[48/64] drbd: Use struct_group() to zero algs

Message ID	20210727205855.411487-49-keescook@chromium.org (mailing list archive)
State	Not Applicable
Headers	show Return-Path: <netdev-owner@kernel.org> From: Kees Cook <keescook@chromium.org> To: linux-hardening@vger.kernel.org Cc: Kees Cook <keescook@chromium.org>, "Gustavo A. R. Silva" <gustavoars@kernel.org>, Keith Packard <keithpac@amazon.com>, Greg Kroah-Hartman <gregkh@linuxfoundation.org>, Andrew Morton <akpm@linux-foundation.org>, linux-kernel@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, dri-devel@lists.freedesktop.org, linux-staging@lists.linux.dev, linux-block@vger.kernel.org, linux-kbuild@vger.kernel.org, clang-built-linux@googlegroups.com Subject: [PATCH 48/64] drbd: Use struct_group() to zero algs Date: Tue, 27 Jul 2021 13:58:39 -0700 Message-Id: <20210727205855.411487-49-keescook@chromium.org> In-Reply-To: <20210727205855.411487-1-keescook@chromium.org> References: <20210727205855.411487-1-keescook@chromium.org> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Precedence: bulk
Series	Introduce strict memcpy() bounds checking \| expand [00/64] Introduce strict memcpy() bounds checking [01/64] media: omap3isp: Extract struct group for memcpy() region [02/64] mac80211: Use flex-array for radiotap header bitmap [03/64] rpmsg: glink: Replace strncpy() with strscpy_pad() [04/64] stddef: Introduce struct_group() helper macro [05/64] skbuff: Switch structure bounds to struct_group() [06/64] bnxt_en: Use struct_group_attr() for memcpy() region [07/64] staging: rtl8192e: Use struct_group() for memcpy() region [08/64] staging: rtl8192u: Use struct_group() for memcpy() region [09/64] staging: rtl8723bs: Avoid field-overflowing memcpy() [10/64] lib80211: Use struct_group() for memcpy() region [11/64] net/mlx5e: Avoid field-overflowing memcpy() [12/64] mwl8k: Use struct_group() for memcpy() region [13/64] libertas: Use struct_group() for memcpy() region [14/64] libertas_tf: Use struct_group() for memcpy() region [15/64] ipw2x00: Use struct_group() for memcpy() region [16/64] thermal: intel: int340x_thermal: Use struct_group() for memcpy() region [17/64] iommu/amd: Use struct_group() for memcpy() region [18/64] cxgb3: Use struct_group() for memcpy() region [19/64] ip: Use struct_group() for memcpy() regions [20/64] intersil: Use struct_group() for memcpy() region [21/64] cxgb4: Use struct_group() for memcpy() region [22/64] bnx2x: Use struct_group() for memcpy() region [23/64] drm/amd/pm: Use struct_group() for memcpy() region [24/64] staging: wlan-ng: Use struct_group() for memcpy() region [25/64] drm/mga/mga_ioc32: Use struct_group() for memcpy() region [26/64] net/mlx5e: Use struct_group() for memcpy() region [27/64] HID: cp2112: Use struct_group() for memcpy() region [28/64] compiler_types.h: Remove __compiletime_object_size() [29/64] lib/string: Move helper functions out of string.c [30/64] fortify: Move remaining fortify helpers into fortify-string.h [31/64] fortify: Explicitly disable Clang support [32/64] fortify: Add compile-time FORTIFY_SOURCE tests [33/64] lib: Introduce CONFIG_TEST_MEMCPY [34/64] fortify: Detect struct member overflows in memcpy() at compile-time [35/64] fortify: Detect struct member overflows in memmove() at compile-time [36/64] scsi: ibmvscsi: Avoid multi-field memset() overflow by aiming at srp [37/64] string.h: Introduce memset_after() for wiping trailing members/padding [38/64] xfrm: Use memset_after() to clear padding [39/64] mac80211: Use memset_after() to clear tx status [40/64] net: 802: Use memset_after() to clear struct fields [41/64] net: dccp: Use memset_after() for TP zeroing [42/64] net: qede: Use memset_after() for counters [43/64] ath11k: Use memset_after() for clearing queue descriptors [44/64] iw_cxgb4: Use memset_after() for cpl_t5_pass_accept_rpl [45/64] intel_th: msu: Use memset_after() for clearing hw header [46/64] IB/mthca: Use memset_after() for clearing mpt_entry [47/64] btrfs: Use memset_after() to clear end of struct [48/64] drbd: Use struct_group() to zero algs [49/64] cm4000_cs: Use struct_group() to zero struct cm4000_dev region [50/64] KVM: x86: Use struct_group() to zero decode cache [51/64] tracing: Use struct_group() to zero struct trace_iterator [52/64] dm integrity: Use struct_group() to zero struct journal_sector [53/64] HID: roccat: Use struct_group() to zero kone_mouse_event [54/64] ipv6: Use struct_group() to zero rt6_info [55/64] RDMA/mlx5: Use struct_group() to zero struct mlx5_ib_mr [56/64] ethtool: stats: Use struct_group() to clear all stats at once [57/64] netfilter: conntrack: Use struct_group() to zero struct nf_conn [58/64] powerpc: Split memset() to avoid multi-field overflow [59/64] fortify: Detect struct member overflows in memset() at compile-time [60/64] fortify: Work around Clang inlining bugs [61/64] Makefile: Enable -Warray-bounds [62/64] netlink: Avoid false-positive memcpy() warning [63/64] iwlwifi: dbg_ini: Split memcpy() to avoid multi-field write [64/64] fortify: Add run-time WARN for cross-field memcpy()

