Message ID | 20220625153439.513559-1-tanzixuan.me@gmail.com (mailing list archive) |
---|---|
State | Not Applicable |
Delegated to: | BPF |
Headers | show |
Series | perf build: Suppress openssl v3 deprecation warnings in libcrypto feature test | expand |
Context | Check | Description |
---|---|---|
netdev/tree_selection | success | Not a local patch |
On Sat, Jun 25, 2022 at 11:34:38PM +0800, Zixuan Tan wrote: > With OpenSSL v3 installed, the libcrypto feature check fails as it use the > deprecated MD5_* API (and is compiled with -Werror). The error message is > as follows. > > $ make tools/perf > ``` > Makefile.config:778: No libcrypto.h found, disables jitted code injection, > please install openssl-devel or libssl-dev > > Auto-detecting system features: > ... dwarf: [ on ] > ... dwarf_getlocations: [ on ] > ... glibc: [ on ] > ... libbfd: [ on ] > ... libbfd-buildid: [ on ] > ... libcap: [ on ] > ... libelf: [ on ] > ... libnuma: [ on ] > ... numa_num_possible_cpus: [ on ] > ... libperl: [ on ] > ... libpython: [ on ] > ... libcrypto: [ OFF ] > ... libunwind: [ on ] > ... libdw-dwarf-unwind: [ on ] > ... zlib: [ on ] > ... lzma: [ on ] > ... get_cpuid: [ on ] > ... bpf: [ on ] > ... libaio: [ on ] > ... libzstd: [ on ] > ... disassembler-four-args: [ on ] > ``` > > This is very confusing because the suggested library (on my Ubuntu 20.04 > it is libssl-dev) is already installed. As the test only checks for the > presence of libcrypto, this commit suppresses the deprecation warning to > allow the test to pass. > > Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com> > --- > tools/build/feature/test-libcrypto.c | 6 ++++++ > 1 file changed, 6 insertions(+) > > diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c > index a98174e0569c..31afff093d0b 100644 > --- a/tools/build/feature/test-libcrypto.c > +++ b/tools/build/feature/test-libcrypto.c > @@ -2,6 +2,12 @@ > #include <openssl/sha.h> > #include <openssl/md5.h> > > +/* > + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the > + * feature test to fail silently. This is a workaround. > + */ then we use these deprecated MD5 calls in util/genelf.c if libcrypto is detected, so I wonder how come the rest of the compilation passed for you.. do you have CONFIG_JITDUMP disabled? thanks, jirka > +#pragma GCC diagnostic ignored "-Wdeprecated-declarations" > + > int main(void) > { > MD5_CTX context; > -- > 2.34.1 >
Em Sun, Jun 26, 2022 at 04:45:39PM +0200, Jiri Olsa escreveu: > On Sat, Jun 25, 2022 at 11:34:38PM +0800, Zixuan Tan wrote: > > With OpenSSL v3 installed, the libcrypto feature check fails as it use the > > deprecated MD5_* API (and is compiled with -Werror). The error message is > > as follows. > > > > $ make tools/perf > > ``` > > Makefile.config:778: No libcrypto.h found, disables jitted code injection, > > please install openssl-devel or libssl-dev > > > > Auto-detecting system features: > > ... dwarf: [ on ] > > ... dwarf_getlocations: [ on ] > > ... glibc: [ on ] > > ... libbfd: [ on ] > > ... libbfd-buildid: [ on ] > > ... libcap: [ on ] > > ... libelf: [ on ] > > ... libnuma: [ on ] > > ... numa_num_possible_cpus: [ on ] > > ... libperl: [ on ] > > ... libpython: [ on ] > > ... libcrypto: [ OFF ] > > ... libunwind: [ on ] > > ... libdw-dwarf-unwind: [ on ] > > ... zlib: [ on ] > > ... lzma: [ on ] > > ... get_cpuid: [ on ] > > ... bpf: [ on ] > > ... libaio: [ on ] > > ... libzstd: [ on ] > > ... disassembler-four-args: [ on ] > > ``` > > > > This is very confusing because the suggested library (on my Ubuntu 20.04 > > it is libssl-dev) is already installed. As the test only checks for the > > presence of libcrypto, this commit suppresses the deprecation warning to > > allow the test to pass. > > > > Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com> > > --- > > tools/build/feature/test-libcrypto.c | 6 ++++++ > > 1 file changed, 6 insertions(+) > > > > diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c > > index a98174e0569c..31afff093d0b 100644 > > --- a/tools/build/feature/test-libcrypto.c > > +++ b/tools/build/feature/test-libcrypto.c > > @@ -2,6 +2,12 @@ > > #include <openssl/sha.h> > > #include <openssl/md5.h> > > > > +/* > > + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the > > + * feature test to fail silently. This is a workaround. > > + */ > > then we use these deprecated MD5 calls in util/genelf.c if libcrypto is detected, > so I wonder how come the rest of the compilation passed for you.. do you have > CONFIG_JITDUMP disabled? So, here, on fedora 36: [acme@quaco perf-urgent]$ m make: Entering directory '/home/acme/git/perf-urgent/tools/perf' BUILD: Doing 'make -j8' parallel build HOSTCC /tmp/build/perf-urgent/fixdep.o HOSTLD /tmp/build/perf-urgent/fixdep-in.o LINK /tmp/build/perf-urgent/fixdep Warning: Kernel ABI header at 'tools/include/uapi/linux/kvm.h' differs from latest version at 'include/uapi/linux/kvm.h' diff -u tools/include/uapi/linux/kvm.h include/uapi/linux/kvm.h Warning: Kernel ABI header at 'tools/arch/x86/include/asm/disabled-features.h' differs from latest version at 'arch/x86/include/asm/disabled-features.h' diff -u