On 4/25/19 7:32 PM, Jann Horn wrote:
> On Thu, Apr 25, 2019 at 6:14 PM Dmitry Safonov <d...@arista.com> wrote:
>>
>> As it has been discussed on timens RFC, adding a new conditional branch
>> `if (inside_time_ns)` on VDSO for all processes is undesirable.
>> It will add a penalty for everybody as branch predictor may mispredict
>> the jump. Also there are instruction cache lines wasted on cmp/jmp.
>>
>> Those effects of introducing time namespace are very much unwanted
>> having in mind how much work have been spent on micro-optimisation
>> vdso code.
>>
>> The propose is to allocate a second vdso code with dynamically
>> patched out (disabled by static_branch) timens code on boot time.
>>
>> Allocate another vdso and copy original code.
> [...]
>> diff --git a/arch/x86/entry/vdso/vma.c b/arch/x86/entry/vdso/vma.c
>> index 80cbb2167eba..6aae9c0d400d 100644
>> --- a/arch/x86/entry/vdso/vma.c
>> +++ b/arch/x86/entry/vdso/vma.c
> [...]
>> static vm_fault_t vdso_fault(const struct vm_special_mapping *sm,
>> struct vm_area_struct *vma, struct vm_fault *vmf)
>> {
>> const struct vdso_image *image = vma->vm_mm->context.vdso_image;
>> + unsigned long offset = vmf->pgoff << PAGE_SHIFT;
>>
>> if (!image || (vmf->pgoff << PAGE_SHIFT) >= image->size)
>> return VM_FAULT_SIGBUS;
>>
>> - vmf->page = virt_to_page(image->text + (vmf->pgoff << PAGE_SHIFT));
>> + if (current_timens_offsets() && image->text_timens)
>
> I'm pretty sure that accessing `current` in here is wrong. AFAIK this
> fault handler can be invoked on remote processes, through interfaces
> like /proc/$pid/mem and process_vm_readv(); in that case, the kernel
> should install a page based on the time namespace of the target
> process, not based on the time namespace of the caller.
Oh yeah, I see, smells bogus.
Will try to redesign it..
>
>> + vmf->page = vmalloc_to_page(image->text_timens + offset);
>> + else
>> + vmf->page = virt_to_page(image->text + offset);
>> +
>> get_page(vmf->page);
>> return 0;
>> }
> [...]
>
Thanks,
Dima