Context	Check	Description
netdev/tree_selection	success	Guessing tree name failed - patch did not apply

Kees Cook July 27, 2021, 8:58 p.m. UTC

In preparation for FORTIFY_SOURCE performing compile-time and run-time
field bounds checking for memset(), avoid intentionally writing across
neighboring fields.

Add a struct_group() for the algs so that memset() can correctly reason
about the size.

Signed-off-by: Kees Cook <keescook@chromium.org>
---
 drivers/block/drbd/drbd_main.c     | 3 ++-
 drivers/block/drbd/drbd_protocol.h | 6 ++++--
 drivers/block/drbd/drbd_receiver.c | 3 ++-
 3 files changed, 8 insertions(+), 4 deletions(-)

Bart Van Assche July 28, 2021, 9:45 p.m. UTC | #1

On 7/27/21 1:58 PM, Kees Cook wrote:
> In preparation for FORTIFY_SOURCE performing compile-time and run-time
> field bounds checking for memset(), avoid intentionally writing across
> neighboring fields.
> 
> Add a struct_group() for the algs so that memset() can correctly reason
> about the size.
> 
> Signed-off-by: Kees Cook <keescook@chromium.org>
> ---
>   drivers/block/drbd/drbd_main.c     | 3 ++-
>   drivers/block/drbd/drbd_protocol.h | 6 ++++--
>   drivers/block/drbd/drbd_receiver.c | 3 ++-
>   3 files changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
> index 55234a558e98..b824679cfcb2 100644
> --- a/drivers/block/drbd/drbd_main.c
> +++ b/drivers/block/drbd/drbd_main.c
> @@ -729,7 +729,8 @@ int drbd_send_sync_param(struct drbd_peer_device *peer_device)
>   	cmd = apv >= 89 ? P_SYNC_PARAM89 : P_SYNC_PARAM;
>   
>   	/* initialize verify_alg and csums_alg */
> -	memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
> +	BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
> +	memset(&p->algs, 0, sizeof(p->algs));
>   
>   	if (get_ldev(peer_device->device)) {
>   		dc = rcu_dereference(peer_device->device->ldev->disk_conf);
> diff --git a/drivers/block/drbd/drbd_protocol.h b/drivers/block/drbd/drbd_protocol.h
> index dea59c92ecc1..a882b65ab5d2 100644
> --- a/drivers/block/drbd/drbd_protocol.h
> +++ b/drivers/block/drbd/drbd_protocol.h
> @@ -283,8 +283,10 @@ struct p_rs_param_89 {
>   
>   struct p_rs_param_95 {
>   	u32 resync_rate;
> -	char verify_alg[SHARED_SECRET_MAX];
> -	char csums_alg[SHARED_SECRET_MAX];
> +	struct_group(algs,
> +		char verify_alg[SHARED_SECRET_MAX];
> +		char csums_alg[SHARED_SECRET_MAX];
> +	);
>   	u32 c_plan_ahead;
>   	u32 c_delay_target;
>   	u32 c_fill_target;
> diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
> index 1f740e42e457..6df2539e215b 100644
> --- a/drivers/block/drbd/drbd_receiver.c
> +++ b/drivers/block/drbd/drbd_receiver.c
> @@ -3921,7 +3921,8 @@ static int receive_SyncParam(struct drbd_connection *connection, struct packet_i
>   
>   	/* initialize verify_alg and csums_alg */
>   	p = pi->data;
> -	memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
> +	BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
> +	memset(&p->algs, 0, sizeof(p->algs));