tools/arch/x86/include/asm/disabled-features.h arch/x86/include/asm/disabled-features.h Warning: Kernel ABI header at 'tools/arch/arm64/include/uapi/asm/perf_regs.h' differs from latest version at 'arch/arm64/include/uapi/asm/perf_regs.h' diff -u tools/arch/arm64/include/uapi/asm/perf_regs.h arch/arm64/include/uapi/asm/perf_regs.h Warning: Kernel ABI header at 'tools/arch/x86/include/uapi/asm/svm.h' differs from latest version at 'arch/x86/include/uapi/asm/svm.h' diff -u tools/arch/x86/include/uapi/asm/svm.h arch/x86/include/uapi/asm/svm.h Warning: Kernel ABI header at 'tools/arch/arm64/include/uapi/asm/kvm.h' differs from latest version at 'arch/arm64/include/uapi/asm/kvm.h' diff -u tools/arch/arm64/include/uapi/asm/kvm.h arch/arm64/include/uapi/asm/kvm.h Warning: Kernel ABI header at 'tools/include/linux/coresight-pmu.h' differs from latest version at 'include/linux/coresight-pmu.h' diff -u tools/include/linux/coresight-pmu.h include/linux/coresight-pmu.h Makefile.config:778: No libcrypto.h found, disables jitted code injection, please install openssl-devel or libssl-dev Makefile.config:1108: No openjdk development package found, please install JDK package, e.g. openjdk-8-jdk, java-1.8.0-openjdk-devel Auto-detecting system features: ... dwarf: [ on ] ... dwarf_getlocations: [ on ] ... glibc: [ on ] ... libbfd: [ on ] ... libbfd-buildid: [ on ] ... libcap: [ on ] ... libelf: [ on ] ... libnuma: [ on ] ... numa_num_possible_cpus: [ on ] ... libperl: [ on ] ... libpython: [ on ] ... libcrypto: [ OFF ] <------------------------------------------------------- ... libunwind: [ on ] ... libdw-dwarf-unwind: [ on ] ... zlib: [ on ] ... lzma: [ on ] ... get_cpuid: [ on ] ... bpf: [ on ] ... libaio: [ on ] ... libzstd: [ on ] ... disassembler-four-args: [ on ] GEN /tmp/build/perf-urgent/common-cmds.h MKDIR /tmp/build/perf-urgent/fd/ CC /tmp/build/perf-urgent/exec-cmd.o CC /tmp/build/perf-urgent/fd/array.o But then: ⬢[acme@toolbox perf-urgent]$ perf -vv perf version 5.19.rc3.gfbec4d8dd3a7 dwarf: [ on ] # HAVE_DWARF_SUPPORT dwarf_getlocations: [ on ] # HAVE_DWARF_GETLOCATIONS_SUPPORT glibc: [ on ] # HAVE_GLIBC_SUPPORT syscall_table: [ on ] # HAVE_SYSCALL_TABLE_SUPPORT libbfd: [ on ] # HAVE_LIBBFD_SUPPORT debuginfod: [ on ] # HAVE_DEBUGINFOD_SUPPORT libelf: [ on ] # HAVE_LIBELF_SUPPORT libnuma: [ on ] # HAVE_LIBNUMA_SUPPORT numa_num_possible_cpus: [ on ] # HAVE_LIBNUMA_SUPPORT libperl: [ on ] # HAVE_LIBPERL_SUPPORT libpython: [ on ] # HAVE_LIBPYTHON_SUPPORT libslang: [ on ] # HAVE_SLANG_SUPPORT libcrypto: [ OFF ] # HAVE_LIBCRYPTO_SUPPORT libunwind: [ on ] # HAVE_LIBUNWIND_SUPPORT libdw-dwarf-unwind: [ on ] # HAVE_DWARF_SUPPORT zlib: [ on ] # HAVE_ZLIB_SUPPORT lzma: [ on ] # HAVE_LZMA_SUPPORT get_cpuid: [ on ] # HAVE_AUXTRACE_SUPPORT bpf: [ on ] # HAVE_LIBBPF_SUPPORT aio: [ on ] # HAVE_AIO_SUPPORT zstd: [ on ] # HAVE_ZSTD_SUPPORT libpfm4: [ OFF ] # HAVE_LIBPFM ⬢[acme@toolbox perf-urgent]$ But...: ⬢[acme@toolbox perf-urgent]$ ldd ~/bin/perf | grep ssl libssl.so.3 => /lib64/libssl.so.3 (0x00007f02dc87e000) ⬢[acme@toolbox perf-urgent]$ - Arnaldo
On Sun, Jun 26, 2022 at 10:45 PM Jiri Olsa <olsajiri@gmail.com> wrote: > > On Sat, Jun 25, 2022 at 11:34:38PM +0800, Zixuan Tan wrote: > > With OpenSSL v3 installed, the libcrypto feature check fails as it use the > > deprecated MD5_* API (and is compiled with -Werror). The error message is > > as follows. > > > > $ make tools/perf > > ``` > > Makefile.config:778: No libcrypto.h found, disables jitted code injection, > > please install openssl-devel or libssl-dev > > > > Auto-detecting system features: > > ... dwarf: [ on ] > > ... dwarf_getlocations: [ on ] > > ... glibc: [ on ] > > ... libbfd: [ on ] > > ... libbfd-buildid: [ on ] > > ... libcap: [ on ] > > ... libelf: [ on ] > > ... libnuma: [ on ] > > ... numa_num_possible_cpus: [ on ] > > ... libperl: [ on ] > > ... libpython: [ on ] > > ... libcrypto: [ OFF ] > > ... libunwind: [ on ] > > ... libdw-dwarf-unwind: [ on ] > > ... zlib: [ on ] > > ... lzma: [ on ] > > ... get_cpuid: [ on ] > > ... bpf: [ on ] > > ... libaio: [ on ] > > ... libzstd: [ on ] > > ... disassembler-four-args: [ on ] > > ``` > > > > This is very confusing because the suggested library (on my Ubuntu 20.04 > > it is libssl-dev) is already installed. As the test only checks for the > > presence of libcrypto, this commit suppresses the deprecation warning to > > allow the test to pass. > > > > Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com> > > --- > > tools/build/feature/test-libcrypto.c | 6 ++++++ > > 1 file changed, 6 insertions(+) > > > > diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c > > index a98174e0569c..31afff093d0b 100644 > > --- a/tools/build/feature/test-libcrypto.c > > +++ b/tools/build/feature/test-libcrypto.c > > @@ -2,6 +2,12 @@ > > #include <openssl/sha.h> > > #include <openssl/md5.h> > > > > +/* > > + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the > > + * feature test to fail silently. This is a workaround. > > + */ > > then we use these deprecated MD5 calls in util/genelf.c if libcrypto is detected, > so I wonder how come the rest of the compilation passed for you.. do you have > CONFIG_JITDUMP disabled? > > thanks, > jirka > No, CONFIG_JITDUMP is not disabled. I am using the default configuration. Yes, you are right. The rest of the compilation should fail, but it doesn't. I checked the verbose build commands. This seems to be the result of another inconsistency. If libcrypto is detected, the macro "HAVE_LIBCRYPTO_SUPPORT" will be defined, but in perf/util/genelf.c, "HAVE_LIBCRYPTO" without the "_SUPPORT" prefix is checked. This causes urandom always be used to create build id rather than MD5 and SHA1, no matter what the detection result is. In perf/Makefile.config, from line 776 ``` ifndef NO_LIBCRYPTO ifneq ($(feature-libcrypto), 1) msg := $(warning No libcrypto.h found, disables jitted code injection, please install openssl-devel or libssl-dev); NO_LIBCRYPTO := 1 else <-- if libcrypto feature detected CFLAGS += -DHAVE_LIBCRYPTO_SUPPORT <-- define this EXTLIBS += -lcrypto $(call detected,CONFIG_CRYPTO) endif endif ``` In perf/util/genelf.c, from line 33 ``` #ifdef HAVE_LIBCRYPTO <-- but check this, it's always false #define BUILD_ID_MD5 #undef BUILD_ID_SHA /* does not seem to work well when linked with Java */ #undef BUILD_ID_URANDOM /* different uuid for each run */ #ifdef BUILD_ID_SHA #include <openssl/sha.h> #endif #ifdef BUILD_ID_MD5 #include <openssl/md5.h> #endif #endif <-- this block will be skipped ``` Maybe we should fix this, to really make use of libcrypto if it is available? Links: This commit include the genelf.c: https://lore.kernel.org/all/1448874143-7269-3-git-send-email-eranian@google.com/T/#mb6d3e18bee4901b71a4d4ef4f406feaaf48346d9 This commit include the feature test: https://lore.kernel.org/all/1448874143-7269-3-git-send-email-eranian@google.com/T/#m12a2ababf8ad3e366d56d9efab870592e6ff60a5 Thanks, Zixuan > > +#pragma GCC diagnostic ignored "-Wdeprecated-declarations" > > + > > int main(void) > > { > > MD5_CTX context; > > -- > > 2.34.1 > >
On Mon, Jun 27, 2022 at 11:08:34AM +0800, 谭梓煊 wrote: > On Sun, Jun 26, 2022 at 10:45 PM Jiri Olsa <olsajiri@gmail.com> wrote: > > > > On Sat, Jun 25, 2022 at 11:34:38PM +0800, Zixuan Tan wrote: > > > With OpenSSL v3 installed, the libcrypto feature check fails as it use the > > > deprecated MD5_* API (and is compiled with -Werror). The error message is > > > as follows. > > > > > > $ make tools/perf > > > ``` > > > Makefile.config:778: No libcrypto.h found, disables jitted code injection, > > > please install openssl-devel or libssl-dev > > > > > > Auto-detecting system features: > > > ... dwarf: [ on ] > > > ... dwarf_getlocations: [ on ] > > > ... glibc: [ on ] > > > ... libbfd: [ on ] > > > ... libbfd-buildid: [ on ] > > > ... libcap: [ on ] > > > ... libelf: [ on ] > > > ... libnuma: [ on ] > > > ... numa_num_possible_cpus: [ on ] > > > ... libperl: [ on ] > > > ... libpython: [ on ] > > > ... libcrypto: [ OFF ] > > > ... libunwind: [ on ] > > > ... libdw-dwarf-unwind: [ on ] > > > ... zlib: [ on ] > > > ... lzma: [ on ] > > > ... get_cpuid: [ on ] > > > ... bpf: [ on ] > > > ... libaio: [ on ] > > > ... libzstd: [ on ] > > > ... disassembler-four-args: [ on ] > > > ``` > > > > > > This is very confusing because the suggested library (on my Ubuntu 20.04 > > > it is libssl-dev) is already installed. As the test only checks for the > > > presence of libcrypto, this commit suppresses the deprecation warning to > > > allow the test to pass. > > > > > > Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com> > > > --- > > > tools/build/feature/test-libcrypto.c | 6 ++++++ > > > 1 file changed, 6 insertions(+) > > > > > > diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c > > > index a98174e0569c..31afff093d0b 100644 > > > --- a/tools/build/feature/test-libcrypto.c > > > +++ b/tools/build/feature/test-libcrypto.c > > > @@ -2,6 +2,12 @@ > > > #include <openssl/sha.h> > > > #include <openssl/md5.h> > > > > > > +/* > > > + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the > > > + * feature test to fail silently. This is a workaround. > > > + */ > > > > then we use these deprecated MD5 calls in util/genelf.c if libcrypto is detected, > > so I wonder how come the rest of the compilation passed for you.. do you have > > CONFIG_JITDUMP disabled? > > > > thanks, > > jirka > > > No, CONFIG_JITDUMP is not disabled. I am using the default configuration. > > Yes, you are right. The rest of the compilation should fail, but it doesn't. > I checked the verbose build commands. This seems to be the result of another > inconsistency. > > If libcrypto is detected, the macro "HAVE_LIBCRYPTO_SUPPORT" will be > defined, but in perf/util/genelf.c, "HAVE_LIBCRYPTO" without the "_SUPPORT" > prefix is checked. This causes urandom always be used to create build id > rather than MD5 and SHA1, no matter what the detection result is. > > In perf/Makefile.config, from line 776 > ``` > ifndef NO_LIBCRYPTO > ifneq ($(feature-libcrypto), 1) > msg := $(warning No libcrypto.h found, disables jitted code injection, > please install openssl-devel or libssl-dev); > NO_LIBCRYPTO := 1 > else <-- if libcrypto feature detected > CFLAGS += -DHAVE_LIBCRYPTO_SUPPORT <-- define this > EXTLIBS += -lcrypto > $(call