# same convention. "uname -m" doesn't result in the correct mapping for +# s390x and aarch64. +# +# No change necessary for x86_64 UNAME_M := $(shell uname -m) +# Set UNAME_M for arm64 compile/install to work +ifeq ($(ARCH),arm64) + UNAME_M := aarch64 +endif +# Set UNAME_M s390x compile/install to work +ifeq ($(ARCH),s390) + UNAME_M := s390x +endif + LIBKVM = lib/assert.c lib/elf.c lib/io.c lib/kvm_util.c lib/sparsebit.c lib/test_util.c LIBKVM_x86_64 = lib/x86_64/processor.c lib/x86_64/vmx.c lib/x86_64/svm.c lib/x86_64/ucall.c LIBKVM_aarch64 = lib/aarch64/processor.c lib/aarch64/ucall.c @@ -53,7 +79,7 @@ LIBKVM += $(LIBKVM_$(UNAME_M)) INSTALL_HDR_PATH = $(top_srcdir)/usr LINUX_HDR_PATH = $(INSTALL_HDR_PATH)/include/ LINUX_TOOL_INCLUDE = $(top_srcdir)/tools/include -LINUX_TOOL_ARCH_INCLUDE = $(top_srcdir)/tools/arch/x86/include +LINUX_TOOL_ARCH_INCLUDE = $(top_srcdir)/tools/arch/$(ARCH)/include CFLAGS += -Wall -Wstrict-prototypes -Wuninitialized -O2 -g -std=gnu99 \ -fno-stack-protector -fno-PIE -I$(LINUX_TOOL_INCLUDE) \ -I$(LINUX_TOOL_ARCH_INCLUDE) -I$(LINUX_HDR_PATH) -Iinclude \ @@ -84,6 +110,7 @@ $(LIBKVM_OBJ): $(OUTPUT)/%.o: %.c $(OUTPUT)/libkvm.a: $(LIBKVM_OBJ) $(AR) crs $@ $^ +x := $(shell mkdir -p $(sort $(dir $(TEST_GEN_PROGS)))) all: $(STATIC_LIBS) $(TEST_GEN_PROGS): $(STATIC_LIBS) -- cgit v1.2.3 From 0b80f9866e6bbfb905140ed8787ff2af03652c0c Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 4 May 2020 19:27:54 -0400 Subject: iocost: protect iocg->abs_vdebt with iocg->waitq.lock abs_vdebt is an atomic_64 which tracks how much over budget a given cgroup is and controls the activation of use_delay mechanism. Once a cgroup goes over budget from forced IOs, it has to pay it back with its future budget. The progress guarantee on debt paying comes from the iocg being active - active iocgs are processed by the periodic timer, which ensures that as time passes the debts dissipate and the iocg returns to normal operation. However, both iocg activation and vdebt handling are asynchronous and a sequence like the following may happen. 1. The iocg is in the process of being deactivated by the periodic timer. 2. A bio enters ioc_rqos_throttle(), calls iocg_activate() which returns without anything because it still sees that the iocg is already active. 3. The iocg is deactivated. 4. The bio from #2 is over budget but needs to be forced. It increases abs_vdebt and goes over the threshold and enables use_delay. 5. IO control is enabled for the iocg's subtree and now IOs are attributed to the descendant cgroups and the iocg itself no longer issues IOs. This leaves the iocg with stuck abs_vdebt - it has debt but inactive and no further IOs which can activate it. This can end up unduly punishing all the descendants cgroups. The usual throttling path has the same issue - the iocg must be active while throttled to ensure that future event will wake it up - and solves the problem by synchronizing the throttling path with a spinlock. abs_vdebt handling is another form of overage handling and shares a lot of characteristics including the fact that it isn't in the hottest path. This patch fixes the above and other possible races by strictly synchronizing abs_vdebt and use_delay handling with iocg->waitq.lock. Signed-off-by: Tejun Heo Reported-by: Vlad Dmitriev Cc: stable@vger.kernel.org # v5.4+ Fixes: e1518f63f246 ("blk-iocost: Don't let merges push vtime into the future") Signed-off-by: Jens Axboe --- block/blk-iocost.c | 117 +++++++++++++++++++++++++---------------- tools/cgroup/iocost_monitor.py | 7 ++- 2 files changed, 77 insertions(+), 47 deletions(-) (limited to 'tools') diff --git a/block/blk-iocost.c b/block/blk-iocost.c index 3ab0c1c704b6..7c1fe605d0d6 100644 --- a/block/blk-iocost.c +++ b/block/blk-iocost.c @@ -466,7 +466,7 @@ struct ioc_gq { */ atomic64_t vtime; atomic64_t done_vtime; - atomic64_t abs_vdebt; + u64 abs_vdebt; u64 last_vtime; /* @@ -1142,7 +1142,7 @@ static void iocg_kick_waitq(struct ioc_gq *iocg, struct ioc_now *now) struct iocg_wake_ctx ctx = { .iocg = iocg }; u64 margin_ns = (u64)(ioc->period_us * WAITQ_TIMER_MARGIN_PCT / 100) * NSEC_PER_USEC; - u64 abs_vdebt, vdebt, vshortage, expires, oexpires; + u64 vdebt, vshortage, expires, oexpires; s64 vbudget; u32 hw_inuse; @@ -1152,18 +1152,15 @@ static void iocg_kick_waitq(struct ioc_gq *iocg, struct ioc_now *now) vbudget = now->vnow - atomic64_read(&iocg->vtime); /* pay off debt */ - abs_vdebt = atomic64_read(&iocg->abs_vdebt); - vdebt = abs_cost_to_cost(abs_vdebt, hw_inuse); + vdebt = abs_cost_to_cost(iocg->abs_vdebt, hw_inuse); if (vdebt && vbudget > 0) { u64 delta = min_t(u64, vbudget, vdebt); u64 abs_delta = min(cost_to_abs_cost(delta, hw_inuse), - abs_vdebt); + iocg->abs_vdebt); atomic64_add(delta, &iocg->vtime); atomic64_add(delta, &iocg->done_vtime); - atomic64_sub(abs_delta, &iocg->abs_vdebt); - if (WARN_ON_ONCE(atomic64_read(&iocg->abs_vdebt) < 0)) - atomic64_set(&iocg->abs_vdebt, 0); + iocg->abs_vdebt -= abs_delta; } /* @@ -1219,12 +1216,18 @@ static bool iocg_kick_delay(struct ioc_gq *iocg, struct ioc_now *now, u64 cost) u64 expires, oexpires; u32 hw_inuse; + lockdep_assert_held(&iocg->waitq.lock); + /* debt-adjust vtime */ current_hweight(iocg, NULL, &hw_inuse); - vtime += abs_cost_to_cost(atomic64_read(&iocg->abs_vdebt), hw_inuse); + vtime += abs_cost_to_cost(iocg->abs_vdebt, hw_inuse); - /* clear or maintain depending on the overage */ - if (time_before_eq64(vtime, now->vnow)) { + /* + * Clear or maintain depending on the overage. Non-zero vdebt is what + * guarantees that @iocg is online and future iocg_kick_delay() will + * clear use_delay. Don't leave it on when there's no vdebt. + */ + if (!iocg->abs_vdebt || time_before_eq64(vtime, now->vnow)) { blkcg_clear_delay(blkg); return false; } @@ -1258,9 +1261,12 @@ static enum hrtimer_restart iocg_delay_timer_fn(struct hrtimer *timer) { struct ioc_gq *iocg = container_of(timer, struct ioc_gq, delay_timer); struct ioc_now now; + unsigned long flags; + spin_lock_irqsave(&iocg->waitq.lock, flags); ioc_now(iocg->ioc, &now); iocg_kick_delay(iocg, &now, 0); + spin_unlock_irqrestore(&iocg->waitq.lock, flags); return HRTIMER_NORESTART; } @@ -1368,14 +1374,13 @@ static void ioc_timer_fn(struct timer_list *timer) * should have woken up in the last period and expire idle iocgs. */ list_for_each_entry_safe(iocg, tiocg, &ioc->active_iocgs, active_list) { - if (!waitqueue_active(&iocg->waitq) && - !atomic64_read(&iocg->abs_vdebt) && !iocg_is_idle(iocg)) + if (!waitqueue_active(&iocg->waitq) && iocg->abs_vdebt && + !iocg_is_idle(iocg)) continue; spin_lock(&iocg->waitq.lock); - if (waitqueue_active(&iocg->waitq) || - atomic64_read(&iocg->abs_vdebt)) { + if (waitqueue_active(&iocg->waitq) || iocg->abs_vdebt) { /* might be oversleeping vtime / hweight changes, kick */ iocg_kick_waitq(iocg, &now); iocg_kick_delay(iocg, &now, 0); @@ -1718,28 +1723,49 @@ static void ioc_rqos_throttle(struct rq_qos *rqos, struct bio *bio) * tests are racy but the races aren't systemic - we only miss once * in a while which is fine. */ - if (!waitqueue_active(&iocg->waitq) && - !atomic64_read(&iocg->abs_vdebt) && + if (!waitqueue_active(&iocg->waitq) && !iocg->abs_vdebt && time_before_eq64(vtime + cost, now.vnow)) { iocg_commit_bio(iocg, bio, cost); return; } /* - * We're over budget. If @bio has to be issued regardless, - * remember the abs_cost instead of advancing vtime. - * iocg_kick_waitq() will pay off the debt before waking more IOs. + * We activated above but w/o any synchronization. Deactivation is + * synchronized with waitq.lock and we won't get deactivated as long + * as we're waiting or has debt, so we're good if we're activated + * here. In the unlikely case that we aren't, just issue the IO. + */ + spin_lock_irq(&iocg->waitq.lock); + + if (unlikely(list_empty(&iocg->active_list))) { + spin_unlock_irq(&iocg->waitq.lock); + iocg_commit_bio(iocg, bio, cost); + return; + } + + /* + * We're over budget. If @bio has to be issued regardless, remember + * the abs_cost instead of advancing vtime. iocg_kick_waitq() will pay + * off the debt before waking more IOs. + * * This way, the debt is continuously paid off each period with the - * actual budget available to the cgroup. If we just wound vtime, - * we would incorrectly use the current hw_inuse for the entire - * amount which, for example, can lead to the cgroup staying - * blocked for a long time even with substantially raised hw_inuse. + * actual budget available to the cgroup. If we just wound vtime, we + * would incorrectly use the current hw_inuse for the entire amount + * which, for example, can lead to the cgroup staying blocked for a + * long time even with substantially raised hw_inuse. + * + * An iocg with vdebt should stay online so that the timer can keep + * deducting its vdebt and [de]activate use_delay mechanism + * accordingly. We don't want to race against the timer trying to + * clear them and leave @iocg inactive w/ dangling use_delay heavily + * penalizing the cgroup and its descendants. */ if (bio_issue_as_root_blkg(bio) || fatal_signal_pending(current)) { - atomic64_add(abs_cost, &iocg->abs_vdebt); + iocg->abs_vdebt += abs_cost; if (iocg_kick_delay(iocg, &now, cost)) blkcg_schedule_throttle(rqos->q, (bio->bi_opf & REQ_SWAP) == REQ_SWAP); + spin_unlock_irq(&iocg->waitq.lock); return; } @@ -1756,20 +1782,6 @@ static void ioc_rqos_throttle(struct rq_qos *rqos, struct bio *bio) * All waiters are on iocg->waitq and the wait states are * synchronized using waitq.lock. */ - spin_lock_irq(&iocg->waitq.lock); - - /* - * We activated above but w/o any synchronization. Deactivation is - * synchronized with waitq.lock and we won't get deactivated as - * long as we're waiting, so we're good if we're activated here. - * In the unlikely case that we are deactivated, just issue the IO. - */ - if (unlikely(list_empty(&iocg->active_list))) { - spin_unlock_irq(&iocg->waitq.lock); - iocg_commit_bio(iocg, bio, cost); - return; - } - init_waitqueue_func_entry(&wait.wait, iocg_wake_fn); wait.wait.private = current; wait.bio = bio; @@ -1801,6 +1813,7 @@ static void ioc_rqos_merge(struct rq_qos *rqos, struct request *rq, struct ioc_now now; u32 hw_inuse; u64 abs_cost, cost; + unsigned long flags; /* bypass if disabled or for root cgroup */ if (!ioc->enabled || !iocg->level) @@ -1820,15 +1833,28 @@ static void ioc_rqos_merge(struct rq_qos *rqos, struct request *rq, iocg->cursor = bio_end; /* - * Charge if there's enough vtime budget and the existing request - * has cost