linux-kernel - Re: [PATCH 2/6] staging: media: intel-ipu3: preferred __aligned(size) over __attribute_

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 13 Apr 2021 12:56:21 +0300
From:   "sakari.ailus@...ux.intel.com" <sakari.ailus@...ux.intel.com>
To:     David Laight <David.Laight@...LAB.COM>
Cc:     'Mitali Borkar' <mitaliborkar810@...il.com>,
        "bingbu.cao@...el.com" <bingbu.cao@...el.com>,
        "tian.shu.qiu@...el.com" <tian.shu.qiu@...el.com>,
        "mchehab@...nel.org" <mchehab@...nel.org>,
        "gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
        "linux-media@...r.kernel.org" <linux-media@...r.kernel.org>,
        "linux-staging@...ts.linux.dev" <linux-staging@...ts.linux.dev>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "outreachy-kernel@...glegroups.com" 
        <outreachy-kernel@...glegroups.com>,
        "mitali_s@...iitr.ac.in" <mitali_s@...iitr.ac.in>
Subject: Re: [PATCH 2/6] staging: media: intel-ipu3: preferred
 __aligned(size) over __attribute__aligned(size)

Hi David,

On Tue, Apr 13, 2021 at 07:40:12AM +0000, David Laight wrote:
> From: Mitali Borkar
> > Sent: 12 April 2021 00:09
> > 
> > This patch fixes the warning identified by checkpatch.pl by replacing
> > __attribute__aligned(size) with __aligned(size)
> > 
> > Signed-off-by: Mitali Borkar <mitaliborkar810@...il.com>
> > ---
> >  .../staging/media/ipu3/include/intel-ipu3.h   | 74 +++++++++----------
> >  1 file changed, 37 insertions(+), 37 deletions(-)
> > 
> > diff --git a/drivers/staging/media/ipu3/include/intel-ipu3.h
> > b/drivers/staging/media/ipu3/include/intel-ipu3.h
> > index 589d5ccee3a7..d95ca9ebfafb 100644
> > --- a/drivers/staging/media/ipu3/include/intel-ipu3.h
> > +++ b/drivers/staging/media/ipu3/include/intel-ipu3.h
> > @@ -84,7 +84,7 @@ struct ipu3_uapi_grid_config {
> >   */
> >  struct ipu3_uapi_awb_raw_buffer {
> >  	__u8 meta_data[IPU3_UAPI_AWB_MAX_BUFFER_SIZE]
> > -		__attribute__((aligned(32)));
> > +		__aligned(32);
> >  } __packed;
> 
> WTF?
> 
> It either has 1-byte alignment because it is just __u8,
> 32-byte because of the aligned(32),
> or 1 byte because of the outer packed.
> 
> What alignment does this (and all the other) structures
> actually need?

32 as noted above. Here packed makes no difference though.

Some of these structs are used embedded in other structs or alone. I
haven't checked this one.

It's also possible to have __packed and __aligned() conflict (in which case
a decent compiler would give you a warning) --- which does not happen
currently AFAIK.

> 
> Specifying 'packed' isn't free.

It may be free if the packed alignment of the fields corresponds to
architecture's packing. Here __aligned() is used to satisfy
hardware alignment requirements and __packed is used to ensure the same
memory layout independently of ABI rules.

-- 
Regards,

Sakari Ailus