detected,CONFIG_CRYPTO) > endif > endif > ``` > > In perf/util/genelf.c, from line 33 > ``` > #ifdef HAVE_LIBCRYPTO <-- but check this, it's always false nice :) > > #define BUILD_ID_MD5 > #undef BUILD_ID_SHA /* does not seem to work well when linked with Java */ > #undef BUILD_ID_URANDOM /* different uuid for each run */ > > #ifdef BUILD_ID_SHA > #include <openssl/sha.h> > #endif > > #ifdef BUILD_ID_MD5 > #include <openssl/md5.h> > #endif > #endif <-- this block will be skipped > ``` > > Maybe we should fix this, to really make use of libcrypto if it is available? yea, I think that was the original idea, let's keep the variable with SUPPORT suffix and use the -Wdeprecated-declarations for genelf.c full fix would be to detect the new API and use it when it's available but.. given that the check was false at least since 2016, perhaps we could remove that code? ;-) Stephane? jirka > > Links: > This commit include the genelf.c: > https://lore.kernel.org/all/1448874143-7269-3-git-send-email-eranian@google.com/T/#mb6d3e18bee4901b71a4d4ef4f406feaaf48346d9 > This commit include the feature test: > https://lore.kernel.org/all/1448874143-7269-3-git-send-email-eranian@google.com/T/#m12a2ababf8ad3e366d56d9efab870592e6ff60a5 > > Thanks, > Zixuan > > > > +#pragma GCC diagnostic ignored "-Wdeprecated-declarations" > > > + > > > int main(void) > > > { > > > MD5_CTX context; > > > -- > > > 2.34.1 > > >
On Tue, Jun 28, 2022 at 08:16:06AM +0200, Jiri Olsa wrote: > On Mon, Jun 27, 2022 at 11:08:34AM +0800, 谭梓煊 wrote: > > On Sun, Jun 26, 2022 at 10:45 PM Jiri Olsa <olsajiri@gmail.com> wrote: > > > > > > On Sat, Jun 25, 2022 at 11:34:38PM +0800, Zixuan Tan wrote: > > > > With OpenSSL v3 installed, the libcrypto feature check fails as it use the > > > > deprecated MD5_* API (and is compiled with -Werror). The error message is > > > > as follows. > > > > > > > > $ make tools/perf > > > > ``` > > > > Makefile.config:778: No libcrypto.h found, disables jitted code injection, > > > > please install openssl-devel or libssl-dev > > > > > > > > Auto-detecting system features: > > > > ... dwarf: [ on ] > > > > ... dwarf_getlocations: [ on ] > > > > ... glibc: [ on ] > > > > ... libbfd: [ on ] > > > > ... libbfd-buildid: [ on ] > > > > ... libcap: [ on ] > > > > ... libelf: [ on ] > > > > ... libnuma: [ on ] > > > > ... numa_num_possible_cpus: [ on ] > > > > ... libperl: [ on ] > > > > ... libpython: [ on ] > > > > ... libcrypto: [ OFF ] > > > > ... libunwind: [ on ] > > > > ... libdw-dwarf-unwind: [ on ] > > > > ... zlib: [ on ] > > > > ... lzma: [ on ] > > > > ... get_cpuid: [ on ] > > > > ... bpf: [ on ] > > > > ... libaio: [ on ] > > > > ... libzstd: [ on ] > > > > ... disassembler-four-args: [ on ] > > > > ``` > > > > > > > > This is very confusing because the suggested library (on my Ubuntu 20.04 > > > > it is libssl-dev) is already installed. As the test only checks for the > > > > presence of libcrypto, this commit suppresses the deprecation warning to > > > > allow the test to pass. > > > > > > > > Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com> > > > > --- > > > > tools/build/feature/test-libcrypto.c | 6 ++++++ > > > > 1 file changed, 6 insertions(+) > > > > > > > > diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c > > > > index a98174e0569c..31afff093d0b 100644 > > > > --- a/tools/build/feature/test-libcrypto.c > > > > +++ b/tools/build/feature/test-libcrypto.c > > > > @@ -2,6 +2,12 @@ > > > > #include <openssl/sha.h> > > > > #include <openssl/md5.h> > > > > > > > > +/* > > > > + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the > > > > + * feature test to fail silently. This is a workaround. > > > > + */ > > > > > > then we use these deprecated MD5 calls in util/genelf.c if libcrypto is detected, > > > so I wonder how come the rest of the compilation passed for you.. do you have > > > CONFIG_JITDUMP disabled? > > > > > > thanks, > > > jirka > > > > > No, CONFIG_JITDUMP is not disabled. I am using the default configuration. > > > > Yes, you are right. The rest of the compilation should fail, but it doesn't. > > I checked the verbose build commands. This seems to be the result of another > > inconsistency. > > > > If libcrypto is detected, the macro "HAVE_LIBCRYPTO_SUPPORT" will be > > defined, but in perf/util/genelf.c, "HAVE_LIBCRYPTO" without the "_SUPPORT" > > prefix is checked. This causes urandom always be used to create build id > > rather than MD5 and SHA1, no matter what the detection result is. > > > > In perf/Makefile.config, from line 776 > > ``` > > ifndef NO_LIBCRYPTO > > ifneq ($(feature-libcrypto), 1) > > msg := $(warning No libcrypto.h found, disables jitted code injection, > > please install openssl-devel or libssl-dev); > > NO_LIBCRYPTO := 1 > > else <-- if libcrypto feature detected > > CFLAGS += -DHAVE_LIBCRYPTO_SUPPORT <-- define this > > EXTLIBS += -lcrypto > > $(call detected,CONFIG_CRYPTO) > > endif > > endif > > ``` > > > > In perf/util/genelf.c, from line 33 > > ``` > > #ifdef HAVE_LIBCRYPTO <-- but check this, it's always false > > nice :) > > > > > #define BUILD_ID_MD5 > > #undef BUILD_ID_SHA /* does not seem to work well when linked with Java */ > > #undef BUILD_ID_URANDOM /* different uuid for each run */ > > > > #ifdef BUILD_ID_SHA > > #include <openssl/sha.h> > > #endif > > > > #ifdef BUILD_ID_MD5 > > #include <openssl/md5.h> > > #endif > > #endif <-- this block will be skipped > > ``` > > > > Maybe we should fix this, to really make use of libcrypto if it is available? > > yea, I think that was the original idea, let's keep the variable with > SUPPORT suffix and use the -Wdeprecated-declarations for genelf.c > > full fix would be to detect the new API and use it when it's available but.. > given that the check was false at least since 2016, perhaps we could remove > that code? ;-) Stephane? ping jirka