assigned. Otherwise, account it as debt. See debt - * handling in ioc_rqos_throttle() for details. + * Charge if there's enough vtime budget and the existing request has + * cost assigned. */ if (rq->bio && rq->bio->bi_iocost_cost && - time_before_eq64(atomic64_read(&iocg->vtime) + cost, now.vnow)) + time_before_eq64(atomic64_read(&iocg->vtime) + cost, now.vnow)) { iocg_commit_bio(iocg, bio, cost); - else - atomic64_add(abs_cost, &iocg->abs_vdebt); + return; + } + + /* + * Otherwise, account it as debt if @iocg is online, which it should + * be for the vast majority of cases. See debt handling in + * ioc_rqos_throttle() for details. + */ + spin_lock_irqsave(&iocg->waitq.lock, flags); + if (likely(!list_empty(&iocg->active_list))) { + iocg->abs_vdebt += abs_cost; + iocg_kick_delay(iocg, &now, cost); + } else { + iocg_commit_bio(iocg, bio, cost); + } + spin_unlock_irqrestore(&iocg->waitq.lock, flags); } static void ioc_rqos_done_bio(struct rq_qos *rqos, struct bio *bio) @@ -1998,7 +2024,6 @@ static void ioc_pd_init(struct blkg_policy_data *pd) iocg->ioc = ioc; atomic64_set(&iocg->vtime, now.vnow); atomic64_set(&iocg->done_vtime, now.vnow); - atomic64_set(&iocg->abs_vdebt, 0); atomic64_set(&iocg->active_period, atomic64_read(&ioc->cur_period)); INIT_LIST_HEAD(&iocg->active_list); iocg->hweight_active = HWEIGHT_WHOLE; diff --git a/tools/cgroup/iocost_monitor.py b/tools/cgroup/iocost_monitor.py index 7427a5ee761b..9d8e9613008a 100644 --- a/tools/cgroup/iocost_monitor.py +++ b/tools/cgroup/iocost_monitor.py @@ -159,7 +159,12 @@ class IocgStat: else: self.inflight_pct = 0 - self.debt_ms = iocg.abs_vdebt.counter.value_() / VTIME_PER_USEC / 1000 + # vdebt used to be an atomic64_t and is now u64, support both + try: + self.debt_ms = iocg.abs_vdebt.counter.value_() / VTIME_PER_USEC / 1000 + except: + self.debt_ms = iocg.abs_vdebt.value_() / VTIME_PER_USEC / 1000 + self.use_delay = blkg.use_delay.counter.value_() self.delay_ms = blkg.delay_nsec.counter.value_() / 1_000_000 -- cgit v1.2.3 From 8ffdaf9155ebe517cdec5edbcca19ba6e7ee9c3c Mon Sep 17 00:00:00 2001 From: Peter Xu Date: Mon, 4 May 2020 18:06:07 -0400 Subject: KVM: selftests: Fix build for evmcs.h I got this error when building kvm selftests: /usr/bin/ld: /home/xz/git/linux/tools/testing/selftests/kvm/libkvm.a(vmx.o):/home/xz/git/linux/tools/testing/selftests/kvm/include/evmcs.h:222: multiple definition of `current_evmcs'; /tmp/cco1G48P.o:/home/xz/git/linux/tools/testing/selftests/kvm/include/evmcs.h:222: first defined here /usr/bin/ld: /home/xz/git/linux/tools/testing/selftests/kvm/libkvm.a(vmx.o):/home/xz/git/linux/tools/testing/selftests/kvm/include/evmcs.h:223: multiple definition of `current_vp_assist'; /tmp/cco1G48P.o:/home/xz/git/linux/tools/testing/selftests/kvm/include/evmcs.h:223: first defined here I think it's because evmcs.h is included both in a test file and a lib file so the structs have multiple declarations when linking. After all it's not a good habit to declare structs in the header files. Cc: Vitaly Kuznetsov Signed-off-by: Peter Xu Message-Id: <20200504220607.99627-1-peterx@redhat.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/include/evmcs.h | 4 ++-- tools/testing/selftests/kvm/lib/x86_64/vmx.c | 3 +++ 2 files changed, 5 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/kvm/include/evmcs.h b/tools/testing/selftests/kvm/include/evmcs.h index d8f4d6bfe05d..a034438b6266 100644 --- a/tools/testing/selftests/kvm/include/evmcs.h +++ b/tools/testing/selftests/kvm/include/evmcs.h @@ -219,8 +219,8 @@ struct hv_enlightened_vmcs { #define HV_X64_MSR_VP_ASSIST_PAGE_ADDRESS_MASK \ (~((1ull << HV_X64_MSR_VP_ASSIST_PAGE_ADDRESS_SHIFT) - 1)) -struct hv_enlightened_vmcs *current_evmcs; -struct hv_vp_assist_page *current_vp_assist; +extern struct hv_enlightened_vmcs *current_evmcs; +extern struct hv_vp_assist_page *current_vp_assist; int vcpu_enable_evmcs(struct kvm_vm *vm, int vcpu_id); diff --git a/tools/testing/selftests/kvm/lib/x86_64/vmx.c b/tools/testing/selftests/kvm/lib/x86_64/vmx.c index 6f17f69394be..4ae104f6ce69 100644 --- a/tools/testing/selftests/kvm/lib/x86_64/vmx.c +++ b/tools/testing/selftests/kvm/lib/x86_64/vmx.c @@ -17,6 +17,9 @@ bool enable_evmcs; +struct hv_enlightened_vmcs *current_evmcs; +struct hv_vp_assist_page *current_vp_assist; + struct eptPageTableEntry { uint64_t readable:1; uint64_t writable:1; -- cgit v1.2.3 From 1119d265bc20226c241e5540fc8a246d9e30b272 Mon Sep 17 00:00:00 2001 From: Josh Poimboeuf Date: Tue, 28 Apr 2020 16:45:16 -0500 Subject: objtool: Fix infinite loop in find_jump_table() Kristen found a hang in objtool when building with -ffunction-sections. It was caused by evergreen_pcie_gen2_enable.cold() being laid out immediately before evergreen_pcie_gen2_enable(). Since their "pfunc" is always the same, find_jump_table() got into an infinite loop because it didn't recognize the boundary between the two functions. Fix that with a new prev_insn_same_sym() helper, which doesn't cross subfunction boundaries. Reported-by: Kristen Carlson Accardi Signed-off-by: Josh Poimboeuf Signed-off-by: Peter Zijlstra (Intel) Link: https://lkml.kernel.org/r/378b51c9d9c894dc3294bc460b4b0869e950b7c5.1588110291.git.jpoimboe@redhat.com --- tools/objtool/check.