[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Mat Martineau]

[-- Attachment #1: Type: text/plain, Size: 1729 bytes --]


On Tue, 7 Jan 2020, Paolo Abeni wrote:

> On Tue, 2020-01-07 at 16:08 +0100, Matthieu Baerts wrote:
>> Hi Paolo,
>>
>> On 07/01/2020 15:05, Paolo Abeni wrote:
>>> We need to update the ebpf accessors for both sk_type and
>>> sk_protocol. The code previously in place has some special
>>> hack to take care such fields being part of a bitfield.
>>>
>>> The fix actually clean-up a bit the existing code.
>>
>> Thank you for looking at that!
>>
>> I have 3 questions below.
>>
>>> Note: I'm not 100% sure about dropping the 'target_size' handling.
>>> This changes the 'target_size' from 1 to 2 for sk_protocol.
>>> Should not matter, as the ebpf target field is 4 bytes wide.

It looks like your handling of target_size is consistent with other 16-bit 
fields (like skc_dport), so it seems like the right way to go.

>>>
>>> To be appended to the orig commit message:
>>> """
>>> Take care of BPF field access for sk_type/sk_protocol.
>>> Both of them are now outside the bitfield, so we can use
>>> load instructions without fourther shifting/masking.
>>
>> (detail: s/fourther/further)
>>
>>> v5 -> v6:
>>>   - update eBPF accessors, too.
>>> """
>>>
>>> Additionally, Eric's Reviewed-by should be dropped.
>>
>> Good point!
>>
>>> ---
>>>   include/net/sock.h | 15 ------------
>>>   net/core/filter.c  | 58 ++++++++++++++++------------------------------
>>>   2 files changed, 20 insertions(+), 53 deletions(-)
>>>

...

>
> I'll send v1 with the cleanup hinted by the 2nd question after doing
> some tests.
>

I don't have anything to add to Matthieu's comments for now, and thanks 
for tracking this down Paolo! Will watch for next version.

--
Mat Martineau
Intel

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 7447 bytes --]

On Tue, 2020-01-07 at 16:08 +0100, Matthieu Baerts wrote:
> Hi Paolo,
> 
> On 07/01/2020 15:05, Paolo Abeni wrote:
> > We need to update the ebpf accessors for both sk_type and
> > sk_protocol. The code previously in place has some special
> > hack to take care such fields being part of a bitfield.
> > 
> > The fix actually clean-up a bit the existing code.
> 
> Thank you for looking at that!
> 
> I have 3 questions below.
> 
> > Note: I'm not 100% sure about dropping the 'target_size' handling.
> > This changes the 'target_size' from 1 to 2 for sk_protocol.
> > Should not matter, as the ebpf target field is 4 bytes wide.
> > 
> > To be appended to the orig commit message:
> > """
> > Take care of BPF field access for sk_type/sk_protocol.
> > Both of them are now outside the bitfield, so we can use
> > load instructions without fourther shifting/masking.
> 
> (detail: s/fourther/further)
> 
> > v5 -> v6:
> >   - update eBPF accessors, too.
> > """
> > 
> > Additionally, Eric's Reviewed-by should be dropped.
> 
> Good point!
> 
> > ---
> >   include/net/sock.h | 15 ------------
> >   net/core/filter.c  | 58 ++++++++++++++++------------------------------
> >   2 files changed, 20 insertions(+), 53 deletions(-)
> > 
> > diff --git a/include/net/sock.h b/include/net/sock.h
> > index c63e9088b4e3..8766f9bc3e70 100644
> > --- a/include/net/sock.h
> > +++ b/include/net/sock.h
> > @@ -436,21 +436,6 @@ struct sock {
> >   	 * Because of non atomicity rules, all
> >   	 * changes are protected by socket lock.
> >   	 */
> > -	unsigned int		__sk_flags_offset[0];
> > -#ifdef __BIG_ENDIAN_BITFIELD
> > -#define SK_FL_PROTO_SHIFT  16
> > -#define SK_FL_PROTO_MASK   0x00ff0000
> > -
> > -#define SK_FL_TYPE_SHIFT   0
> > -#define SK_FL_TYPE_MASK    0x0000ffff
> > -#else
> > -#define SK_FL_PROTO_SHIFT  8
> > -#define SK_FL_PROTO_MASK   0x0000ff00
> > -
> > -#define SK_FL_TYPE_SHIFT   16
> > -#define SK_FL_TYPE_MASK    0xffff0000
> > -#endif
> > -
> >   	u8			sk_padding : 1,
> >   				sk_kern_sock : 1,
> >   				sk_no_check_tx : 1,
> > diff --git a/net/core/filter.c b/net/core/filter.c
> > index 42fd17c48c5f..37759d6ce3ea 100644
> > --- a/net/core/filter.c
> > +++ b/net/core/filter.c
> > @@ -7607,21 +7607,19 @@ u32 bpf_sock_convert_ctx_access(enum bpf_access_type type,
> >   		break;
> >   
> >   	case offsetof(struct bpf_sock, type):
> > -		BUILD_BUG_ON(HWEIGHT32(SK_FL_TYPE_MASK) != BITS_PER_BYTE * 2);
> 
> Should we not keep it?
> 
> I see that some others are still doing that:
> 
>      BUILD_BUG_ON(sizeof_field(struct sock, sk_mark) != 4);

