Re: [PATCH v2 1/2] bpf: extend stackmap to save binary_build_id+offset instead of address

[Song Liu]

> On Mar 6, 2018, at 10:25 AM, Peter Zijlstra  wrote:
> 
> On Tue, Mar 06, 2018 at 10:09:13AM -0800, Song Liu wrote:
>> +/* Parse build ID of ELF file mapped to vma */
>> +static int stack_map_get_build_id(struct vm_area_struct *vma,
>> +  unsigned char *build_id)
>> +{
>> +Elf32_Ehdr *ehdr;
>> +struct page *page;
>> +int ret;
>> +
>> +/*
>> + * vm_start is user memory, so we need to be careful with it.
>> + * We don't want too many copy_from_user to reduce overhead.
>> + * Most ELF file is at least one page long, and the build_id
>> + * is stored in the first page. Therefore, we limit the search of
>> + * build_id to the first page only. This can be made safe with
>> + * get_user_pages_fast(). If the file is smaller than PAGE_SIZE,
>> + * or the build_id is not in the first page, the look up fails.
>> + */
>> +if (vma->vm_end - vma->vm_start < PAGE_SIZE ||
>> +vma->vm_start & (PAGE_SIZE - 1))/* page aligned */
>> +return -EINVAL;
>> +
>> +if (get_user_pages_fast(vma->vm_start, 1, 0, ) != 1)
>> +return -EFAULT;
>> +
>> +ret = -EINVAL;
>> +ehdr = (Elf32_Ehdr *)vma->vm_start;
>> +
> 
> You're still directly accessing a user pointer afaict. This will
> insta-fault on SMAP enabled hardware due to the lack of STAC (or PAN on
> ARM).
> 
> You _really_ cannot just access user pointers without the proper APIs.
> 
> Did you maybe mean to use:
> 
>   ehdr = (Elf32_Ehdr *)page_address(page);
> 
> ?

Yeah, I missed this part. Let me fix it. 

Thanks again!
Song


>> +/* compare magic x7f "ELF" */
>> +if (memcmp(ehdr->e_ident, ELFMAG, SELFMAG) != 0)
>> +goto out;
>> +
>> +/* only support executable file and shared object file */
>> +if (ehdr->e_type != ET_EXEC && ehdr->e_type != ET_DYN)
>> +goto out;
>> +
>> +if (ehdr->e_ident[EI_CLASS] == 1)
>> +ret = stack_map_get_build_id_32(vma, build_id);
>> +else if (ehdr->e_ident[EI_CLASS] == 2)
>> +ret = stack_map_get_build_id_64(vma, build_id);
>> +out:
>> +put_page(page);
>> +return ret;
>> +}

Re: [PATCH v2 1/2] bpf: extend stackmap to save binary_build_id+offset instead of address

[Peter Zijlstra]

On Tue, Mar 06, 2018 at 10:09:13AM -0800, Song Liu wrote:
> +/* Parse build ID of ELF file mapped to vma */
> +static int stack_map_get_build_id(struct vm_area_struct *vma,
> +   unsigned char *build_id)
> +{
> + Elf32_Ehdr *ehdr;
> + struct page *page;
> + int ret;
> +
> + /*
> +  * vm_start is user memory, so we need to be careful with it.
> +  * We don't want too many copy_from_user to reduce overhead.
> +  * Most ELF file is at least one page long, and the build_id
> +  * is stored in the first page. Therefore, we limit the search of
> +  * build_id to the first page only. This can be made safe with
> +  * get_user_pages_fast(). If the file is smaller than PAGE_SIZE,
> +  * or the build_id is not in the first page, the look up fails.
> +  */
> + if (vma->vm_end - vma->vm_start < PAGE_SIZE ||
> + vma->vm_start & (PAGE_SIZE - 1))/* page aligned */
> + return -EINVAL;
> +
> + if (get_user_pages_fast(vma->vm_start, 1, 0, ) != 1)
> + return -EFAULT;
> +
> + ret = -EINVAL;
> + ehdr = (Elf32_Ehdr *)vma->vm_start;
> +

You're still directly accessing a user pointer afaict. This will
insta-fault on SMAP enabled hardware due to the lack of STAC (or PAN on
ARM).

You _really_ cannot just access user pointers without the proper APIs.

Did you maybe mean to use:

ehdr = (Elf32_Ehdr *)page_address(page);

?

> + /* compare magic x7f "ELF" */
> + if (memcmp(ehdr->e_ident, ELFMAG, SELFMAG) != 0)
> + goto out;
> +
> + /* only support executable file and shared object file */
> + if (ehdr->e_type != ET_EXEC && ehdr->e_type != ET_DYN)
> + goto out;
> +
> + if (ehdr->e_ident[EI_CLASS] == 1)
> + ret = stack_map_get_build_id_32(vma, build_id);
> + else if (ehdr->e_ident[EI_CLASS] == 2)
> + ret = stack_map_get_build_id_64(vma, build_id);
> +out:
> + put_page(page);
> + return ret;
> +}