Using struct_group() introduces complexity. Has it been considered not 
to modify struct p_rs_param_95 and instead to use two memset() calls 
instead of one (one memset() call per member)?

Thanks,

Bart.

Kees Cook July 30, 2021, 2:31 a.m. UTC | #2

On Wed, Jul 28, 2021 at 02:45:55PM -0700, Bart Van Assche wrote:
> On 7/27/21 1:58 PM, Kees Cook wrote:
> > In preparation for FORTIFY_SOURCE performing compile-time and run-time
> > field bounds checking for memset(), avoid intentionally writing across
> > neighboring fields.
> > 
> > Add a struct_group() for the algs so that memset() can correctly reason
> > about the size.
> > 
> > Signed-off-by: Kees Cook <keescook@chromium.org>
> > ---
> >   drivers/block/drbd/drbd_main.c     | 3 ++-
> >   drivers/block/drbd/drbd_protocol.h | 6 ++++--
> >   drivers/block/drbd/drbd_receiver.c | 3 ++-
> >   3 files changed, 8 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
> > index 55234a558e98..b824679cfcb2 100644
> > --- a/drivers/block/drbd/drbd_main.c
> > +++ b/drivers/block/drbd/drbd_main.c
> > @@ -729,7 +729,8 @@ int drbd_send_sync_param(struct drbd_peer_device *peer_device)
> >   	cmd = apv >= 89 ? P_SYNC_PARAM89 : P_SYNC_PARAM;
> >   	/* initialize verify_alg and csums_alg */
> > -	memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
> > +	BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
> > +	memset(&p->algs, 0, sizeof(p->algs));
> >   	if (get_ldev(peer_device->device)) {
> >   		dc = rcu_dereference(peer_device->device->ldev->disk_conf);
> > diff --git a/drivers/block/drbd/drbd_protocol.h b/drivers/block/drbd/drbd_protocol.h
> > index dea59c92ecc1..a882b65ab5d2 100644
> > --- a/drivers/block/drbd/drbd_protocol.h
> > +++ b/drivers/block/drbd/drbd_protocol.h
> > @@ -283,8 +283,10 @@ struct p_rs_param_89 {
> >   struct p_rs_param_95 {
> >   	u32 resync_rate;
> > -	char verify_alg[SHARED_SECRET_MAX];
> > -	char csums_alg[SHARED_SECRET_MAX];
> > +	struct_group(algs,
> > +		char verify_alg[SHARED_SECRET_MAX];
> > +		char csums_alg[SHARED_SECRET_MAX];
> > +	);
> >   	u32 c_plan_ahead;
> >   	u32 c_delay_target;
> >   	u32 c_fill_target;
> > diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
> > index 1f740e42e457..6df2539e215b 100644
> > --- a/drivers/block/drbd/drbd_receiver.c
> > +++ b/drivers/block/drbd/drbd_receiver.c
> > @@ -3921,7 +3921,8 @@ static int receive_SyncParam(struct drbd_connection *connection, struct packet_i
> >   	/* initialize verify_alg and csums_alg */
> >   	p = pi->data;
> > -	memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
> > +	BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
> > +	memset(&p->algs, 0, sizeof(p->algs));
> 
> Using struct_group() introduces complexity. Has it been considered not to
> modify struct p_rs_param_95 and instead to use two memset() calls instead of
> one (one memset() call per member)?

