| 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 linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20200317220136.srrt6rpxdjhptu23@google.com>
Date: Tue, 17 Mar 2020 15:01:36 -0700
From: Fangrui Song <maskray@...gle.com>
To: Stanislav Fomichev <sdf@...ichev.me>
Cc: bpf <bpf@...r.kernel.org>, Networking <netdev@...r.kernel.org>,
Andrii Nakryiko <andrii.nakryiko@...il.com>,
Daniel Borkmann <daniel@...earbox.net>,
Nick Desaulniers <ndesaulniers@...gle.com>,
Nathan Chancellor <natechancellor@...il.com>,
clang-built-linux@...glegroups.com,
Stanislav Fomichev <sdf@...gle.com>,
"David S. Miller" <davem@...emloft.net>,
Alexei Starovoitov <ast@...nel.org>,
Kees Cook <keescook@...omium.org>
Subject: Re: [PATCH bpf-next v5] bpf: Support llvm-objcopy and llvm-objdump
for vmlinux BTF
On 2020-03-17, Stanislav Fomichev wrote:
>On 03/17, Fangrui Song wrote:
>> Simplify gen_btf logic to make it work with llvm-objcopy and llvm-objdump.
>> The existing 'file format' and 'architecture' parsing logic is brittle
>> and does not work with llvm-objcopy/llvm-objdump.
>Thanks, it all makes sense and looks much better/portable (too much
>dependence on binutils :-).
>I've left a bunch of questions/nits below.
>
>>
>> .BTF in .tmp_vmlinux.btf is non-SHF_ALLOC. Add the SHF_ALLOC flag and
>> rename .BTF to BTF so that C code can reference the section via linker
>> synthesized __start_BTF and __stop_BTF. This fixes a small problem that
>> previous .BTF had the SHF_WRITE flag. Additionally, `objcopy -I binary`
>> synthesized symbols _binary__btf_vmlinux_bin_start and
>> _binary__btf_vmlinux_bin_start (not used elsewhere) are replaced with
>> more common __start_BTF and __stop_BTF.
>>
>> Add 2>/dev/null because GNU objcopy (but not llvm-objcopy) warns
>> "empty loadable segment detected at vaddr=0xffffffff81000000, is this intentional?"
>>
>> We use a dd command to change the e_type field in the ELF header from
>> ET_EXEC to ET_REL so that lld will accept .btf.vmlinux.bin.o. Accepting
>> ET_EXEC as an input file is an extremely rare GNU ld feature that lld
>> does not intend to support, because this is error-prone.
>Please keep small changelog here, for example:
>
>v5:
>* rebased on top of bpfnext
Thanks for the tip. Add them at the bottom?
>Btw, I tried to pull and test it and failed:
>$ curl -LO https://lore.kernel.org/bpf/20200317211649.o4fzaxrzy6qxvz4f@google.com/raw
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
>100 9627 100 9627 0 0 47191 0 --:--:-- --:--:-- --:--:-- 47191
>$ git am raw
>warning: Patch sent with format=flowed; space at the end of lines might
>be lost.
>Applying: bpf: Support llvm-objcopy and llvm-objdump for vmlinux BTF
>error: corrupt patch at line 20
>Patch failed at 0001 bpf: Support llvm-objcopy and llvm-objdump for
>vmlinux BTF
>hint: Use 'git am --show-current-patch' to see the failed patch
>When you have resolved this problem, run "git am --continue".
>If you prefer to skip this patch, run "git am --skip" instead.
>To restore the original branch and stop patching, run "git am --abort".
>$ git describe
>v5.6-rc5-1621-g230021539e8c
>
>Are you sure it's on top of the bpf-next? Or am I doing something wrong?
It can be my fault when messing with various patches..
Will be more cautious.
>> Fixes: df786c9b9476 ("bpf: Force .BTF section start to zero when dumping from vmlinux")
>> Cc: Stanislav Fomichev <sdf@...gle.com>
>> Cc: Nick Desaulniers <ndesaulniers@...gle.com>
>> Reported-by: Nathan Chancellor <natechancellor@...il.com>
>> Link: https://github.com/ClangBuiltLinux/linux/issues/871
>> Signed-off-by: Fangrui Song <maskray@...gle.com>
>> ---
>> kernel/bpf/btf.c | 9 ++++-----
>> kernel/bpf/sysfs_btf.c | 11 +++++------
>> scripts/link-vmlinux.sh | 17 ++++++-----------
>> 3 files changed, 15 insertions(+), 22 deletions(-)
>>
>> diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c
>> index 50080add2ab9..6f397c4da05e 100644
>> --- a/kernel/bpf/btf.c
>> +++ b/kernel/bpf/btf.c
>> @@ -3477,8 +3477,8 @@ static struct btf *btf_parse(void __user *btf_data, u32 btf_data_size,
>> return ERR_PTR(err);
>> }
>> -extern char __weak _binary__btf_vmlinux_bin_start[];
>> -extern char __weak _binary__btf_vmlinux_bin_end[];
>> +extern char __weak __start_BTF[];
>> +extern char __weak __stop_BTF[];
>> extern struct btf *btf_vmlinux;
>> #define BPF_MAP_TYPE(_id, _ops)
>> @@ -3605,9 +3605,8 @@ struct btf *btf_parse_vmlinux(void)
>> }
>> env->btf = btf;
>> - btf->data = _binary__btf_vmlinux_bin_start;
>> - btf->data_size = _binary__btf_vmlinux_bin_end -
>> - _binary__btf_vmlinux_bin_start;
>> + btf->data = __start_BTF;
>> + btf->data_size = __stop_BTF - __start_BTF;
>> err = btf_parse_hdr(env);
>> if (err)
>> diff --git a/kernel/bpf/sysfs_btf.c b/kernel/bpf/sysfs_btf.c
>> index 7ae5dddd1fe6..3b495773de5a 100644
>> --- a/kernel/bpf/sysfs_btf.c
>> +++ b/kernel/bpf/sysfs_btf.c
>> @@ -9,15 +9,15 @@
>> #include <linux/sysfs.h>
>> /* See scripts/link-vmlinux.sh, gen_btf() func for details */
>> -extern char __weak _binary__btf_vmlinux_bin_start[];
>> -extern char __weak _binary__btf_vmlinux_bin_end[];
>> +extern char __weak __start_BTF[];
>> +extern char __weak __stop_BTF[];
>> static ssize_t
>> btf_vmlinux_read(struct file *file, struct kobject *kobj,
>> struct bin_attribute *bin_attr,
>> char *buf, loff_t off, size_t len)
>> {
>> - memcpy(buf, _binary__btf_vmlinux_bin_start + off, len);
>> + memcpy(buf, __start_BTF + off, len);
>> return len;
>> }
>> @@ -30,15 +30,14 @@ static struct kobject *btf_kobj;
>> static int __init btf_vmlinux_init(void)
>> {
>> - if (!_binary__btf_vmlinux_bin_start)
>> + if (!__start_BTF)
>> return 0;
>> btf_kobj = kobject_create_and_add("btf", kernel_kobj);
>> if (!btf_kobj)
>> return -ENOMEM;
>> - bin_attr_btf_vmlinux.size = _binary__btf_vmlinux_bin_end -
>> - _binary__btf_vmlinux_bin_start;
>> + bin_attr_btf_vmlinux.size = __stop_BTF - __start_BTF;
>> return sysfs_create_bin_file(btf_kobj, &bin_attr_btf_vmlinux);
>> }
>> diff --git a/scripts/link-vmlinux.sh b/scripts/link-vmlinux.sh
>> index ac569e197bfa..ae2048625f1e 100755
>> --- a/scripts/link-vmlinux.sh
>> +++ b/scripts/link-vmlinux.sh
>> @@ -133,17 +133,12 @@ gen_btf()
>> info "BTF" ${2}
>> LLVM_OBJCOPY=${OBJCOPY} ${PAHOLE} -J ${1}
>> - # dump .BTF section into raw binary file to link with final vmlinux
>> - bin_arch=$(LANG=C ${OBJDUMP} -f ${1} | grep architecture | \
>> - cut -d, -f1 | cut -d' ' -f2)
>> - bin_format=$(LANG=C ${OBJDUMP} -f ${1} | grep 'file format' | \
>> - awk '{print $4}')
>> - bin_file=.btf.vmlinux.bin
>We still have the following in gen_btf that you need to remove:
> local bin_arch
> local bin_format
> local bin_file
Thanks. Will delete them.
>> - ${OBJCOPY} --change-section-address .BTF=0 \
>> - --set-section-flags .BTF=alloc -O binary \
>> - --only-section=.BTF ${1} $bin_file
>> - ${OBJCOPY} -I binary -O ${bin_format} -B ${bin_arch} \
>> - --rename-section .data=.BTF $bin_file ${2}
>> + # Extract .BTF, add SHF_ALLOC, rename to BTF so that we can reference
>> + # it via linker synthesized __start_BTF and __stop_BTF. Change e_type
>> + # to ET_REL so that it can be used to link final vmlinux.
>> + ${OBJCOPY} --only-section=.BTF --set-section-flags .BTF=alloc,readonly \
>> + --rename-section .BTF=BTF ${1} ${2} 2>/dev/null && \
>> + printf '\1' | dd of=${2} conv=notrunc bs=1 seek=16 status=none
> We have 'set -e' so && can be omitted. Maybe it will be a bit
> more clear:
Agree.
> # Extract .BTF, add SHF_ALLOC, rename to BTF so that we can reference
> # it via linker synthesized __start_BTF and __stop_BTF.
> ${OBJCOPY} --only-section=.BTF --set-section-flags .BTF=alloc,readonly \
> --rename-section .BTF=BTF ${1} ${2} 2>/dev/null
>
> # Change e_type to ET_REL (0x01) so that it can be used to link final
> # vmlinux.
> printf '\1' | dd of=${2} conv=notrunc bs=1 seek=16 status=none
>
> ^^^ maybe also point out that this is required for llvm's ldd.
> I know you point it out in the commit message, but still nice
> to have as a comment here.
Will mention lld in v6.
>> }
>> # Create ${2} .o file with all symbols from the ${1} object file
>> --
>> 2.25.1.481.gfbce0eb801-goog
Powered by blists - more mailing lists