The first BUILD_BUG reported above regards only the type mask. This
patch remove such definition, so we can't keep the build bug as-is.

The second one instead is still present, via the 'bpf_target_off()'
macro call.

> But not all, e.g. for sk_state, sk_num, sk_dport.
> 
> > -		*insn++ = BPF_LDX_MEM(BPF_W, si->dst_reg, si->src_reg,
> > -				      offsetof(struct sock, __sk_flags_offset));
> > -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_TYPE_MASK);
> > -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg, SK_FL_TYPE_SHIFT);
> > -		*target_size = 2;
> > +		*insn++ = BPF_LDX_MEM(BPF_H, si->dst_reg, si->src_reg,
> 
> Instead of BPF_H, should we use this macro?
> 
>      BPF_FIELD_SIZEOF(struct sock, sk_type)

Right u r, its' more consistent with the nearby code.

> > +				      bpf_target_off(struct sock, sk_type,
> > +						     sizeof_field(struct sock,
> > +								  sk_type),
> > +						     target_size));
> >   		break;
> >   
> >   	case offsetof(struct bpf_sock, protocol):
> > -		BUILD_BUG_ON(HWEIGHT32(SK_FL_PROTO_MASK) != BITS_PER_BYTE);
> > -		*insn++ = BPF_LDX_MEM(BPF_W, si->dst_reg, si->src_reg,
> > -				      offsetof(struct sock, __sk_flags_offset));
> > -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_PROTO_MASK);
> > -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg, SK_FL_PROTO_SHIFT);
> > -		*target_size = 1;
> > +		*insn++ = BPF_LDX_MEM(BPF_H, si->dst_reg, si->src_reg,
> > +				      bpf_target_off(struct sock, sk_protocol,
> > +						     sizeof_field(struct sock,
> > +								  sk_protocol),
> > +						     target_size));
> >   		break;
> >   
> >   	case offsetof(struct bpf_sock, src_ip4):
> > @@ -7903,20 +7901,13 @@ static u32 sock_addr_convert_ctx_access(enum bpf_access_type type,
> >   		break;
> >   
> >   	case offsetof(struct bpf_sock_addr, type):
> > -		SOCK_ADDR_LOAD_NESTED_FIELD_SIZE_OFF(
> > -			struct bpf_sock_addr_kern, struct sock, sk,
> > -			__sk_flags_offset, BPF_W, 0);
> > -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_TYPE_MASK);
> > -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg, SK_FL_TYPE_SHIFT);
> > +		SOCK_ADDR_LOAD_NESTED_FIELD(struct bpf_sock_addr_kern,
> > +					    struct sock, sk, sk_type);
> >   		break;
> >   
> >   	case offsetof(struct bpf_sock_addr, protocol):
> > -		SOCK_ADDR_LOAD_NESTED_FIELD_SIZE_OFF(
> > -			struct bpf_sock_addr_kern, struct sock, sk,
> > -			__sk_flags_offset, BPF_W, 0);
> > -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_PROTO_MASK);
> > -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg,
> > -					SK_FL_PROTO_SHIFT);
> > +		SOCK_ADDR_LOAD_NESTED_FIELD(struct bpf_sock_addr_kern,
> > +					    struct sock, sk, sk_protocol);
> >   		break;
> >   
> >   	case offsetof(struct bpf_sock_addr, msg_src_ip4):
> > @@ -8835,11 +8826,11 @@ sk_reuseport_is_valid_access(int off, int size,
> >   				    skb,				\
> >   				    SKB_FIELD)
> >   
> > -#define SK_REUSEPORT_LOAD_SK_FIELD_SIZE_OFF(SK_FIELD, BPF_SIZE, EXTRA_OFF) \
> > -	SOCK_ADDR_LOAD_NESTED_FIELD_SIZE_OFF(struct sk_reuseport_kern,	\
> > -					     struct sock,		\
> > -					     sk,			\
> > -					     SK_FIELD, BPF_SIZE, EXTRA_OFF)
> > +#define SK_REUSEPORT_LOAD_SK_FIELD(SK_FIELD)				\
> > +	SOCK_ADDR_LOAD_NESTED_FIELD(struct sk_reuseport_kern,		\
> > +				    struct sock,			\
> > +				    sk,					\
> > +				    SK_FIELD)
> 
> Why did you have to change that too?

