netdev - Re: [PATCH v5 bpf-next 1/2] bpf: introduce helper bpf_find

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Mon, 8 Nov 2021 23:08:16 +0000
From:   Song Liu <songliubraving@...com>
To:     Eric Dumazet <eric.dumazet@...il.com>
CC:     bpf <bpf@...r.kernel.org>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "ast@...nel.org" <ast@...nel.org>,
        "daniel@...earbox.net" <daniel@...earbox.net>,
        "andrii@...nel.org" <andrii@...nel.org>,
        Kernel Team <Kernel-team@...com>,
        "kpsingh@...nel.org" <kpsingh@...nel.org>,
        "hengqi.chen@...il.com" <hengqi.chen@...il.com>,
        Yonghong Song <yhs@...com>
Subject: Re: [PATCH v5 bpf-next 1/2] bpf: introduce helper bpf_find_vma



> On Nov 8, 2021, at 2:56 PM, Eric Dumazet <eric.dumazet@...il.com> wrote:
> 
> 
> 
> On 11/8/21 2:43 PM, Eric Dumazet wrote:
>> 
>> 
>> On 11/8/21 2:27 PM, Eric Dumazet wrote:
>>> 
>>> 
>>> On 11/8/21 1:59 PM, Song Liu wrote:
>>>> 
>>>> 
>>>>> On Nov 8, 2021, at 10:36 AM, Eric Dumazet <eric.dumazet@...il.com> wrote:
>>>>> 
>>>>> 
>>>>> 
>>>>> On 11/5/21 4:23 PM, Song Liu wrote:
>>>>>> In some profiler use cases, it is necessary to map an address to the
>>>>>> backing file, e.g., a shared library. bpf_find_vma helper provides a
>>>>>> flexible way to achieve this. bpf_find_vma maps an address of a task to
>>>>>> the vma (vm_area_struct) for this address, and feed the vma to an callback
>>>>>> BPF function. The callback function is necessary here, as we need to
>>>>>> ensure mmap_sem is unlocked.
>>>>>> 
>>>>>> It is necessary to lock mmap_sem for find_vma. To lock and unlock mmap_sem
>>>>>> safely when irqs are disable, we use the same mechanism as stackmap with
>>>>>> build_id. Specifically, when irqs are disabled, the unlocked is postponed
>>>>>> in an irq_work. Refactor stackmap.c so that the irq_work is shared among
>>>>>> bpf_find_vma and stackmap helpers.
>>>>>> 
>>>>>> Acked-by: Yonghong Song <yhs@...com>
>>>>>> Tested-by: Hengqi Chen <hengqi.chen@...il.com>
>>>>>> Signed-off-by: Song Liu <songliubraving@...com>
>>>>>> ---
>>>>> 
>>>>> ...
>>>>> 
>>>>>> diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c
>>>>>> index dbc3ad07e21b6..cdb0fba656006 100644
>>>>>> --- a/kernel/bpf/btf.c
>>>>>> +++ b/kernel/bpf/btf.c
>>>>>> @@ -6342,7 +6342,10 @@ const struct bpf_func_proto bpf_btf_find_by_name_kind_proto = {
>>>>>> 	.arg4_type	= ARG_ANYTHING,
>>>>>> };
>>>>>> 
>>>>>> -BTF_ID_LIST_GLOBAL_SINGLE(btf_task_struct_ids, struct, task_struct)
>>>>>> +BTF_ID_LIST_GLOBAL(btf_task_struct_ids)
>>>>>> +BTF_ID(struct, task_struct)
>>>>>> +BTF_ID(struct, file)
>>>>>> +BTF_ID(struct, vm_area_struct)
>>>>> 
>>>>> $ nm -v vmlinux |grep -A3 btf_task_struct_ids
>>>>> ffffffff82adfd9c R btf_task_struct_ids
>>>>> ffffffff82adfda0 r __BTF_ID__struct__file__715
>>>>> ffffffff82adfda4 r __BTF_ID__struct__vm_area_struct__716
>>>>> ffffffff82adfda8 r bpf_skb_output_btf_ids
>>>>> 
>>>>> KASAN thinks btf_task_struct_ids has 4 bytes only.
>>>> 
>>>> I have KASAN enabled, but couldn't repro this issue. I think
>>>> btf_task_struct_ids looks correct:
>>>> 
>>>> nm -v vmlinux | grep -A3 -B1 btf_task_struct_ids
>>>> ffffffff83cf8260 r __BTF_ID__struct__task_struct__1026
>>>> ffffffff83cf8260 R btf_task_struct_ids
>>>> ffffffff83cf8264 r __BTF_ID__struct__file__1027
>>>> ffffffff83cf8268 r __BTF_ID__struct__vm_area_struct__1028
>>>> ffffffff83cf826c r bpf_skb_output_btf_ids
>>>> 
>>>> Did I miss something?
>>>> 
>>>> Thanks,
>>>> Song
>>>> 
>>> 
>>> I will release the syzbot bug, so that you can use its .config
>>> 
>>> Basically, we have
>>> 
>>> u32 btf_task_struct_ids[1];
>> 
>> That is, if  
>> 
>> # CONFIG_DEBUG_INFO_BTF is not set
>> 
> 
> This is how btf_sock_ids gets defined :
> 
> #ifdef CONFIG_DEBUG_INFO_BTF
> BTF_ID_LIST_GLOBAL(btf_sock_ids)
> #define BTF_SOCK_TYPE(name, type) BTF_ID(struct, type)
> BTF_SOCK_TYPE_xxx
> #undef BTF_SOCK_TYPE
> #else
> u32 btf_sock_ids[MAX_BTF_SOCK_TYPE];
> #endif
> 
> 
> Perhaps do the same for btf_task_struct_ids ?

Yeah, I was testing something below, but this one looks better. 

Shall I include syzbot link for the fix?

Thanks,
Song




diff --git i/include/linux/btf_ids.h w/include/linux/btf_ids.h
index 47d9abfbdb556..4153264c1236b 100644
--- i/include/linux/btf_ids.h
+++ w/include/linux/btf_ids.h
@@ -149,7 +149,7 @@ extern struct btf_id_set name;
 #define BTF_ID_LIST(name) static u32 name[5];
 #define BTF_ID(prefix, name)
 #define BTF_ID_UNUSED
-#define BTF_ID_LIST_GLOBAL(name) u32 name[1];
+#define BTF_ID_LIST_GLOBAL(name) u32 name[3];
 #define BTF_ID_LIST_SINGLE(name, prefix, typename) static u32 name[1];
 #define BTF_ID_LIST_GLOBAL_SINGLE(name, prefix, typename) u32 name[1];
 #define BTF_SET_START(name) static struct btf_id_set name = { 0 };