2019-10-14 16:18:07

by Tuowen Zhao

[permalink] [raw]
Subject: [PATCH v4 1/2] lib: devres: add a helper function for ioremap_uc

Implement a resource managed strongly uncachable ioremap function.

Cc: <[email protected]>
Tested-by: AceLan Kao <[email protected]>
Signed-off-by: Tuowen Zhao <[email protected]>
Acked-by: Mika Westerberg <[email protected]>
Acked-by: Andy Shevchenko <[email protected]>
Acked-by: Luis Chamberlain <[email protected]>
Acked-for-MFD-by: Lee Jones <[email protected]>
---
Changes from previous version:

* Add Cc stable

.../driver-api/driver-model/devres.rst | 1 +
include/linux/io.h | 2 ++
lib/devres.c | 19 +++++++++++++++++++
3 files changed, 22 insertions(+)

diff --git a/Documentation/driver-api/driver-model/devres.rst b/Documentation/driver-api/driver-model/devres.rst
index a100bef54952..92628fdc2f11 100644
--- a/Documentation/driver-api/driver-model/devres.rst
+++ b/Documentation/driver-api/driver-model/devres.rst
@@ -314,6 +314,7 @@ IOMAP
devm_ioport_unmap()
devm_ioremap()
devm_ioremap_nocache()
+ devm_ioremap_uc()
devm_ioremap_wc()
devm_ioremap_resource() : checks resource, requests memory region, ioremaps
devm_iounmap()
diff --git a/include/linux/io.h b/include/linux/io.h
index accac822336a..a59834bc0a11 100644
--- a/include/linux/io.h
+++ b/include/linux/io.h
@@ -64,6 +64,8 @@ static inline void devm_ioport_unmap(struct device *dev, void __iomem *addr)

void __iomem *devm_ioremap(struct device *dev, resource_size_t offset,
resource_size_t size);
+void __iomem *devm_ioremap_uc(struct device *dev, resource_size_t offset,
+ resource_size_t size);
void __iomem *devm_ioremap_nocache(struct device *dev, resource_size_t offset,
resource_size_t size);
void __iomem *devm_ioremap_wc(struct device *dev, resource_size_t offset,
diff --git a/lib/devres.c b/lib/devres.c
index 6a0e9bd6524a..17624d35e82d 100644
--- a/lib/devres.c
+++ b/lib/devres.c
@@ -9,6 +9,7 @@
enum devm_ioremap_type {
DEVM_IOREMAP = 0,
DEVM_IOREMAP_NC,
+ DEVM_IOREMAP_UC,
DEVM_IOREMAP_WC,
};

@@ -39,6 +40,9 @@ static void __iomem *__devm_ioremap(struct device *dev, resource_size_t offset,
case DEVM_IOREMAP_NC:
addr = ioremap_nocache(offset, size);
break;
+ case DEVM_IOREMAP_UC:
+ addr = ioremap_uc(offset, size);
+ break;
case DEVM_IOREMAP_WC:
addr = ioremap_wc(offset, size);
break;
@@ -68,6 +72,21 @@ void __iomem *devm_ioremap(struct device *dev, resource_size_t offset,
}
EXPORT_SYMBOL(devm_ioremap);

+/**
+ * devm_ioremap_uc - Managed ioremap_uc()
+ * @dev: Generic device to remap IO address for
+ * @offset: Resource address to map
+ * @size: Size of map
+ *
+ * Managed ioremap_uc(). Map is automatically unmapped on driver detach.
+ */
+void __iomem *devm_ioremap_uc(struct device *dev, resource_size_t offset,
+ resource_size_t size)
+{
+ return __devm_ioremap(dev, offset, size, DEVM_IOREMAP_UC);
+}
+EXPORT_SYMBOL_GPL(devm_ioremap_uc);
+
/**
* devm_ioremap_nocache - Managed ioremap_nocache()
* @dev: Generic device to remap IO address for
--
2.23.0


2019-10-14 16:21:26

by Tuowen Zhao

[permalink] [raw]
Subject: [PATCH v4 2/2] mfd: intel-lpss: use devm_ioremap_uc for MMIO

Some BIOS erroneously specifies write-combining BAR for intel-lpss-pci
in MTRR. This will cause the system to hang during boot. If possible,
this bug could be corrected with a firmware update.