SK_REUSEPORT_LOAD_SK_FIELD_SIZE_OFF() is unused after this patch.
So I crafted another helper to make sk_protocol access easier.

> 
> >   
> >   static u32 sk_reuseport_convert_ctx_access(enum bpf_access_type type,
> >   					   const struct bpf_insn *si,
> > @@ -8863,16 +8854,7 @@ static u32 sk_reuseport_convert_ctx_access(enum bpf_access_type type,
> >   		break;
> >   
> >   	case offsetof(struct sk_reuseport_md, ip_protocol):
> > -		BUILD_BUG_ON(HWEIGHT32(SK_FL_PROTO_MASK) != BITS_PER_BYTE);
> > -		SK_REUSEPORT_LOAD_SK_FIELD_SIZE_OFF(__sk_flags_offset,
> > -						    BPF_W, 0);
> > -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_PROTO_MASK);
> > -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg,
> > -					SK_FL_PROTO_SHIFT);
> > -		/* SK_FL_PROTO_MASK and SK_FL_PROTO_SHIFT are endian
> > -		 * aware.  No further narrowing or masking is needed.
> > -		 */
> > -		*target_size = 1;
> > +		SK_REUSEPORT_LOAD_SK_FIELD(sk_protocol);
> 
> Detail for me: why did they not modify the bitfield instead of adding 
> __sk_flags_offset[0] in include/net/sock.h ?

I guess to avoid touching a long established and delicate piece of
code. You know, you get quite serious feedback when doint that kind of
changes ;)

I'll send v1 with the cleanup hinted by the 2nd question after doing
some tests.

/P

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 6576 bytes --]

Hi Paolo,

On 07/01/2020 15:05, Paolo Abeni wrote:
> We need to update the ebpf accessors for both sk_type and
> sk_protocol. The code previously in place has some special
> hack to take care such fields being part of a bitfield.
> 
> The fix actually clean-up a bit the existing code.

Thank you for looking at that!

I have 3 questions below.

> Note: I'm not 100% sure about dropping the 'target_size' handling.
> This changes the 'target_size' from 1 to 2 for sk_protocol.
> Should not matter, as the ebpf target field is 4 bytes wide.
> 
> To be appended to the orig commit message:
> """
> Take care of BPF field access for sk_type/sk_protocol.
> Both of them are now outside the bitfield, so we can use
> load instructions without fourther shifting/masking.

(detail: s/fourther/further)

> 
> v5 -> v6:
>   - update eBPF accessors, too.
> """
> 
> Additionally, Eric's Reviewed-by should be dropped.

Good point!