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/objtool/check.c b/tools/objtool/check.c index 4b170fd08a28..0e8f9a32e4d1 100644 --- a/tools/objtool/check.c +++ b/tools/objtool/check.c @@ -72,6 +72,17 @@ static struct instruction *next_insn_same_func(struct objtool_file *file, return find_insn(file, func->cfunc->sec, func->cfunc->offset); } +static struct instruction *prev_insn_same_sym(struct objtool_file *file, + struct instruction *insn) +{ + struct instruction *prev = list_prev_entry(insn, list); + + if (&prev->list != &file->insn_list && prev->func == insn->func) + return prev; + + return NULL; +} + #define func_for_each_insn(file, func, insn) \ for (insn = find_insn(file, func->sec, func->offset); \ insn; \ @@ -1050,8 +1061,8 @@ static struct rela *find_jump_table(struct objtool_file *file, * it. */ for (; - &insn->list != &file->insn_list && insn->func && insn->func->pfunc == func; - insn = insn->first_jump_src ?: list_prev_entry(insn, list)) { + insn && insn->func && insn->func->pfunc == func; + insn = insn->first_jump_src ?: prev_insn_same_sym(file, insn)) { if (insn != orig_insn && insn->type == INSN_JUMP_DYNAMIC) break; -- cgit v1.2.3 From 8842604446d1f005abcbf8c63c12eabdb5695094 Mon Sep 17 00:00:00 2001 From: Yunfeng Ye Date: Thu, 7 May 2020 17:23:36 +0800 Subject: tools/bootconfig: Fix resource leak in apply_xbc() Fix the @data and @fd allocations that are leaked in the error path of apply_xbc(). Link: http://lkml.kernel.org/r/583a49c9-c27a-931d-e6c2-6f63a4b18bea@huawei.com Acked-by: Masami Hiramatsu Signed-off-by: Yunfeng Ye Signed-off-by: Steven Rostedt (VMware) --- tools/bootconfig/main.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/bootconfig/main.c b/tools/bootconfig/main.c index 16b9a420e6fd..001076c51712 100644 --- a/tools/bootconfig/main.c +++ b/tools/bootconfig/main.c @@ -314,6 +314,7 @@ int apply_xbc(const char *path, const char *xbc_path) ret = delete_xbc(path); if (ret < 0) { pr_err("Failed to delete previous boot config: %d\n", ret); + free(data); return ret; } @@ -321,24 +322,26 @@ int apply_xbc(const char *path, const char *xbc_path) fd = open(path, O_RDWR | O_APPEND); if (fd < 0) { pr_err("Failed to open %s: %d\n", path, fd); + free(data); return fd; } /* TODO: Ensure the @path is initramfs/initrd image */ ret = write(fd, data, size + 8); if (ret < 0) { pr_err("Failed to apply a boot config: %d\n", ret); - return ret; + goto out; } /* Write a magic word of the bootconfig */ ret = write(fd, BOOTCONFIG_MAGIC, BOOTCONFIG_MAGIC_LEN); if (ret < 0) { pr_err("Failed to apply a boot config magic: %d\n", ret); - return ret; + goto out; } +out: close(fd); free(data); - return 0; + return ret; } int usage(void) -- cgit v1.2.3 From 474328c06e3ee75bb6b92826fec90fdc8ef3c573 Mon Sep 17 00:00:00 2001 From: Roman Penyaev Date: Thu, 7 May 2020 18:36:13 -0700 Subject: kselftests: introduce new epoll60 testcase for catching lost wakeups This test case catches lost wake up introduced by commit 339ddb53d373 ("fs/epoll: remove unnecessary wakeups of nested epoll") The test is simple: we have 10 threads and 10 event fds. Each thread can harvest only 1 event. 1 producer fires all 10 events at once and waits that all 10 events will be observed by 10 threads. In case of lost wakeup epoll_wait() will timeout and 0 will be returned. Test case catches two sort of problems: forgotten wakeup on event, which hits the ->ovflist list, this problem was fixed by: 5a2513239750 ("eventpoll: fix missing wakeup for ovflist in ep_poll_callback") the other problem is when several sequential events hit the same waiting thread, thus other waiters get no wakeups. Problem is fixed in the following patch. Signed-off-by: Roman Penyaev Signed-off-by: Andrew Morton Cc: Khazhismel Kumykov Cc: Alexander Viro Cc: Heiher Cc: Jason Baron Link: http://lkml.kernel.org/r/20200430130326.1368509-1-rpenyaev@suse.de Signed-off-by: Linus Torvalds --- .../filesystems/epoll/epoll_wakeup_test.c | 146 +++++++++++++++++++++ 1 file changed, 146 insertions(+) (limited to 'tools') diff --git a/tools/testing/selftests/filesystems/epoll/epoll_wakeup_test.c b/tools/testing/selftests/filesystems/epoll/epoll_wakeup_test.c index 11eee0b60040..d979ff14775a 100644 --- a/tools/testing/selftests/filesystems/epoll/epoll_wakeup_test.c +++ b/tools/testing/selftests/filesystems/epoll/epoll_wakeup_test.c @@ -3,6 +3,7 @@ #define _GNU_SOURCE #include #include +#include #include #include #include @@ -3136,4 +3137,149 @@ TEST(epoll59) close(ctx.sfd[0]); } +enum { + EPOLL60_EVENTS_NR = 10, +}; + +struct epoll60_ctx { + volatile int stopped; + int ready; + int waiters; + int epfd; + int evfd[EPOLL60_EVENTS_NR]; +}; + +static void *epoll60_wait_thread(void *ctx_) +{ + struct epoll60_ctx *ctx = ctx_; + struct epoll_event e; + sigset_t sigmask; + uint64_t v; + int ret; + + /* Block SIGUSR1 */ + sigemptyset(&sigmask); + sigaddset(&sigmask, SIGUSR1); + sigprocmask(SIG_SETMASK, &sigmask, NULL); + + /* Prepare empty mask for epoll_pwait() */ + sigemptyset(&sigmask); + + while (!ctx->stopped) { + /* Mark we are ready */ + __atomic_fetch_add(&ctx->ready, 1, __ATOMIC_ACQUIRE); + + /* Start when all are ready */ + while (__atomic_load_n(&ctx->ready, __ATOMIC_ACQUIRE) && + !ctx->stopped); + + /* Account this waiter */ + __atomic_fetch_add(&ctx->waiters, 1, __ATOMIC_ACQUIRE); + + ret = epoll_pwait(ctx->epfd, &e, 1, 2000, &sigmask); + if (ret != 1) { + /* We expect only signal delivery on stop */ + assert(ret < 0 && errno == EINTR && "Lost wakeup!