I went this direction because using two memset()s (or memcpy()s in other
patches) changes the machine code. It's not much of a change, but it
seems easier to justify "no binary changes" via the use of struct_group().

If splitting the memset() is preferred, I can totally do that instead.
:)

-Kees

Bart Van Assche July 30, 2021, 2:57 a.m. UTC | #3

On 7/29/21 7:31 PM, Kees Cook wrote:
> On Wed, Jul 28, 2021 at 02:45:55PM -0700, Bart Van Assche wrote:
>> On 7/27/21 1:58 PM, Kees Cook wrote:
>>> In preparation for FORTIFY_SOURCE performing compile-time and run-time
>>> field bounds checking for memset(), avoid intentionally writing across
>>> neighboring fields.
>>>
>>> Add a struct_group() for the algs so that memset() can correctly reason
>>> about the size.
>>>
>>> Signed-off-by: Kees Cook <keescook@chromium.org>
>>> ---
>>>   drivers/block/drbd/drbd_main.c     | 3 ++-
>>>   drivers/block/drbd/drbd_protocol.h | 6 ++++--
>>>   drivers/block/drbd/drbd_receiver.c | 3 ++-
>>>   3 files changed, 8 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
>>> index 55234a558e98..b824679cfcb2 100644
>>> --- a/drivers/block/drbd/drbd_main.c
>>> +++ b/drivers/block/drbd/drbd_main.c
>>> @@ -729,7 +729,8 @@ int drbd_send_sync_param(struct drbd_peer_device *peer_device)
>>>   	cmd = apv >= 89 ? P_SYNC_PARAM89 : P_SYNC_PARAM;
>>>   	/* initialize verify_alg and csums_alg */
>>> -	memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
>>> +	BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
>>> +	memset(&p->algs, 0, sizeof(p->algs));
>>>   	if (get_ldev(peer_device->device)) {
>>>   		dc = rcu_dereference(peer_device->device->ldev->disk_conf);
>>> diff --git a/drivers/block/drbd/drbd_protocol.h b/drivers/block/drbd/drbd_protocol.h
>>> index dea59c92ecc1..a882b65ab5d2 100644
>>> --- a/drivers/block/drbd/drbd_protocol.h
>>> +++ b/drivers/block/drbd/drbd_protocol.h
>>> @@ -283,8 +283,10 @@ struct p_rs_param_89 {
>>>   struct p_rs_param_95 {
>>>   	u32 resync_rate;
>>> -	char verify_alg[SHARED_SECRET_MAX];
>>> -	char csums_alg[SHARED_SECRET_MAX];
>>> +	struct_group(algs,
>>> +		char verify_alg[SHARED_SECRET_MAX];
>>> +		char csums_alg[SHARED_SECRET_MAX];
>>> +	);
>>>   	u32 c_plan_ahead;
>>>   	u32 c_delay_target;
>>>   	u32 c_fill_target;
>>> diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
>>> index 1f740e42e457..6df2539e215b 100644
>>> --- a/drivers/block/drbd/drbd_receiver.c
>>> +++ b/drivers/block/drbd/drbd_receiver.c
>>> @@ -3921,7 +3921,8 @@ static int receive_SyncParam(struct drbd_connection *connection, struct packet_i
>>>   	/* initialize verify_alg and csums_alg */
>>>   	p = pi->data;
>>> -	memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
>>> +	BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
>>> +	memset(&p->algs, 0, sizeof(p->algs));
>>
>> Using struct_group() introduces complexity. Has it been considered not to
>> modify struct p_rs_param_95 and instead to use two memset() calls instead of
>> one (one memset() call per member)?
> 
> I went this direction because using two memset()s (or memcpy()s in other
> patches) changes the machine code. It's not much of a change, but it
> seems easier to justify "no binary changes" via the use of struct_group().
> 
> If splitting the memset() is preferred, I can totally do that instead.
> :)

I don't have a strong opinion about this. Lars, do you want to comment
on this patch?