> ---
>   include/net/sock.h | 15 ------------
>   net/core/filter.c  | 58 ++++++++++++++++------------------------------
>   2 files changed, 20 insertions(+), 53 deletions(-)
> 
> diff --git a/include/net/sock.h b/include/net/sock.h
> index c63e9088b4e3..8766f9bc3e70 100644
> --- a/include/net/sock.h
> +++ b/include/net/sock.h
> @@ -436,21 +436,6 @@ struct sock {
>   	 * Because of non atomicity rules, all
>   	 * changes are protected by socket lock.
>   	 */
> -	unsigned int		__sk_flags_offset[0];
> -#ifdef __BIG_ENDIAN_BITFIELD
> -#define SK_FL_PROTO_SHIFT  16
> -#define SK_FL_PROTO_MASK   0x00ff0000
> -
> -#define SK_FL_TYPE_SHIFT   0
> -#define SK_FL_TYPE_MASK    0x0000ffff
> -#else
> -#define SK_FL_PROTO_SHIFT  8
> -#define SK_FL_PROTO_MASK   0x0000ff00
> -
> -#define SK_FL_TYPE_SHIFT   16
> -#define SK_FL_TYPE_MASK    0xffff0000
> -#endif
> -
>   	u8			sk_padding : 1,
>   				sk_kern_sock : 1,
>   				sk_no_check_tx : 1,
> diff --git a/net/core/filter.c b/net/core/filter.c
> index 42fd17c48c5f..37759d6ce3ea 100644
> --- a/net/core/filter.c
> +++ b/net/core/filter.c
> @@ -7607,21 +7607,19 @@ u32 bpf_sock_convert_ctx_access(enum bpf_access_type type,
>   		break;
>   
>   	case offsetof(struct bpf_sock, type):
> -		BUILD_BUG_ON(HWEIGHT32(SK_FL_TYPE_MASK) != BITS_PER_BYTE * 2);

Should we not keep it?

I see that some others are still doing that:

     BUILD_BUG_ON(sizeof_field(struct sock, sk_mark) != 4);

But not all, e.g. for sk_state, sk_num, sk_dport.

> -		*insn++ = BPF_LDX_MEM(BPF_W, si->dst_reg, si->src_reg,
> -				      offsetof(struct sock, __sk_flags_offset));
> -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_TYPE_MASK);
> -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg, SK_FL_TYPE_SHIFT);
> -		*target_size = 2;
> +		*insn++ = BPF_LDX_MEM(BPF_H, si->dst_reg, si->src_reg,

Instead of BPF_H, should we use this macro?

     BPF_FIELD_SIZEOF(struct sock, sk_type)

> +				      bpf_target_off(struct sock, sk_type,
> +						     sizeof_field(struct sock,
> +								  sk_type),
> +						     target_size));
>   		break;
>   
>   	case offsetof(struct bpf_sock, protocol):
> -		BUILD_BUG_ON(HWEIGHT32(SK_FL_PROTO_MASK) != BITS_PER_BYTE);
> -		*insn++ = BPF_LDX_MEM(BPF_W, si->dst_reg, si->src_reg,
> -				      offsetof(struct sock, __sk_flags_offset));
> -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_PROTO_MASK);
> -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg, SK_FL_PROTO_SHIFT);
> -		*target_size = 1;
> +		*insn++ = BPF_LDX_MEM(BPF_H, si->dst_reg, si->src_reg,
> +				      bpf_target_off(struct sock, sk_protocol,
> +						     sizeof_field(struct sock,
> +								  sk_protocol),
> +						     target_size));
>   		break;
>   
>   	case offsetof(struct bpf_sock, src_ip4):
> @@ -7903,20 +7901,13 @@ static u32 sock_addr_convert_ctx_access(enum bpf_access_type type,
>   		break;
>   
>   	case offsetof(struct bpf_sock_addr, type):
> -		SOCK_ADDR_LOAD_NESTED_FIELD_SIZE_OFF(
> -			struct bpf_sock_addr_kern, struct sock, sk,
> -			__sk_flags_offset, BPF_W, 0);
> -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_TYPE_MASK);
> -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg, SK_FL_TYPE_SHIFT);
> +		SOCK_ADDR_LOAD_NESTED_FIELD(struct bpf_sock_addr_kern,
> +					    struct sock, sk, sk_type);
>   		break;
>   
>   	case offsetof(struct bpf_sock_addr, protocol):
> -		SOCK_ADDR_LOAD_NESTED_FIELD_SIZE_OFF(
> -			struct bpf_sock_addr_kern, struct sock, sk,
> -			__sk_flags_offset, BPF_W, 0);
> -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_PROTO_MASK);
> -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg,
> -					SK_FL_PROTO_SHIFT);
> +		SOCK_ADDR_LOAD_NESTED_FIELD(struct bpf_sock_addr_kern,
> +					    struct sock, sk, sk_protocol);
>   		break;
>   
>   	case offsetof(struct bpf_sock_addr, msg_src_ip4):
> @@ -8835,11 +8826,11 @@ sk_reuseport_is_valid_access(int off, int size,
>   				    skb,				\
>   				    SKB_FIELD)
>   
> -#define SK_REUSEPORT_LOAD_SK_FIELD_SIZE_OFF(SK_FIELD, BPF_SIZE, EXTRA_OFF) \
> -	SOCK_ADDR_LOAD_NESTED_FIELD_SIZE_OFF(struct sk_reuseport_kern,	\
> -					     struct sock,		\
> -					     sk,			\
> -					     SK_FIELD, BPF_SIZE, EXTRA_OFF)
> +#define SK_REUSEPORT_LOAD_SK_FIELD(SK_FIELD)				\
> +	SOCK_ADDR_LOAD_NESTED_FIELD(struct sk_reuseport_kern,		\
> +				    struct sock,			\
> +				    sk,					\
> +				    SK_FIELD)

Why did you have to change that too?

