From: Daniel Borkmann
Date: Mon, 15 May 2017 23:55:47 +0200
> I'm actually wondering about the min_align/aux_off/aux_off_align and
> given this is not really related to varlen_map_access and we currently
> just skip this.
>
> We should make sure that when env->strict_alignment is false that we
>
On 05/15/2017 05:34 PM, David Miller wrote:
From: Daniel Borkmann
Date: Mon, 15 May 2017 15:10:02 +0200
What are the semantics of using id here? In ptr_to_pkt, we have it,
so that eventually, in find_good_pkt_pointers() we can match on id
and update the range for all such regs with the same id
Just like packet pointers, track the known alignment of MAP pointers.
In order to facilitate the state tracking, move the register offset
field into where there is an unused 32-bit padding slot on 64-bit.
The check logic is the same as for packet pointers, except we do not
apply NET_IP_ALIGN to
From: Daniel Borkmann
Date: Mon, 15 May 2017 15:10:02 +0200
>>> What are the semantics of using id here? In ptr_to_pkt, we have it,
>>> so that eventually, in find_good_pkt_pointers() we can match on id
>>> and update the range for all such regs with the same id. I'm just
>>> wondering as the sid
On 05/15/2017 03:00 AM, David Miller wrote:
From: Daniel Borkmann
Date: Sun, 14 May 2017 16:31:10 +0200
On 05/13/2017 04:28 AM, David Miller wrote:
@@ -823,10 +825,27 @@ static int check_pkt_ptr_alignment(const struct
bpf_reg_state *reg,
}
static int check_val_ptr_alignment(const struc
From: Daniel Borkmann
Date: Sun, 14 May 2017 16:31:10 +0200
> On 05/13/2017 04:28 AM, David Miller wrote:
>> @@ -823,10 +825,27 @@ static int check_pkt_ptr_alignment(const struct
>> bpf_reg_state *reg,
>> }
>>
>> static int check_val_ptr_alignment(const struct bpf_reg_state *reg,
>> -
On 05/13/2017 04:28 AM, David Miller wrote:
Just like packet pointers, track the known alignment of MAP pointers.
In order to facilitate the state tracking, move the register offset
field into where there is an unused 32-bit padding slot on 64-bit.
The check logic is the same as for packet
Just like packet pointers, track the known alignment of MAP pointers.
In order to facilitate the state tracking, move the register offset
field into where there is an unused 32-bit padding slot on 64-bit.
The check logic is the same as for packet pointers, except we do not
apply NET_IP_ALIGN to
Daniel, this continues our discussion from yesterday where you
rightly pointed out that map pointers don't have their state
adjusted like packet pointers do currently.
I'm working on a patch sketched below to deal with this.
My understanding is that packet pointers are validated using the
accumu