Em Mon, Aug 01, 2022 at 03:24:18PM +0200, Jiri Olsa escreveu: > On Tue, Jun 28, 2022 at 08:16:06AM +0200, Jiri Olsa wrote: > > On Mon, Jun 27, 2022 at 11:08:34AM +0800, 谭梓煊 wrote: > > > On Sun, Jun 26, 2022 at 10:45 PM Jiri Olsa <olsajiri@gmail.com> wrote: > > > > > > > > On Sat, Jun 25, 2022 at 11:34:38PM +0800, Zixuan Tan wrote: > > > > > With OpenSSL v3 installed, the libcrypto feature check fails as it use the > > > > > deprecated MD5_* API (and is compiled with -Werror). The error message is > > > > > as follows. > > > > > > > > > > $ make tools/perf > > > > > ``` > > > > > Makefile.config:778: No libcrypto.h found, disables jitted code injection, > > > > > please install openssl-devel or libssl-dev > > > > > > > > > > Auto-detecting system features: > > > > > ... dwarf: [ on ] > > > > > ... dwarf_getlocations: [ on ] > > > > > ... glibc: [ on ] > > > > > ... libbfd: [ on ] > > > > > ... libbfd-buildid: [ on ] > > > > > ... libcap: [ on ] > > > > > ... libelf: [ on ] > > > > > ... libnuma: [ on ] > > > > > ... numa_num_possible_cpus: [ on ] > > > > > ... libperl: [ on ] > > > > > ... libpython: [ on ] > > > > > ... libcrypto: [ OFF ] > > > > > ... libunwind: [ on ] > > > > > ... libdw-dwarf-unwind: [ on ] > > > > > ... zlib: [ on ] > > > > > ... lzma: [ on ] > > > > > ... get_cpuid: [ on ] > > > > > ... bpf: [ on ] > > > > > ... libaio: [ on ] > > > > > ... libzstd: [ on ] > > > > > ... disassembler-four-args: [ on ] > > > > > ``` > > > > > > > > > > This is very confusing because the suggested library (on my Ubuntu 20.04 > > > > > it is libssl-dev) is already installed. As the test only checks for the > > > > > presence of libcrypto, this commit suppresses the deprecation warning to > > > > > allow the test to pass. > > > > > > > > > > Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com> > > > > > --- > > > > > tools/build/feature/test-libcrypto.c | 6 ++++++ > > > > > 1 file changed, 6 insertions(+) > > > > > > > > > > diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c > > > > > index a98174e0569c..31afff093d0b 100644 > > > > > --- a/tools/build/feature/test-libcrypto.c > > > > > +++ b/tools/build/feature/test-libcrypto.c > > > > > @@ -2,6 +2,12 @@ > > > > > #include <openssl/sha.h> > > > > > #include <openssl/md5.h> > > > > > > > > > > +/* > > > > > + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the > > > > > + * feature test to fail silently. This is a workaround. > > > > > + */ > > > > > > > > then we use these deprecated MD5 calls in util/genelf.c if libcrypto is detected, > > > > so I wonder how come the rest of the compilation passed for you.. do you have > > > > CONFIG_JITDUMP disabled? > > > > > > > > thanks, > > > > jirka > > > > > > > No, CONFIG_JITDUMP is not disabled. I am using the default configuration. > > > > > > Yes, you are right. The rest of the compilation should fail, but it doesn't. > > > I checked the verbose build commands. This seems to be the result of another > > > inconsistency. > > > > > > If libcrypto is detected, the macro "HAVE_LIBCRYPTO_SUPPORT" will be > > > defined, but in perf/util/genelf.c, "HAVE_LIBCRYPTO" without the "_SUPPORT" > > > prefix is checked. This causes urandom always be used to create build id > > > rather than MD5 and SHA1, no matter what the detection result is. > > > > > > In perf/Makefile.config, from line 776 > > > ``` > > > ifndef NO_LIBCRYPTO > > > ifneq ($(feature-libcrypto), 1) > > > msg := $(warning No libcrypto.h found, disables jitted code injection, > > > please install openssl-devel or libssl-dev); > > > NO_LIBCRYPTO := 1 > > > else <-- if libcrypto feature detected > > > CFLAGS += -DHAVE_LIBCRYPTO_SUPPORT <-- define this > > > EXTLIBS += -lcrypto > > > $(call detected,CONFIG_CRYPTO) > > > endif > > > endif > > > ``` > > > > > > In perf/util/genelf.c, from line 33 > > > ``` > > > #ifdef HAVE_LIBCRYPTO <-- but check this, it's always false > > > > nice :) > > > > > > > > #define BUILD_ID_MD5 > > > #undef BUILD_ID_SHA /* does not seem to work well when linked with Java */ > > > #undef BUILD_ID_URANDOM /* different uuid for each run */ > > > > > > #ifdef BUILD_ID_SHA > > > #include <openssl/sha.h> > > > #endif > > > > > > #ifdef BUILD_ID_MD5 > > > #include <openssl/md5.h> > > > #endif > > > #endif <-- this block will be skipped > > > ``` > > > > > > Maybe we should fix this, to really make use of libcrypto if it is available? > > > > yea, I think that was the original idea, let's keep the variable with > > SUPPORT suffix and use the -Wdeprecated-declarations for genelf.c > > > > full fix would be to detect the new API and use it when it's available but.. > > given that the check was false at least since 2016, perhaps we could remove > > that code? ;-) Stephane? > > ping So, we should start with 谭梓煊 patch, then fix that ifdef and go on from there? - Arnaldo