>   
>   static u32 sk_reuseport_convert_ctx_access(enum bpf_access_type type,
>   					   const struct bpf_insn *si,
> @@ -8863,16 +8854,7 @@ static u32 sk_reuseport_convert_ctx_access(enum bpf_access_type type,
>   		break;
>   
>   	case offsetof(struct sk_reuseport_md, ip_protocol):
> -		BUILD_BUG_ON(HWEIGHT32(SK_FL_PROTO_MASK) != BITS_PER_BYTE);
> -		SK_REUSEPORT_LOAD_SK_FIELD_SIZE_OFF(__sk_flags_offset,
> -						    BPF_W, 0);
> -		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_PROTO_MASK);
> -		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg,
> -					SK_FL_PROTO_SHIFT);
> -		/* SK_FL_PROTO_MASK and SK_FL_PROTO_SHIFT are endian
> -		 * aware.  No further narrowing or masking is needed.
> -		 */
> -		*target_size = 1;
> +		SK_REUSEPORT_LOAD_SK_FIELD(sk_protocol);

Detail for me: why did they not modify the bitfield instead of adding 
__sk_flags_offset[0] in include/net/sock.h ?

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 2068 bytes --]

Hi Paolo,

On 17/12/2019 12:08, Paolo Abeni wrote:
> On Tue, 2019-12-17 at 11:02 +0100, Matthieu Baerts wrote:
>> move sk_type to a separate field to:
>> - avoid extra operations to fetch sk->sk_type on some arches (Eric)
>> - allow compilers to emit better code to access/manipulate native types
>>    compared to bitfields. (Paolo)
>>
>> Suggested-by: Eric Dumazet <edumazet(a)google.com>
>> Suggested-by: Paolo Abeni <pabeni(a)redhat.com>
>> Signed-off-by: Matthieu Baerts <matthieu.baerts(a)tessares.net>
>> ---
>>
>> Notes:
>>      to be squashed in "sock: Make sk_protocol a 16-bit value"
>>
>>   include/net/sock.h | 6 +++---
>>   1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/include/net/sock.h b/include/net/sock.h
>> index 0930f46c600c..b93cadba1a3b 100644
>> --- a/include/net/sock.h
>> +++ b/include/net/sock.h
>> @@ -451,13 +451,13 @@ struct sock {
>>   #define SK_FL_TYPE_MASK    0xffff0000
>>   #endif
>>   
>> -	unsigned int		sk_padding : 1,
>> +	u8			sk_padding : 1,
>>   				sk_kern_sock : 1,
>>   				sk_no_check_tx : 1,
>>   				sk_no_check_rx : 1,
>> -				sk_userlocks : 4,
>> -				sk_type      : 16;
>> +				sk_userlocks : 4;
>>   	u8			sk_pacing_shift;
>> +	u16			sk_type;
>>   	u16			sk_protocol;
>>   	u16			sk_gso_max_segs;
>>   	unsigned long	        sk_lingertime;
> 
> LGTM.

Thank you for the review and the suggestion not to simply move it to the 
beginning of the 'unsigned int' but to move it to a separated field.

- 32121b5387c7: "squashed" in "sock: Make sk_protocol a 16-bit value"
- 1e3599433335: "Signed-off-by" + "Co-developed-by"
- 3a1b579608d1..bf07050cd31b: result

Tests are still OK!

> @list, please do not reply directly to the orig messages, or trim the
> recipients list, as eric email address unintentionally slipped into it.