\n"); + assert(ctx->stopped); + break; + } + + ret = read(e.data.fd, &v, sizeof(v)); + /* Since we are on ET mode, thus each thread gets its own fd. */ + assert(ret == sizeof(v)); + + __atomic_fetch_sub(&ctx->waiters, 1, __ATOMIC_RELEASE); + } + + return NULL; +} + +static inline unsigned long long msecs(void) +{ + struct timespec ts; + unsigned long long msecs; + + clock_gettime(CLOCK_REALTIME, &ts); + msecs = ts.tv_sec * 1000ull; + msecs += ts.tv_nsec / 1000000ull; + + return msecs; +} + +static inline int count_waiters(struct epoll60_ctx *ctx) +{ + return __atomic_load_n(&ctx->waiters, __ATOMIC_ACQUIRE); +} + +TEST(epoll60) +{ + struct epoll60_ctx ctx = { 0 }; + pthread_t waiters[ARRAY_SIZE(ctx.evfd)]; + struct epoll_event e; + int i, n, ret; + + signal(SIGUSR1, signal_handler); + + ctx.epfd = epoll_create1(0); + ASSERT_GE(ctx.epfd, 0); + + /* Create event fds */ + for (i = 0; i < ARRAY_SIZE(ctx.evfd); i++) { + ctx.evfd[i] = eventfd(0, EFD_NONBLOCK); + ASSERT_GE(ctx.evfd[i], 0); + + e.events = EPOLLIN | EPOLLET; + e.data.fd = ctx.evfd[i]; + ASSERT_EQ(epoll_ctl(ctx.epfd, EPOLL_CTL_ADD, ctx.evfd[i], &e), 0); + } + + /* Create waiter threads */ + for (i = 0; i < ARRAY_SIZE(waiters); i++) + ASSERT_EQ(pthread_create(&waiters[i], NULL, + epoll60_wait_thread, &ctx), 0); + + for (i = 0; i < 300; i++) { + uint64_t v = 1, ms; + + /* Wait for all to be ready */ + while (__atomic_load_n(&ctx.ready, __ATOMIC_ACQUIRE) != + ARRAY_SIZE(ctx.evfd)) + ; + + /* Steady, go */ + __atomic_fetch_sub(&ctx.ready, ARRAY_SIZE(ctx.evfd), + __ATOMIC_ACQUIRE); + + /* Wait all have gone to kernel */ + while (count_waiters(&ctx) != ARRAY_SIZE(ctx.evfd)) + ; + + /* 1ms should be enough to schedule away */ + usleep(1000); + + /* Quickly signal all handles at once */ + for (n = 0; n < ARRAY_SIZE(ctx.evfd); n++) { + ret = write(ctx.evfd[n], &v, sizeof(v)); + ASSERT_EQ(ret, sizeof(v)); + } + + /* Busy loop for 1s and wait for all waiters to wake up */ + ms = msecs(); + while (count_waiters(&ctx) && msecs() < ms + 1000) + ; + + ASSERT_EQ(count_waiters(&ctx), 0); + } + ctx.stopped = 1; + /* Stop waiters */ + for (i = 0; i < ARRAY_SIZE(waiters); i++) + ret = pthread_kill(waiters[i], SIGUSR1); + for (i = 0; i < ARRAY_SIZE(waiters); i++) + pthread_join(waiters[i], NULL); + + for (i = 0; i < ARRAY_SIZE(waiters); i++) + close(ctx.evfd[i]); + close(ctx.epfd); +} + TEST_HARNESS_MAIN -- cgit v1.2.3 From 8d58f222e85f01da0c0e1fc1e77986c86de889e2 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 7 May 2020 18:36:23 -0700 Subject: ubsan: disable UBSAN_ALIGNMENT under COMPILE_TEST The documentation for UBSAN_ALIGNMENT already mentions that it should not be used on all*config builds (and for efficient-unaligned-access architectures), so just refactor the Kconfig to correctly implement this so randconfigs will stop creating insane images that freak out objtool under CONFIG_UBSAN_TRAP (due to the false positives producing functions that never return, etc). Link: http://lkml.kernel.org/r/202005011433.C42EA3E2D@keescook Fixes: 0887a7ebc977 ("ubsan: add trap instrumentation option") Signed-off-by: Kees Cook Reported-by: Randy Dunlap Link: https://lore.kernel.org/linux-next/202004231224.D6B3B650@keescook/ Signed-off-by: Andrew Morton Cc: Josh Poimboeuf Cc: Stephen Rothwell Cc: Peter Zijlstra Cc: Andrey Ryabinin Signed-off-by: Linus Torvalds --- lib/Kconfig.ubsan | 15 ++++++--------- tools/testing/selftests/wireguard/qemu/debug.config | 1 - 2 files changed, 6 insertions(+), 10 deletions(-) (limited to 'tools') diff --git a/lib/Kconfig.ubsan b/lib/Kconfig.ubsan index 48469c95d78e..929211039bac 100644 --- a/lib/Kconfig.ubsan +++ b/lib/Kconfig.ubsan @@ -60,18 +60,15 @@ config UBSAN_SANITIZE_ALL Enabling this option will get kernel image size increased significantly. -config UBSAN_NO_ALIGNMENT - bool "Disable checking of pointers alignment" - default y if HAVE_EFFICIENT_UNALIGNED_ACCESS +config UBSAN_ALIGNMENT + bool "Enable checks for pointers alignment" + default !HAVE_EFFICIENT_UNALIGNED_ACCESS + depends on !X86 || !COMPILE_TEST help - This option disables the check of unaligned memory accesses. - This option should be used when building allmodconfig. - Disabling this option on architectures that support unaligned + This option enables the check of unaligned memory accesses. + Enabling this option on architectures that support unaligned accesses may produce a lot of false positives. -config UBSAN_ALIGNMENT - def_bool !UBSAN_NO_ALIGNMENT - config TEST_UBSAN tristate "Module for testing for undefined behavior detection" depends on m diff --git a/tools/testing/selftests/wireguard/qemu/debug.config b/tools/testing/selftests/wireguard/qemu/debug.config index 5909e7ef2a5c..9803dbb54181 100644 --- a/tools/testing/selftests/wireguard/qemu/debug.config +++ b/tools/testing/selftests/wireguard/qemu/debug.config @@ -25,7 +25,6 @@ CONFIG_KASAN=y CONFIG_KASAN_INLINE=y CONFIG_UBSAN=y CONFIG_UBSAN_SANITIZE_ALL=y -CONFIG_UBSAN_NO_ALIGNMENT=y CONFIG_UBSAN_NULL=y CONFIG_DEBUG_KMEMLEAK=y CONFIG_DEBUG_KMEMLEAK_EARLY_LOG_SIZE=8192 -- cgit v1.2.3 From 4bb9d46d47b105a774f9dca642f5271375bca4b2 Mon Sep 17 00:00:00 2001 From: John Stultz Date: Fri, 8 May 2020 04:56:10 +0000 Subject: kselftests: dmabuf-heaps: Fix confused return value on expected error testing When I added the expected error testing, I forgot I need to set the return to zero when we successfully see an error. Without this change we only end up testing a single heap before the test quits. Cc: Shuah Khan Cc: Sumit Semwal Cc: Benjamin Gaignard Cc: Brian Starkey Cc: Laura Abbott Cc: "Andrew F. Davis" Cc: linux-kselftest@vger.kernel.org Signed-off-by: John Stultz Signed-off-by: Shuah Khan --- tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c | 1 + 1 file changed, 1 insertion(+) (limited to 'tools') diff --git a/tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c b/tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c index cd5e1f602ac9..909da9cdda97 100644 --- a/tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c +++ b/tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c @@ -351,6 +351,7 @@ static int test_alloc_errors(char *heap_name) } printf("Expected error checking passed\n"); + ret = 0; out: if (dmabuf_fd >= 0) close(dmabuf_fd); -- cgit v1.2.3 From d8238f9eb6e0c175c8b657af20164eef6b30c71c Mon Sep 17 00:00:00 2001 From: "Gustavo A. R. Silva" Date: Thu, 7 May 2020 13:56:08 -0500 Subject: tools/testing: Replace zero-length array with flexible-array The current codebase makes use of the zero-length array language extension to the C90 standard, but the preferred mechanism to declare variable-length types such as these ones is a flexible array member[1][2], introduced in C99: struct foo { int stuff; struct boo array[]; }; By making use of the mechanism above, we will get a compiler warning in case the flexible array does not occur last in the structure, which will help us prevent some kind of undefined behavior bugs from being inadvertently introduced[3] to the codebase from now on. Also, notice that, dynamic memory allocations won't be affected by this change: "Flexible array members have incomplete type, and so the sizeof operator may not be applied. As a quirk of the original implementation of zero-length arrays, sizeof evaluates to zero."[1] sizeof(flexible-array-member) triggers a warning because flexible array members have incomplete type[1]. There are some instances of code in which the sizeof operator is being incorrectly/erroneously applied to zero-length arrays and the result is zero. Such instances may be hiding some bugs. So, this work (flexible-array member conversions) will also help to get completely rid of those sorts of issues. This issue was found with the help of Coccinelle. [1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html [2] https://github.com/KSPP/linux/issues/21 [3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour") Signed-off-by: Gustavo A. R. Silva Signed-off-by: Shuah Khan --- tools/testing/selftests/nsfs/pidns.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/nsfs/pidns.c b/tools/testing/selftests/nsfs/pidns.c index e0d86e1668c0..e3c772c6a7c7 100644 --- a/tools/testing/selftests/nsfs/pidns.c +++ b/tools/testing/selftests/nsfs/pidns.c @@ -27,7 +27,7 @@ #define __stack_aligned__ __attribute__((aligned(16))) struct cr_clone_arg { char stack[128] __stack_aligned__; - char stack_ptr[0]; + char stack_ptr[]; }; static int child(void *args) -- cgit v1.2.3 From adb571649c7ce7ee16022fa302b62043d1812b4b Mon Sep 17 00:00:00 2001 From: Po-Hsu Lin Date: Tue, 5 May 2020 18:14:45 +0800 Subject: selftests/ftrace: mark irqsoff_tracer.tc test as unresolved if the test module does not exist The UNRESOLVED state is much more apporiate than the UNSUPPORTED state for the absence of the test module, as it matches "test was set up incorrectly" situation in the README file. A possible scenario is that the function was enabled (supported by the kernel) but the module was not installed properly, in this case we cannot call this as UNSUPPORTED. This change also make it consistent with other module-related tests in ftrace. Signed-off-by: Po-Hsu Lin Acked-by: Steven Rostedt (VMware) Acked-by: Masami Hiramatsu Signed-off-by: Shuah Khan --- .../testing/selftests/ftrace/test.d/preemptirq/irqsoff_tracer.tc | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/ftrace/test.d/preemptirq/irqsoff_tracer.tc b/tools/testing/selftests/ftrace/test.d/preemptirq/irqsoff_tracer.tc index cbd174334a48..2b82c80edf69 100644 --- a/tools/testing/selftests/ftrace/test.d/preemptirq/irqsoff_tracer.tc +++ b/tools/testing/selftests/ftrace/test.d/preemptirq/irqsoff_tracer.tc @@ -17,7 +17,14 @@ unsup() { #msg exit_unsupported } -modprobe $MOD || unsup "$MOD module not available" +unres() { #msg + reset_tracer + rmmod $MOD || true + echo $1 + exit_unresolved +} + +modprobe $MOD || unres "$MOD module not available" rmmod $MOD grep -q "preemptoff" available_tracers || unsup "preemptoff tracer not enabled" -- cgit v1.2.3 From f131d9edc29d527df4b8f6840c340415f559f095 Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Fri, 8 May 2020 16:53:55 +1000 Subject: selftests/lkdtm: Don't clear dmesg when running tests It is Very Rude to clear dmesg in test scripts. That's because the script may be part of a larger test run, and clearing dmesg