diff mbox series

[v2] devres: Really align data field to unsigned long long

Message ID 20180708175621.6951-1-abrodkin@synopsys.com
State New
Headers show
Series [v2] devres: Really align data field to unsigned long long | expand

Commit Message

Alexey Brodkin July 8, 2018, 5:56 p.m. UTC
Depending on ABI "long long" type of a particular 32-bit CPU
might be aligned by either word (32-bits) or double word (64-bits).
Make sure "data" is really 64-bit aligned for any 32-bit CPU.

At least for 32-bit ARC cores ABI requires "long long" types
to be aligned by normal 32-bit word. This makes "data" field aligned to
12 bytes. Which is still OK as long as we use 32-bit data only.

But once we want to use native atomic64_t type (i.e. when we use special
instructions LLOCKD/SCONDD for accessing 64-bit data) we easily hit
misaligned access exception.

That's because even on CPUs capable of non-aligned data access LL/SC
instructions require strict alignment.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
---

Changes v1 -> v2:

 * Reworded commit message
 * Inserted comment right in source [Thomas]

 drivers/base/devres.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

Comments

Greg KH July 8, 2018, 6:40 p.m. UTC | #1
On Sun, Jul 08, 2018 at 08:56:21PM +0300, Alexey Brodkin wrote:
> Depending on ABI "long long" type of a particular 32-bit CPU
> might be aligned by either word (32-bits) or double word (64-bits).
> Make sure "data" is really 64-bit aligned for any 32-bit CPU.
> 
> At least for 32-bit ARC cores ABI requires "long long" types
> to be aligned by normal 32-bit word. This makes "data" field aligned to
> 12 bytes. Which is still OK as long as we use 32-bit data only.
> 
> But once we want to use native atomic64_t type (i.e. when we use special
> instructions LLOCKD/SCONDD for accessing 64-bit data) we easily hit
> misaligned access exception.
> 
> That's because even on CPUs capable of non-aligned data access LL/SC
> instructions require strict alignment.
> 
> Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: stable@vger.kernel.org
> ---
> 
> Changes v1 -> v2:
> 
>  * Reworded commit message
>  * Inserted comment right in source [Thomas]
> 
>  drivers/base/devres.c | 8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)

Always use scripts/get_maintainer.pl to properly cc: the needed
developer/maintainer.  As it is, this patch is going to get dropped on
the floor, sorry...

greg k-h
Alexey Brodkin July 8, 2018, 6:45 p.m. UTC | #2
Hi Greg,

вс, 8 июл. 2018 г. в 21:40, Greg KH <greg@kroah.com>:
>
> On Sun, Jul 08, 2018 at 08:56:21PM +0300, Alexey Brodkin wrote:
> > Depending on ABI "long long" type of a particular 32-bit CPU
> > might be aligned by either word (32-bits) or double word (64-bits).
> > Make sure "data" is really 64-bit aligned for any 32-bit CPU.
> >
> > At least for 32-bit ARC cores ABI requires "long long" types
> > to be aligned by normal 32-bit word. This makes "data" field aligned to
> > 12 bytes. Which is still OK as long as we use 32-bit data only.
> >
> > But once we want to use native atomic64_t type (i.e. when we use special
> > instructions LLOCKD/SCONDD for accessing 64-bit data) we easily hit
> > misaligned access exception.
> >
> > That's because even on CPUs capable of non-aligned data access LL/SC
> > instructions require strict alignment.
> >
> > Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
> > Cc: Thomas Gleixner <tglx@linutronix.de>
> > Cc: stable@vger.kernel.org
> > ---
> >
> > Changes v1 -> v2:
> >
> >  * Reworded commit message
> >  * Inserted comment right in source [Thomas]
> >
> >  drivers/base/devres.c | 8 ++++++--
> >  1 file changed, 6 insertions(+), 2 deletions(-)
>
> Always use scripts/get_maintainer.pl to properly cc: the needed
> developer/maintainer.  As it is, this patch is going to get dropped on
> the floor, sorry...

Right I was way too relaxed dealing with very generic stuff which might get
not that much of attention as more narrow topics or subsystems. But anyways
get_maintainer.pl says you're the guy so do I need to do anything extra still or
it's OK for this time?

-Alexey
Greg KH July 9, 2018, 4:18 a.m. UTC | #3
On Sun, Jul 08, 2018 at 09:45:31PM +0300, Алексей Бродкин wrote:
> Hi Greg,
> 
> вс, 8 июл. 2018 г. в 21:40, Greg KH <greg@kroah.com>:
> >
> > On Sun, Jul 08, 2018 at 08:56:21PM +0300, Alexey Brodkin wrote:
> > > Depending on ABI "long long" type of a particular 32-bit CPU
> > > might be aligned by either word (32-bits) or double word (64-bits).
> > > Make sure "data" is really 64-bit aligned for any 32-bit CPU.
> > >
> > > At least for 32-bit ARC cores ABI requires "long long" types
> > > to be aligned by normal 32-bit word. This makes "data" field aligned to
> > > 12 bytes. Which is still OK as long as we use 32-bit data only.
> > >
> > > But once we want to use native atomic64_t type (i.e. when we use special
> > > instructions LLOCKD/SCONDD for accessing 64-bit data) we easily hit
> > > misaligned access exception.
> > >
> > > That's because even on CPUs capable of non-aligned data access LL/SC
> > > instructions require strict alignment.
> > >
> > > Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
> > > Cc: Thomas Gleixner <tglx@linutronix.de>
> > > Cc: stable@vger.kernel.org
> > > ---
> > >
> > > Changes v1 -> v2:
> > >
> > >  * Reworded commit message
> > >  * Inserted comment right in source [Thomas]
> > >
> > >  drivers/base/devres.c | 8 ++++++--
> > >  1 file changed, 6 insertions(+), 2 deletions(-)
> >
> > Always use scripts/get_maintainer.pl to properly cc: the needed
> > developer/maintainer.  As it is, this patch is going to get dropped on
> > the floor, sorry...
> 
> Right I was way too relaxed dealing with very generic stuff which might get
> not that much of attention as more narrow topics or subsystems. But anyways
> get_maintainer.pl says you're the guy so do I need to do anything extra still or
> it's OK for this time?

Please resend it properly, it is not in my patch queue anywhere...

greg k-h
diff mbox series

Patch

diff --git a/drivers/base/devres.c b/drivers/base/devres.c
index f98a097e73f2..466fa59c866a 100644
--- a/drivers/base/devres.c
+++ b/drivers/base/devres.c
@@ -24,8 +24,12 @@  struct devres_node {
 
 struct devres {
 	struct devres_node		node;
-	/* -- 3 pointers */
-	unsigned long long		data[];	/* guarantee ull alignment */
+	/*
+	 * Depending on ABI "long long" type of a particular 32-bit CPU
+	 * might be aligned by either word (32-bits) or double word (64-bits).
+	 * Make sure "data" is really 64-bit aligned for any 32-bit CPU.
+	 */
+	unsigned long long		data[] __aligned(sizeof(unsigned long long));
 };
 
 struct devres_group {