Sorry for that :(

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 1415 bytes --]

On Tue, 2019-12-17 at 11:02 +0100, Matthieu Baerts wrote:
> move sk_type to a separate field to:
> - avoid extra operations to fetch sk->sk_type on some arches (Eric)
> - allow compilers to emit better code to access/manipulate native types
>   compared to bitfields. (Paolo)
> 
> Suggested-by: Eric Dumazet <edumazet(a)google.com>
> Suggested-by: Paolo Abeni <pabeni(a)redhat.com>
> Signed-off-by: Matthieu Baerts <matthieu.baerts(a)tessares.net>
> ---
> 
> Notes:
>     to be squashed in "sock: Make sk_protocol a 16-bit value"
> 
>  include/net/sock.h | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/include/net/sock.h b/include/net/sock.h
> index 0930f46c600c..b93cadba1a3b 100644
> --- a/include/net/sock.h
> +++ b/include/net/sock.h
> @@ -451,13 +451,13 @@ struct sock {
>  #define SK_FL_TYPE_MASK    0xffff0000
>  #endif
>  
> -	unsigned int		sk_padding : 1,
> +	u8			sk_padding : 1,
>  				sk_kern_sock : 1,
>  				sk_no_check_tx : 1,
>  				sk_no_check_rx : 1,
> -				sk_userlocks : 4,
> -				sk_type      : 16;
> +				sk_userlocks : 4;
>  	u8			sk_pacing_shift;
> +	u16			sk_type;
>  	u16			sk_protocol;
>  	u16			sk_gso_max_segs;
>  	unsigned long	        sk_lingertime;

LGTM.

@list, please do not reply directly to the orig messages, or trim the
recipients list, as eric email address unintentionally slipped into it.

/P

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Mat Martineau]

[-- Attachment #1: Type: text/plain, Size: 1583 bytes --]


On Mon, 16 Dec 2019, Paolo Abeni wrote:

> On Mon, 2019-12-16 at 08:51 -0800, Mat Martineau wrote:
>> On Mon, 16 Dec 2019, Paolo Abeni wrote:
>>
>>> move sk_pacing_shift to a separate field, so we can READ_ONCE() it.
>>>
>>> Binary layout of struct sock dose not change (except for the
>>> sk_pacing_shift position)
>>> ---
>>> This addresses Eric's upstream feedback on v1
>>> ---
>>> include/net/sock.h | 2 +-
>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/include/net/sock.h b/include/net/sock.h
>>> index 9dd225f62012..0930f46c600c 100644
>>> --- a/include/net/sock.h
>>> +++ b/include/net/sock.h
>>> @@ -456,8 +456,8 @@ struct sock {
>>> 				sk_no_check_tx : 1,
>>> 				sk_no_check_rx : 1,
>>> 				sk_userlocks : 4,
>>> -				sk_pacing_shift : 8,
>>> 				sk_type      : 16;
>>> +	u8			sk_pacing_shift;
>>> 	u16			sk_protocol;
>>> 	u16			sk_gso_max_segs;
>>
>> sk_pacing_shift was here (between sk_gso_max_segs and sk_lingertime)
>> before, would be good to put it back where it was for a clean diff.
>
> But we can't move back there, otherwise the 'struct sock' will grow
> quite a bit (we will have 7 bytes hole on 64 bits arch).
>
> With the proposed patch 'struct sock' size does not change.
>

I didn't know gcc was smart enough to pack the u8 in with the previous 
bitfield! Learned something new.

I think Paolo's fix is ok as-is, unless it makes sense to worry about 
2-byte alignment of sk_type. It does appear to land on an odd byte now 
(checked with objdump).


Thanks

--
Mat Martineau
Intel

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 1677 bytes --]

Hi Paolo, Mat,

On 16/12/2019 18:47, Paolo Abeni wrote:
> On Mon, 2019-12-16 at 08:51 -0800, Mat Martineau wrote:
>> On Mon, 16 Dec 2019, Paolo Abeni wrote:
>>
>>> move sk_pacing_shift to a separate field, so we can READ_ONCE() it.
>>>
>>> Binary layout of struct sock dose not change (except for the
>>> sk_pacing_shift position)
>>> ---
>>> This addresses Eric's upstream feedback on v1
>>> ---
>>> include/net/sock.h | 2 +-
>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/include/net/sock.h b/include/net/sock.h
>>> index 9dd225f62012..0930f46c600c 100644
>>> --- a/include/net/sock.h
>>> +++ b/include/net/sock.h
>>> @@ -456,8 +456,8 @@ struct sock {
>>> 				sk_no_check_tx : 1,
>>> 				sk_no_check_rx : 1,
>>> 				sk_userlocks : 4,
>>> -				sk_pacing_shift : 8,
>>> 				sk_type      : 16;
>>> +	u8			sk_pacing_shift;
>>> 	u16			sk_protocol;
>>> 	u16			sk_gso_max_segs;
>>
>> sk_pacing_shift was here (between sk_gso_max_segs and sk_lingertime)
>> before, would be good to put it back where it was for a clean diff.
> 
> But we can't move back there, otherwise the 'struct sock' will grow
> quite a bit (we will have 7 bytes hole on 64 bits arch).
> 
> With the proposed patch 'struct sock' size does not change.