potentially destroys the output of other tests. We can avoid using dmesg -c by saving the content of dmesg before the test, and then using diff to compare that to the dmesg afterward, producing a log with just the added lines. Signed-off-by: Michael Ellerman Acked-by: Kees Cook Signed-off-by: Shuah Khan --- tools/testing/selftests/lkdtm/run.sh | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/lkdtm/run.sh b/tools/testing/selftests/lkdtm/run.sh index dadf819148a4..0b409e187c7b 100755 --- a/tools/testing/selftests/lkdtm/run.sh +++ b/tools/testing/selftests/lkdtm/run.sh @@ -59,23 +59,25 @@ if [ -z "$expect" ]; then expect="call trace:" fi -# Clear out dmesg for output reporting -dmesg -c >/dev/null - # Prepare log for report checking -LOG=$(mktemp --tmpdir -t lkdtm-XXXXXX) +LOG=$(mktemp --tmpdir -t lkdtm-log-XXXXXX) +DMESG=$(mktemp --tmpdir -t lkdtm-dmesg-XXXXXX) cleanup() { - rm -f "$LOG" + rm -f "$LOG" "$DMESG" } trap cleanup EXIT +# Save existing dmesg so we can detect new content below +dmesg > "$DMESG" + # Most shells yell about signals and we're expecting the "cat" process # to usually be killed by the kernel. So we have to run it in a sub-shell # and silence errors. ($SHELL -c 'cat <(echo '"$test"') >'"$TRIGGER" 2>/dev/null) || true # Record and dump the results -dmesg -c >"$LOG" +dmesg | diff --changed-group-format='%>' --unchanged-group-format='' "$DMESG" - > "$LOG" || true + cat "$LOG" # Check for expected output if egrep -qi "$expect" "$LOG" ; then -- cgit v1.2.3 From 851c4df54dc1bcae41d07e46e3d89e035b0a7140 Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Fri, 8 May 2020 16:53:56 +1000 Subject: selftests/lkdtm: Use grep -E instead of egrep shellcheck complains that egrep is deprecated, and the grep man page agrees. Use grep -E instead. Signed-off-by: Michael Ellerman Acked-by: Kees Cook Signed-off-by: Shuah Khan --- tools/testing/selftests/lkdtm/run.sh | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/lkdtm/run.sh b/tools/testing/selftests/lkdtm/run.sh index 0b409e187c7b..ee64ff8df8f4 100755 --- a/tools/testing/selftests/lkdtm/run.sh +++ b/tools/testing/selftests/lkdtm/run.sh @@ -25,13 +25,13 @@ fi # Figure out which test to run from our script name. test=$(basename $0 .sh) # Look up details about the test from master list of LKDTM tests. -line=$(egrep '^#?'"$test"'\b' tests.txt) +line=$(grep -E '^#?'"$test"'\b' tests.txt) if [ -z "$line" ]; then echo "Skipped: missing test '$test' in tests.txt" exit $KSELFTEST_SKIP_TEST fi # Check that the test is known to LKDTM. -if ! egrep -q '^'"$test"'$' "$TRIGGER" ; then +if ! grep -E -q '^'"$test"'$' "$TRIGGER" ; then echo "Skipped: test '$test' missing in $TRIGGER!" exit $KSELFTEST_SKIP_TEST fi @@ -80,11 +80,11 @@ dmesg | diff --changed-group-format='%>' --unchanged-group-format='' "$DMESG" - cat "$LOG" # Check for expected output -if egrep -qi "$expect" "$LOG" ; then +if grep -E -qi "$expect" "$LOG" ; then echo "$test: saw '$expect': ok" exit 0 else - if egrep -qi XFAIL: "$LOG" ; then + if grep -E -qi XFAIL: "$LOG" ; then echo "$test: saw 'XFAIL': [SKIP]" exit $KSELFTEST_SKIP_TEST else -- cgit v1.2.3 From 9d82ccda2bc5c148060543d249d54f8703741bb4 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (VMware)" Date: Fri, 8 May 2020 11:07:56 -0400 Subject: tools/bootconfig: Fix apply_xbc() to return zero on success The return of apply_xbc() returns the result of the last write() call, which is not what is expected. It should only return zero on success. Link: https://lore.kernel.org/r/20200508093059.GF9365@kadam Fixes: 8842604446d1 ("tools/bootconfig: Fix resource leak in apply_xbc()") Reported-by: Dan Carpenter Acked-by: Masami Hiramatsu Tested-by: Masami Hiramatsu Signed-off-by: Steven Rostedt (VMware) --- tools/bootconfig/main.c | 1 + 1 file changed, 1 insertion(+) (limited to 'tools') diff --git a/tools/bootconfig/main.c b/tools/bootconfig/main.c index 001076c51712..0efaf45f7367 100644 --- a/tools/bootconfig/main.c +++ b/tools/bootconfig/main.c @@ -337,6 +337,7 @@ int apply_xbc(const char *path, const char *xbc_path) pr_err("Failed to apply a boot config magic: %d\n", ret); goto out; } + ret = 0; out: close(fd); free(data); -- cgit v1.2.3 From 333291ce5055f2039afc907badaf5b66bc1adfdc Mon Sep 17 00:00:00 2001 From: Andrii Nakryiko Date: Tue, 12 May 2020 16:59:25 -0700 Subject: bpf: Fix bug in mmap() implementation for BPF array map mmap() subsystem allows user-space application to memory-map region with initial page offset. This wasn't taken into account in initial implementation of BPF array memory-mapping. This would result in wrong pages, not taking into account requested page shift, being memory-mmaped into user-space. This patch fixes this gap and adds a test for such scenario. Fixes: fc9702273e2e ("bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY") Signed-off-by: Andrii Nakryiko Signed-off-by: Alexei Starovoitov Acked-by: Yonghong Song Link: https://lore.kernel.org/bpf/20200512235925.3817805-1-andriin@fb.com --- kernel/bpf/arraymap.c | 7 ++++++- tools/testing/selftests/bpf/prog_tests/mmap.c | 8 ++++++++ 2 files changed, 14 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/kernel/bpf/arraymap.c b/kernel/bpf/arraymap.c index 95d77770353c..1d6120fd5ba6 100644 --- a/kernel/bpf/arraymap.c +++ b/kernel/bpf/arraymap.c @@ -486,7 +486,12 @@ static int array_map_mmap(struct bpf_map *map, struct vm_area_struct *vma) if (!