Thanks,

Bart.

Lars Ellenberg July 30, 2021, 9:25 a.m. UTC | #4

On Thu, Jul 29, 2021 at 07:57:47PM -0700, Bart Van Assche wrote:
> On 7/29/21 7:31 PM, Kees Cook wrote:
> > On Wed, Jul 28, 2021 at 02:45:55PM -0700, Bart Van Assche wrote:
> >> On 7/27/21 1:58 PM, Kees Cook wrote:
> >>> In preparation for FORTIFY_SOURCE performing compile-time and run-time
> >>> field bounds checking for memset(), avoid intentionally writing across
> >>> neighboring fields.
> >>>
> >>> Add a struct_group() for the algs so that memset() can correctly reason
> >>> about the size.
> >>>
> >>> Signed-off-by: Kees Cook <keescook@chromium.org>
> >>> ---
> >>>   drivers/block/drbd/drbd_main.c     | 3 ++-
> >>>   drivers/block/drbd/drbd_protocol.h | 6 ++++--
> >>>   drivers/block/drbd/drbd_receiver.c | 3 ++-
> >>>   3 files changed, 8 insertions(+), 4 deletions(-)
> >>>
> >>> diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
> >>> index 55234a558e98..b824679cfcb2 100644
> >>> --- a/drivers/block/drbd/drbd_main.c
> >>> +++ b/drivers/block/drbd/drbd_main.c
> >>> @@ -729,7 +729,8 @@ int drbd_send_sync_param(struct drbd_peer_device *peer_device)
> >>>   	cmd = apv >= 89 ? P_SYNC_PARAM89 : P_SYNC_PARAM;
> >>>   	/* initialize verify_alg and csums_alg */
> >>> -	memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
> >>> +	BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
> >>> +	memset(&p->algs, 0, sizeof(p->algs));
> >>>   	if (get_ldev(peer_device->device)) {
> >>>   		dc = rcu_dereference(peer_device->device->ldev->disk_conf);
> >>> diff --git a/drivers/block/drbd/drbd_protocol.h b/drivers/block/drbd/drbd_protocol.h
> >>> index dea59c92ecc1..a882b65ab5d2 100644
> >>> --- a/drivers/block/drbd/drbd_protocol.h
> >>> +++ b/drivers/block/drbd/drbd_protocol.h
> >>> @@ -283,8 +283,10 @@ struct p_rs_param_89 {
> >>>   struct p_rs_param_95 {
> >>>   	u32 resync_rate;
> >>> -	char verify_alg[SHARED_SECRET_MAX];
> >>> -	char csums_alg[SHARED_SECRET_MAX];
> >>> +	struct_group(algs,
> >>> +		char verify_alg[SHARED_SECRET_MAX];
> >>> +		char csums_alg[SHARED_SECRET_MAX];
> >>> +	);
> >>>   	u32 c_plan_ahead;
> >>>   	u32 c_delay_target;
> >>>   	u32 c_fill_target;
> >>> diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
> >>> index 1f740e42e457..6df2539e215b 100644
> >>> --- a/drivers/block/drbd/drbd_receiver.c
> >>> +++ b/drivers/block/drbd/drbd_receiver.c
> >>> @@ -3921,7 +3921,8 @@ static int receive_SyncParam(struct drbd_connection *connection, struct packet_i
> >>>   	/* initialize verify_alg and csums_alg */
> >>>   	p = pi->data;
> >>> -	memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
> >>> +	BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
> >>> +	memset(&p->algs, 0, sizeof(p->algs));
> >>
> >> Using struct_group() introduces complexity. Has it been considered not to
> >> modify struct p_rs_param_95 and instead to use two memset() calls instead of
> >> one (one memset() call per member)?
> > 
> > I went this direction because using two memset()s (or memcpy()s in other
> > patches) changes the machine code. It's not much of a change, but it
> > seems easier to justify "no binary changes" via the use of struct_group().
> > 
> > If splitting the memset() is preferred, I can totally do that instead.
> > :)
> 
> I don't have a strong opinion about this. Lars, do you want to comment
> on this patch?


Fine either way. "no binary changes" sounds good ;-)

Thanks,
    Lars

Nick Desaulniers July 30, 2021, 3:32 p.m. UTC | #5

On Thu, Jul 29, 2021 at 7:31 PM Kees Cook <keescook@chromium.org> wrote:
>
> On Wed, Jul 28, 2021 at 02:45:55PM -0700, Bart Van Assche wrote:
> > On 7/27/21 1:58 PM, Kees Cook wrote:
> > > In preparation for FORTIFY_SOURCE performing compile-time and run-time
> > > field bounds checking for memset(), avoid intentionally writing across
> > > neighboring fields.
> > >
> > > Add a struct_group() for the algs so that memset() can correctly reason
> > > about the size.
> > >
> > > Signed-off-by: Kees Cook <keescook@chromium.org>
> > > ---
> > >   drivers/block/drbd/drbd_main.c     | 3 ++-
> > >   drivers/block/drbd/drbd_protocol.h | 6 ++++--
> > >   drivers/block/drbd/drbd_receiver.c | 3 ++-
> > >   3 files changed, 8 insertions(+), 4 deletions(-)
> > >
> > > diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
> > > index 55234a558e98..b824679cfcb2 100644
> > > --- a/drivers/block/drbd/drbd_main.c
> > > +++ b/drivers/block/drbd/drbd_main.c
> > > @@ -729,7 +729,8 @@ int drbd_send_sync_param(struct drbd_peer_device *peer_device)
> > >     cmd = apv >= 89 ? P_SYNC_PARAM89 : P_SYNC_PARAM;
> > >     /* initialize verify_alg and csums_alg */
> > > -   memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
> > > +   BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
> > > +   memset(&p->algs, 0, sizeof(p->algs));
> > >     if (get_ldev(peer_device->device)) {
> > >             dc = rcu_dereference(peer_device->device->ldev->disk_conf);
> > > diff --git a/drivers/block/drbd/drbd_protocol.h b/drivers/block/drbd/drbd_protocol.h
> > > index dea59c92ecc1..a882b65ab5d2 100644
> > > --- a/drivers/block/drbd/drbd_protocol.h
> > > +++ b/drivers/block/drbd/drbd_protocol.h
> > > @@ -283,8 +283,10 @@ struct p_rs_param_89 {
> > >   struct p_rs_param_95 {
> > >     u32 resync_rate;
> > > -   char verify_alg[SHARED_SECRET_MAX];
> > > -   char csums_alg[SHARED_SECRET_MAX];
> > > +   struct_group(algs,
> > > +           char verify_alg[SHARED_SECRET_MAX];
> > > +           char csums_alg[SHARED_SECRET_MAX];
> > > +   );
> > >     u32 c_plan_ahead;
> > >     u32 c_delay_target;
> > >     u32 c_fill_target;
> > > diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
> > > index 1f740e42e457..6df2539e215b 100644
> > > --- a/drivers/block/drbd/drbd_receiver.c
> > > +++ b/drivers/block/drbd/drbd_receiver.c
> > > @@ -3921,7 +3921,8 @@ static int receive_SyncParam(struct drbd_connection *connection, struct packet_i
> > >     /* initialize verify_alg and csums_alg */
> > >     p = pi->data;
> > > -   memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX);
> > > +   BUILD_BUG_ON(sizeof(p->algs) != 2 * SHARED_SECRET_MAX);
> > > +   memset(&p->algs, 0, sizeof(p->algs));
> >
> > Using struct_group() introduces complexity. Has it been considered not to
> > modify struct p_rs_param_95 and instead to use two memset() calls instead of
> > one (one memset() call per member)?
>
> I went this direction because using two memset()s (or memcpy()s in other
> patches) changes the machine code. It's not much of a change, but it
> seems easier to justify "no binary changes" via the use of struct_group().
>
> If splitting the memset() is preferred, I can totally do that instead.
> :)

I'm not sure that compilers can fold memsets of adjacent members. It
might not matter, but you could wrap these members in a _named_ struct
then simply use assignment for optimal codegen.

[48/64] drbd: Use struct_group() to zero algs

Checks

Commit Message

Comments

Patch