- 548181045475: "squashed" in "sock: Make sk_protocol a 16-bit value"
- 4d09491bafaf: "Signed-off-by" + "Co-developed-by"
- 66880827b48f..e999bb17232e: result

Tests + export will be launched soon.

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 1226 bytes --]

On Mon, 2019-12-16 at 08:51 -0800, Mat Martineau wrote:
> On Mon, 16 Dec 2019, Paolo Abeni wrote:
> 
> > move sk_pacing_shift to a separate field, so we can READ_ONCE() it.
> > 
> > Binary layout of struct sock dose not change (except for the
> > sk_pacing_shift position)
> > ---
> > This addresses Eric's upstream feedback on v1
> > ---
> > include/net/sock.h | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/include/net/sock.h b/include/net/sock.h
> > index 9dd225f62012..0930f46c600c 100644
> > --- a/include/net/sock.h
> > +++ b/include/net/sock.h
> > @@ -456,8 +456,8 @@ struct sock {
> > 				sk_no_check_tx : 1,
> > 				sk_no_check_rx : 1,
> > 				sk_userlocks : 4,
> > -				sk_pacing_shift : 8,
> > 				sk_type      : 16;
> > +	u8			sk_pacing_shift;
> > 	u16			sk_protocol;
> > 	u16			sk_gso_max_segs;
> 
> sk_pacing_shift was here (between sk_gso_max_segs and sk_lingertime) 
> before, would be good to put it back where it was for a clean diff.

But we can't move back there, otherwise the 'struct sock' will grow
quite a bit (we will have 7 bytes hole on 64 bits arch).

With the proposed patch 'struct sock' size does not change.

Cheers,

Paolo

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Mat Martineau]

[-- Attachment #1: Type: text/plain, Size: 987 bytes --]


On Mon, 16 Dec 2019, Paolo Abeni wrote:

> move sk_pacing_shift to a separate field, so we can READ_ONCE() it.
>
> Binary layout of struct sock dose not change (except for the
> sk_pacing_shift position)
> ---
> This addresses Eric's upstream feedback on v1
> ---
> include/net/sock.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/include/net/sock.h b/include/net/sock.h
> index 9dd225f62012..0930f46c600c 100644
> --- a/include/net/sock.h
> +++ b/include/net/sock.h
> @@ -456,8 +456,8 @@ struct sock {
> 				sk_no_check_tx : 1,
> 				sk_no_check_rx : 1,
> 				sk_userlocks : 4,
> -				sk_pacing_shift : 8,
> 				sk_type      : 16;
> +	u8			sk_pacing_shift;
> 	u16			sk_protocol;
> 	u16			sk_gso_max_segs;

sk_pacing_shift was here (between sk_gso_max_segs and sk_lingertime) 
before, would be good to put it back where it was for a clean diff.

> 	unsigned long	        sk_lingertime;
> -- 
> 2.21.0

--
Mat Martineau
Intel

[MPTCP] Re: [PATCH] Squash-to: "sock: Make sk_protocol a 16-bit value"

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 979 bytes --]

Hi Paolo,

On 16/12/2019 13:31, Paolo Abeni wrote:
> move sk_pacing_shift to a separate field, so we can READ_ONCE() it.

Thank you addressing Eric's comments!

> Binary layout of struct sock dose not change (except for the
> sk_pacing_shift position)

Just to be sure, does the compiler will optimize the structure to remove 
the "hole" we have here? We use 24/32 bits of the "unsigned int" and 
then we declare the u8. But because we don't declare the "__packed__" 
attribute, the compiler will do his job and we don't increase the size. 
Right? :)

> ---
> This addresses Eric's upstream feedback on v1

I guess we should add this in the commit message.

Or:

   v1:
   - move sk_pacing_shift to a separate field, so we can READ_ONCE() it 
(Eric Dumazet)

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium