lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening PHC | |
Open Source and information security mailing list archives
| ||
|
Date: Wed, 15 Jan 2020 14:31:37 -0800 From: Andrii Nakryiko <andrii.nakryiko@...il.com> To: Toke Høiland-Jørgensen <toke@...hat.com> Cc: Alexei Starovoitov <ast@...nel.org>, Daniel Borkmann <daniel@...earbox.net>, Martin KaFai Lau <kafai@...com>, Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>, Andrii Nakryiko <andriin@...com>, Doug Ledford <dledford@...hat.com>, Jason Gunthorpe <jgg@...pe.ca>, "David S. Miller" <davem@...emloft.net>, Jakub Kicinski <jakub.kicinski@...ronome.com>, Jesper Dangaard Brouer <brouer@...hat.com>, John Fastabend <john.fastabend@...il.com>, Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...hat.com>, Arnaldo Carvalho de Melo <acme@...nel.org>, Mark Rutland <mark.rutland@....com>, Alexander Shishkin <alexander.shishkin@...ux.intel.com>, Jiri Olsa <jolsa@...hat.com>, Namhyung Kim <namhyung@...nel.org>, Shuah Khan <shuah@...nel.org>, Networking <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>, open list <linux-kernel@...r.kernel.org>, linux-rdma@...r.kernel.org, "open list:KERNEL SELFTEST FRAMEWORK" <linux-kselftest@...r.kernel.org>, clang-built-linux@...glegroups.com Subject: Re: [PATCH bpf-next v2 02/10] tools/bpf/runqslower: Fix override option for VMLINUX_BTF On Wed, Jan 15, 2020 at 2:06 PM Toke Høiland-Jørgensen <toke@...hat.com> wrote: > > Andrii Nakryiko <andrii.nakryiko@...il.com> writes: > > > On Wed, Jan 15, 2020 at 6:13 AM Toke Høiland-Jørgensen <toke@...hat.com> wrote: > >> > >> From: Toke Høiland-Jørgensen <toke@...hat.com> > >> > >> The runqslower tool refuses to build without a file to read vmlinux BTF > >> from. The build fails with an error message to override the location by > >> setting the VMLINUX_BTF variable if autodetection fails. However, the > >> Makefile doesn't actually work with that override - the error message is > >> still emitted. > > > > Do you have example command with VMLINUX_BTF override that didn't work > > (and what error message was emitted)? > > Before this patch: > > $ cd ~/build/linux/tools/bpf/runqslower > $ make > Makefile:18: *** "Can't detect kernel BTF, use VMLINUX_BTF to specify it explicitly". Stop. > > $ make VMLINUX_BTF=~/build/linux/vmlinux > Makefile:18: *** "Can't detect kernel BTF, use VMLINUX_BTF to specify it explicitly". Stop. Ok, so this is strange. Try make clean and run with V=1, it might help to debug this. This could happen if ~/build/linux/vmlinux doesn't exist, but I assume you double-checked that. It works for me just fine (Makefile won't do VMLINUX_BTF := assignment, if it's defined through make invocation, so your change should be a no-op in that regard): $ make clean $ make VMLINUX_BTF=~/linux-build/default/vmlinux V=1 ... .output/sbin/bpftool btf dump file ~/linux-build/default/vmlinux format c > .output/vmlinux.h ... Wonder what your output looks like? > > >> Fix this by only doing auto-detection if no override is set. And while > >> we're at it, also look for a vmlinux file in the current kernel build dir > >> if none if found on the running kernel. > >> > >> Fixes: 9c01546d26d2 ("tools/bpf: Add runqslower tool to tools/bpf") > >> Signed-off-by: Toke Høiland-Jørgensen <toke@...hat.com> > >> --- > >> tools/bpf/runqslower/Makefile | 16 ++++++++++------ > >> 1 file changed, 10 insertions(+), 6 deletions(-) > >> > >> diff --git a/tools/bpf/runqslower/Makefile b/tools/bpf/runqslower/Makefile > >> index cff2fbcd29a8..fb93ce2bf2fe 100644 > >> --- a/tools/bpf/runqslower/Makefile > >> +++ b/tools/bpf/runqslower/Makefile > >> @@ -10,12 +10,16 @@ CFLAGS := -g -Wall > >> > >> # Try to detect best kernel BTF source > >> KERNEL_REL := $(shell uname -r) > >> -ifneq ("$(wildcard /sys/kernel/btf/vmlinux)","") > >> -VMLINUX_BTF := /sys/kernel/btf/vmlinux > >> -else ifneq ("$(wildcard /boot/vmlinux-$(KERNEL_REL))","") > >> -VMLINUX_BTF := /boot/vmlinux-$(KERNEL_REL) > >> -else > >> -$(error "Can't detect kernel BTF, use VMLINUX_BTF to specify it explicitly") > >> +ifeq ("$(VMLINUX_BTF)","") > >> + ifneq ("$(wildcard /sys/kernel/btf/vmlinux)","") > >> + VMLINUX_BTF := /sys/kernel/btf/vmlinux > >> + else ifneq ("$(wildcard /boot/vmlinux-$(KERNEL_REL))","") > >> + VMLINUX_BTF := /boot/vmlinux-$(KERNEL_REL) > >> + else ifneq ("$(wildcard $(abspath ../../../vmlinux))","") > >> + VMLINUX_BTF := $(abspath ../../../vmlinux) > > > > I'm planning to mirror runqslower into libbpf Github repo and this > > ../../../vmlinux piece will be completely out of place in that > > context. Also it only will help when building kernel in-tree. So I'd > > rather not add this. > > Well building the kernel in-tree is something people sometimes want to do ;) > > Specifically, the selftests depend on this, so we should at least fix > those; but I guess it could work to just pass in VMLINUX_BTF as part of > the make -C from the selftests dir? I'll try that... Yes, it can be handled through VMLINUX_BTF override for selftests. As I said, this will be a self-contained example in libbpf's Github repo, so this "in kernel tree" assumption doesn't stand there. > > -Toke >
Powered by blists - more mailing lists