Em Tue, Aug 02, 2022 at 03:09:05PM -0300, Arnaldo Carvalho de Melo escreveu: > Em Mon, Aug 01, 2022 at 03:24:18PM +0200, Jiri Olsa escreveu: > > On Tue, Jun 28, 2022 at 08:16:06AM +0200, Jiri Olsa wrote: > > > On Mon, Jun 27, 2022 at 11:08:34AM +0800, 谭梓煊 wrote: > > > > On Sun, Jun 26, 2022 at 10:45 PM Jiri Olsa <olsajiri@gmail.com> wrote: > > > > > > > > > > On Sat, Jun 25, 2022 at 11:34:38PM +0800, Zixuan Tan wrote: > > > > > > With OpenSSL v3 installed, the libcrypto feature check fails as it use the > > > > > > deprecated MD5_* API (and is compiled with -Werror). The error message is > > > > > > as follows. > > > > > > > > > > > > $ make tools/perf > > > > > > ``` > > > > > > Makefile.config:778: No libcrypto.h found, disables jitted code injection, > > > > > > please install openssl-devel or libssl-dev > > > > > > > > > > > > Auto-detecting system features: > > > > > > ... dwarf: [ on ] > > > > > > ... dwarf_getlocations: [ on ] > > > > > > ... glibc: [ on ] > > > > > > ... libbfd: [ on ] > > > > > > ... libbfd-buildid: [ on ] > > > > > > ... libcap: [ on ] > > > > > > ... libelf: [ on ] > > > > > > ... libnuma: [ on ] > > > > > > ... numa_num_possible_cpus: [ on ] > > > > > > ... libperl: [ on ] > > > > > > ... libpython: [ on ] > > > > > > ... libcrypto: [ OFF ] > > > > > > ... libunwind: [ on ] > > > > > > ... libdw-dwarf-unwind: [ on ] > > > > > > ... zlib: [ on ] > > > > > > ... lzma: [ on ] > > > > > > ... get_cpuid: [ on ] > > > > > > ... bpf: [ on ] > > > > > > ... libaio: [ on ] > > > > > > ... libzstd: [ on ] > > > > > > ... disassembler-four-args: [ on ] > > > > > > ``` > > > > > > > > > > > > This is very confusing because the suggested library (on my Ubuntu 20.04 > > > > > > it is libssl-dev) is already installed. As the test only checks for the > > > > > > presence of libcrypto, this commit suppresses the deprecation warning to > > > > > > allow the test to pass. > > > > > > > > > > > > Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com> > > > > > > --- > > > > > > tools/build/feature/test-libcrypto.c | 6 ++++++ > > > > > > 1 file changed, 6 insertions(+) > > > > > > > > > > > > diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c > > > > > > index a98174e0569c..31afff093d0b 100644 > > > > > > --- a/tools/build/feature/test-libcrypto.c > > > > > > +++ b/tools/build/feature/test-libcrypto.c > > > > > > @@ -2,6 +2,12 @@ > > > > > > #include <openssl/sha.h> > > > > > > #include <openssl/md5.h> > > > > > > > > > > > > +/* > > > > > > + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the > > > > > > + * feature test to fail silently. This is a workaround. > > > > > > + */ > > > > > > > > > > then we use these deprecated MD5 calls in util/genelf.c if libcrypto is detected, > > > > > so I wonder how come the rest of the compilation passed for you.. do you have > > > > > CONFIG_JITDUMP disabled? > > > > > > > > > No, CONFIG_JITDUMP is not disabled. I am using the default configuration. > > > > > > > > Yes, you are right. The rest of the compilation should fail, but it doesn't. > > > > I checked the verbose build commands. This seems to be the result of another > > > > inconsistency. > > > > > > > > If libcrypto is detected, the macro "HAVE_LIBCRYPTO_SUPPORT" will be > > > > defined, but in perf/util/genelf.c, "HAVE_LIBCRYPTO" without the "_SUPPORT" > > > > prefix is checked. This causes urandom always be used to create build id > > > > rather than MD5 and SHA1, no matter what the detection result is. > > > > > > > > In perf/Makefile.config, from line 776 > > > > ``` > > > > ifndef NO_LIBCRYPTO > > > > ifneq ($(feature-libcrypto), 1) > > > > msg := $(warning No libcrypto.h found, disables jitted code injection, > > > > please install openssl-devel or libssl-dev); > > > > NO_LIBCRYPTO := 1 > > > > else <-- if libcrypto feature detected > > > > CFLAGS += -DHAVE_LIBCRYPTO_SUPPORT <-- define this > > > > EXTLIBS += -lcrypto > > > > $(call detected,CONFIG_CRYPTO) > > > > endif > > > > endif > > > > ``` > > > > > > > > In perf/util/genelf.c, from line 33 > > > > ``` > > > > #ifdef HAVE_LIBCRYPTO <-- but check this, it's always false > > > > > > nice :) > > > > > > > > > > > #define BUILD_ID_MD5 > > > > #undef BUILD_ID_SHA /* does not seem to work well when linked with Java */ > > > > #undef BUILD_ID_URANDOM /* different uuid for each run */ > > > > > > > > #ifdef BUILD_ID_SHA > > > > #include <openssl/sha.h> > > > > #endif > > > > > > > > #ifdef BUILD_ID_MD5 > > > > #include <openssl/md5.h> > > > > #endif > > > > #endif <-- this block will be skipped > > > > ``` > > > > > > > > Maybe we should fix this, to really make use of libcrypto if it is available? > > > > > > yea, I think that was the original idea, let's keep the variable with > > > SUPPORT suffix and use the -Wdeprecated-declarations for genelf.c > > > > > > full fix would be to detect the new API and use it when it's available but.. > > > given that the check was false at least since 2016, perhaps we could remove > > > that code? ;-) Stephane? > > > > ping > > So, we should start with 谭梓煊 patch, then fix that ifdef and go on > from there? I.e. with this: diff --git a/tools/perf/util/genelf.c b/tools/perf/util/genelf.c index aed49806a09bab8f..953338b9e887e26f 100644 --- a/tools/perf/util/genelf.c +++ b/tools/perf/util/genelf.c @@ -30,7 +30,11 @@ #define BUILD_ID_URANDOM /* different uuid for each run */ -#ifdef HAVE_LIBCRYPTO +// FIXME, remove this and fix the deprecation warnings before its removed and +// We'll break for good here... +#pragma GCC diagnostic ignored "-Wdeprecated-declarations" + +#ifdef HAVE_LIBCRYPTO_SUPPORT #define BUILD_ID_MD5 #undef BUILD_ID_SHA /* does not seem to work well when linked with Java */
On Wed, Aug 3, 2022 at 2:13 AM Arnaldo Carvalho de Melo <arnaldo.melo@gmail.com> wrote: > > Em Tue, Aug 02, 2022 at 03:09:05PM -0300, Arnaldo Carvalho de Melo escreveu: > > > > So, we should start with 谭梓煊 patch, then fix that ifdef and go on > > from there? > > I.e. with this: > > > diff --git a/tools/perf/util/genelf.c b/tools/perf/util/genelf.c > index aed49806a09bab8f..953338b9e887e26f 100644 > --- a/tools/perf/util/genelf.c > +++ b/tools/perf/util/genelf.c > @@ -30,7 +30,11 @@ > > #define BUILD_ID_URANDOM /* different uuid for each run */ > > -#ifdef HAVE_LIBCRYPTO > +// FIXME, remove this and fix the deprecation warnings before its removed and > +// We'll break for good here... > +#pragma GCC diagnostic ignored "-Wdeprecated-declarations" > + > +#ifdef HAVE_LIBCRYPTO_SUPPORT > > #define BUILD_ID_MD5 > #undef BUILD_ID_SHA /* does not seem to work well when linked with Java */ yea, i think that's ok, thank you Zixuan
On Tue, Aug 02, 2022 at 03:09:05PM -0300, Arnaldo Carvalho de Melo wrote: > Em Mon, Aug 01, 2022 at 03:24:18PM +0200, Jiri Olsa escreveu: > > On Tue, Jun 28, 2022 at 08:16:06AM +0200, Jiri Olsa wrote: > > > On Mon, Jun 27, 2022 at 11:08:34AM +0800, 谭梓煊 wrote: > > > > On Sun, Jun 26, 2022 at 10:45 PM Jiri Olsa <olsajiri@gmail.com> wrote: > > > > > > > > > > On Sat, Jun 25, 2022 at 11:34:38PM +0800, Zixuan Tan wrote: > > > > > > With OpenSSL v3 installed, the libcrypto feature check fails as it use the > > > > > > deprecated MD5_* API (and is compiled with -Werror). The error message is > > > > > > as follows. > > > > > > > > > > > > $ make tools/perf > > > > > > ``` > > > > > > Makefile.config:778: No libcrypto.h found, disables jitted code injection, > > > > > > please install openssl-devel or libssl-dev > > > > > > > > > > > > Auto-detecting system features: > > > > > > ... dwarf: [ on ] > > > > > > ... dwarf_getlocations: [ on ] > > > > > > ... glibc: [ on ] > > > > > > ... libbfd: [ on ] > > > > > > ... libbfd-buildid: [ on ] > > > > > > ... libcap: [ on ] > > > > > > ... libelf: [ on ] > > > > > > ... libnuma: [ on ] > > > > > > ... numa_num_possible_cpus: [ on ] > > > > > > ... libperl: [ on ] > > > > > > ... libpython: [ on ] > > > > > > ... libcrypto: [ OFF ] > > > > > > ... libunwind: [ on ] > > > > > > ... libdw-dwarf-unwind: [ on ] > > > > > > ... zlib: [ on ] > > > > > > ... lzma: [ on ] > > > > > > ... get_cpuid: [ on ] > > > > > > ... bpf: [ on ] > > > > > > ... libaio: [ on ] > > > > > > ... libzstd: [ on ] > > > > > > ... disassembler-four-args: [ on ] > > > > > > ``` > > > > > > > > > > > > This is very confusing because the suggested library (on my Ubuntu 20.04 > > > > > > it is libssl-dev) is already installed. As the test only checks for the > > > > > > presence of libcrypto, this commit suppresses the deprecation warning to > > > > > > allow the test to pass. > > > > > > > > > > > > Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com> > > > > > > --- > > > > > > tools/build/feature/test-libcrypto.c | 6 ++++++ > > > > > > 1 file changed, 6 insertions(+) > > > > > > > > > > > > diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c > > > > > > index a98174e0569c..31afff093d0b 100644 > > > > > > --- a/tools/build/feature/test-libcrypto.c > > > > > > +++ b/tools/build/feature/test-libcrypto.c > > > > > > @@ -2,6 +2,12 @@ > > > > > > #include <openssl/sha.h> > > > > > > #include <openssl/md5.h> > > > > > > > > > > > > +/* > > > > > > + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the > > > > > > + * feature test to fail silently. This is a workaround. > > > > > > + */ > > > > > > > > > > then we use these deprecated MD5 calls in util/genelf.c if libcrypto is detected, > > > > > so I wonder how come the rest of the compilation passed for you.. do you have > > > > > CONFIG_JITDUMP disabled? > > > > > > > > > > thanks, > > > > > jirka > > > > > > > > > No, CONFIG_JITDUMP is not disabled. I am using the default configuration. > > > > > > > > Yes, you are right. The rest of the compilation should fail, but it doesn't. > > > > I checked the verbose build commands. This seems to be the result of another > > > > inconsistency. > > > > > > > > If libcrypto is detected, the macro "HAVE_LIBCRYPTO_SUPPORT" will be > > > > defined, but in perf/util/genelf.c, "HAVE_LIBCRYPTO" without the "_SUPPORT" > > > > prefix is checked. This causes urandom always be used to create build id > > > > rather than MD5 and SHA1, no matter what the detection result is. > > > > > > > > In perf/Makefile.config, from line 776 > > > > ``` > > > > ifndef NO_LIBCRYPTO > > > > ifneq ($(feature-libcrypto), 1) > > > > msg := $(warning No libcrypto.h found, disables jitted code injection, > > > > please install openssl-devel or libssl-dev); > > > > NO_LIBCRYPTO := 1 > > > > else <-- if libcrypto feature detected > > > > CFLAGS += -DHAVE_LIBCRYPTO_SUPPORT <-- define this > > > > EXTLIBS += -lcrypto > > > > $(call detected,CONFIG_CRYPTO) > > > > endif > > > > endif > > > > ``` > > > > > > > > In perf/util/genelf.c, from line 33 > > > > ``` > > > > #ifdef HAVE_LIBCRYPTO <-- but check this, it's always false > > > > > > nice :) > > > > > > > > > > > #define BUILD_ID_MD5 > > > > #undef BUILD_ID_SHA /* does not seem to work well when linked with Java */ > > > > #undef BUILD_ID_URANDOM /* different uuid for each run */ > > > > > > > > #ifdef BUILD_ID_SHA > > > > #include <openssl/sha.h> > > > > #endif > > > > > > > > #ifdef BUILD_ID_MD5 > > > > #include <openssl/md5.h> > > > > #endif > > > > #endif <-- this block will be skipped > > > > ``` > > > > > > > > Maybe we should fix this, to really make use of libcrypto if it is available? > > > > > > yea, I think that was the original idea, let's keep the variable with > > > SUPPORT suffix and use the -Wdeprecated-declarations for genelf.c > > > > > > full fix would be to detect the new API and use it when it's available but.. > > > given that the check was false at least since 2016, perhaps we could remove > > > that code? ;-) Stephane? > > > > ping > > So, we should start with 谭梓煊 patch, then fix that ifdef and go on > from there? yes, I thought we could remove that, but there's no reply from Stephane so let's fix that jirka
Em Wed, Aug 03, 2022 at 09:32:48AM +0200, Jiri Olsa escreveu: > On Tue, Aug 02, 2022 at 03:09:05PM -0300, Arnaldo Carvalho de Melo wrote: > > Em Mon, Aug 01, 2022 at 03:24:18PM +0200, Jiri Olsa escreveu: > > > On Tue, Jun 28, 2022 at 08:16:06AM +0200, Jiri Olsa wrote: > > > > On Mon, Jun 27, 2022 at 11:08:34AM +0800, 谭梓煊 wrote: > > > > > #ifdef HAVE_LIBCRYPTO <-- but check this, it's always false > > > > nice :) > > > > > #define BUILD_ID_MD5 > > > > > #undef BUILD_ID_SHA /* does not seem to work well when linked with Java */ > > > > > #undef BUILD_ID_URANDOM /* different uuid for each run */ > > > > > #ifdef BUILD_ID_SHA > > > > > #include <openssl/sha.h> > > > > > #endif > > > > > #ifdef BUILD_ID_MD5 > > > > > #include <openssl/md5.h> > > > > > #endif > > > > > #endif <-- this block will be skipped > > > > > ``` > > > > > Maybe we should fix this, to really make use of libcrypto if it is available? > > > > yea, I think that was the original idea, let's keep the variable with > > > > SUPPORT suffix and use the -Wdeprecated-declarations for genelf.c > > > > > > > > full fix would be to detect the new API and use it when it's available but.. > > > > given that the check was false at least since 2016, perhaps we could remove > > > > that code? ;-) Stephane? > > > > > > ping > > > > So, we should start with 谭梓煊 patch, then fix that ifdef and go on > > from there? > > yes, I thought we could remove that, but there's no reply from > Stephane so let's fix that Yeah, I did it and it seems to build, so lets ship it :-) - Arnaldo
diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c index a98174e0569c..31afff093d0b 100644 --- a/tools/build/feature/test-libcrypto.c +++ b/tools/build/feature/test-libcrypto.c @@ -2,6 +2,12 @@ #include <openssl/sha.h> #include <openssl/md5.h> +/* + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the + * feature test to fail silently. This is a workaround. + */ +#pragma GCC diagnostic ignored "-Wdeprecated-declarations" + int main(void) { MD5_CTX context;
With OpenSSL v3 installed, the libcrypto feature check fails as it use the deprecated MD5_* API (and is compiled with -Werror). The error message is as follows. $ make tools/perf ``` Makefile.config:778: No libcrypto.h found, disables jitted code injection, please install openssl-devel or libssl-dev Auto-detecting system features: ... dwarf: [ on ] ... dwarf_getlocations: [ on ] ... glibc: [ on ] ... libbfd: [ on ] ... libbfd-buildid: [ on ] ... libcap: [ on ] ... libelf: [ on ] ... libnuma: [ on ] ... numa_num_possible_cpus: [ on ] ... libperl: [ on ] ... libpython: [ on ] ... libcrypto: [ OFF ] ... libunwind: [ on ] ... libdw-dwarf-unwind: [ on ] ... zlib: [ on ] ... lzma: [ on ] ... get_cpuid: [ on ] ... bpf: [ on ] ... libaio: [ on ] ... libzstd: [ on ] ... disassembler-four-args: [ on ] ``` This is very confusing because the suggested library (on my Ubuntu 20.04 it is libssl-dev) is already installed. As the test only checks for the presence of libcrypto, this commit suppresses the deprecation warning to allow the test to pass. Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com> --- tools/build/feature/test-libcrypto.c | 6 ++++++ 1 file changed, 6 insertions(+)