(map->map_flags & BPF_F_MMAPABLE)) return -EINVAL; - return remap_vmalloc_range(vma, array_map_vmalloc_addr(array), pgoff); + if (vma->vm_pgoff * PAGE_SIZE + (vma->vm_end - vma->vm_start) > + PAGE_ALIGN((u64)array->map.max_entries * array->elem_size)) + return -EINVAL; + + return remap_vmalloc_range(vma, array_map_vmalloc_addr(array), + vma->vm_pgoff + pgoff); } const struct bpf_map_ops array_map_ops = { diff --git a/tools/testing/selftests/bpf/prog_tests/mmap.c b/tools/testing/selftests/bpf/prog_tests/mmap.c index 56d80adcf4bd..6b9dce431d41 100644 --- a/tools/testing/selftests/bpf/prog_tests/mmap.c +++ b/tools/testing/selftests/bpf/prog_tests/mmap.c @@ -217,6 +217,14 @@ void test_mmap(void) munmap(tmp2, 4 * page_size); + /* map all 4 pages, but with pg_off=1 page, should fail */ + tmp1 = mmap(NULL, 4 * page_size, PROT_READ, MAP_SHARED | MAP_FIXED, + data_map_fd, page_size /* initial page shift */); + if (CHECK(tmp1 != MAP_FAILED, "adv_mmap7", "unexpected success")) { + munmap(tmp1, 4 * page_size); + goto cleanup; + } + tmp1 = mmap(NULL, map_sz, PROT_READ, MAP_SHARED, data_map_fd, 0); if (CHECK(tmp1 == MAP_FAILED, "last_mmap", "failed %d\n", errno)) goto cleanup; -- cgit v1.2.3 From 516d8d497c017a6820019c76acbb7912a24ec57b Mon Sep 17 00:00:00 2001 From: Sumanth Korikkar Date: Wed, 13 May 2020 17:44:14 +0200 Subject: libbpf: Fix register naming in PT_REGS s390 macros Fix register naming in PT_REGS s390 macros Fixes: b8ebce86ffe6 ("libbpf: Provide CO-RE variants of PT_REGS macros") Signed-off-by: Sumanth Korikkar Signed-off-by: Alexei Starovoitov Reviewed-by: Julian Wiedmann Acked-by: Andrii Nakryiko Link: https://lore.kernel.org/bpf/20200513154414.29972-1-sumanthk@linux.ibm.com --- tools/lib/bpf/bpf_tracing.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/lib/bpf/bpf_tracing.h b/tools/lib/bpf/bpf_tracing.h index f3f3c3fb98cb..48a9c7c69ef1 100644 --- a/tools/lib/bpf/bpf_tracing.h +++ b/tools/lib/bpf/bpf_tracing.h @@ -148,11 +148,11 @@ struct pt_regs; #define PT_REGS_PARM3_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), gprs[4]) #define PT_REGS_PARM4_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), gprs[5]) #define PT_REGS_PARM5_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), gprs[6]) -#define PT_REGS_RET_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), grps[14]) +#define PT_REGS_RET_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), gprs[14]) #define PT_REGS_FP_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), gprs[11]) #define PT_REGS_RC_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), gprs[2]) #define PT_REGS_SP_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), gprs[15]) -#define PT_REGS_IP_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), pdw.addr) +#define PT_REGS_IP_CORE(x) BPF_CORE_READ((PT_REGS_S390 *)(x), psw.addr) #elif defined(bpf_target_arm) -- cgit v1.2.3 From 6d74f64b922b8394dccc52576659cb0dc0a1da7b Mon Sep 17 00:00:00 2001 From: Yonghong Song Date: Wed, 13 May 2020 22:32:07 -0700 Subject: selftests/bpf: Enforce returning 0 for fentry/fexit programs There are a few fentry/fexit programs returning non-0. The tests with these programs will break with the previous patch which enfoced return-0 rules. Fix them properly. Fixes: ac065870d928 ("selftests/bpf: Add BPF_PROG, BPF_KPROBE, and BPF_KRETPROBE macros") Signed-off-by: Yonghong Song Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Link: https://lore.kernel.org/bpf/20200514053207.1298479-1-yhs@fb.com --- tools/testing/selftests/bpf/progs/test_overhead.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/testing/selftests/bpf/progs/test_overhead.c b/tools/testing/selftests/bpf/progs/test_overhead.c index 56a50b25cd33..abb7344b531f 100644 --- a/tools/testing/selftests/bpf/progs/test_overhead.c +++ b/tools/testing/selftests/bpf/progs/test_overhead.c @@ -30,13 +30,13 @@ int prog3(struct bpf_raw_tracepoint_args *ctx) SEC("fentry/__set_task_comm") int BPF_PROG(prog4, struct task_struct *tsk, const char *buf, bool exec) { - return !tsk; + return 0; } SEC("fexit/__set_task_comm") int BPF_PROG(prog5, struct task_struct *tsk, const char *buf, bool exec) { - return !tsk; + return 0; } char _license[] SEC("license") = "GPL"; -- cgit v1.2.3 From 9a2dbb59ebd17a91f99a20b0286dfb6445981ecf Mon Sep 17 00:00:00 2001 From: Matthieu Baerts Date: Fri, 15 May 2020 17:54:41 +0200 Subject: selftests: mptcp: pm: rm the right tmp file "$err" is a variable pointing to a temp file. "$out" is not: only used as a local variable in "check()" and representing the output of a command line. Fixes: eedbc685321b (selftests: add PM netlink functional tests) Signed-off-by: Matthieu Baerts Signed-off-by: David S. Miller --- tools/testing/selftests/net/mptcp/pm_netlink.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/selftests/net/mptcp/pm_netlink.sh b/tools/testing/selftests/net/mptcp/pm_netlink.sh index 9172746b6cf0..15f4f46ca3a9 100755 --- a/tools/testing/selftests/net/mptcp/pm_netlink.sh +++ b/tools/testing/selftests/net/mptcp/pm_netlink.sh @@ -30,7 +30,7 @@ ret=0 cleanup() { - rm -f $out + rm -f $err ip netns del $ns1 } -- cgit v1.2.3