This patch use devm_ioremap_uc to overwrite/ignore the MTRR settings
by forcing the use of strongly uncachable pages for intel-lpss.

The BIOS bug is present on Dell XPS 13 7390 2-in-1:

[ 0.001734] 5 base 4000000000 mask 6000000000 write-combining

4000000000-7fffffffff : PCI Bus 0000:00
4000000000-400fffffff : 0000:00:02.0 (i915)
4010000000-4010000fff : 0000:00:15.0 (intel-lpss-pci)

Link: https://bugzilla.kernel.org/show_bug.cgi?id=203485
Cc: <[email protected]>
Tested-by: AceLan Kao <[email protected]>
Signed-off-by: Tuowen Zhao <[email protected]>
Acked-by: Mika Westerberg <[email protected]>
Acked-by: Andy Shevchenko <[email protected]>
Acked-for-MFD-by: Lee Jones <[email protected]>
---
drivers/mfd/intel-lpss.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/mfd/intel-lpss.c b/drivers/mfd/intel-lpss.c
index bfe4ff337581..b0f0781a6b9c 100644
--- a/drivers/mfd/intel-lpss.c
+++ b/drivers/mfd/intel-lpss.c
@@ -384,7 +384,7 @@ int intel_lpss_probe(struct device *dev,
if (!lpss)
return -ENOMEM;

- lpss->priv = devm_ioremap(dev, info->mem->start + LPSS_PRIV_OFFSET,
+ lpss->priv = devm_ioremap_uc(dev, info->mem->start + LPSS_PRIV_OFFSET,
LPSS_PRIV_SIZE);
if (!lpss->priv)
return -ENOMEM;
--
2.23.0

2019-10-14 23:40:39

by kernel test robot

[permalink] [raw]
Subject: Re: [PATCH v4 1/2] lib: devres: add a helper function for ioremap_uc

Hi Tuowen,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on linus/master]
[cannot apply to v5.4-rc3 next-20191014]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system. BTW, we also suggest to use '--base' option to specify the
base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url: https://github.com/0day-ci/linux/commits/Tuowen-Zhao/lib-devres-add-a-helper-function-for-ioremap_uc/20191015-000416
config: sparc64-allmodconfig (attached as .config)
compiler: sparc64-linux-gcc (GCC) 7.4.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
GCC_VERSION=7.4.0 make.cross ARCH=sparc64

If you fix the issue, kindly add following tag
Reported-by: kbuild test robot <[email protected]>

All error/warnings (new ones prefixed by >>):

lib/devres.c: In function '__devm_ioremap':
>> lib/devres.c:44:10: error: implicit declaration of function 'ioremap_uc'; did you mean 'ioremap_wc'? [-Werror=implicit-function-declaration]
addr = ioremap_uc(offset, size);
^~~~~~~~~~
ioremap_wc
>> lib/devres.c:44:8: warning: assignment makes pointer from integer without a cast [-Wint-conversion]
addr = ioremap_uc(offset, size);
^
cc1: some warnings being treated as errors

vim +44 lib/devres.c

25
26 static void __iomem *__devm_ioremap(struct device *dev, resource_size_t offset,
27 resource_size_t size,
28 enum devm_ioremap_type type)
29 {
30 void __iomem **ptr, *addr = NULL;
31
32 ptr = devres_alloc(devm_ioremap_release, sizeof(*ptr), GFP_KERNEL);
33 if (!ptr)
34 return NULL;
35
36 switch (type) {
37 case DEVM_IOREMAP:
38 addr = ioremap(offset, size);
39 break;
40 case DEVM_IOREMAP_NC:
41 addr = ioremap_nocache(offset, size);
42 break;
43 case DEVM_IOREMAP_UC:
> 44 addr = ioremap_uc(offset, size);
45 break;
46 case DEVM_IOREMAP_WC:
47 addr = ioremap_wc(offset, size);
48 break;
49 }
50
51 if (addr) {
52 *ptr = addr;
53 devres_add(dev, ptr);
54 } else
55 devres_free(ptr);
56
57 return addr;
58 }
59

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation


Attachments:
(No filename) (2.59 kB)
.config.gz (57.70 kB)
Download all attachments

2019-10-14 23:50:12

by Tuowen Zhao

[permalink] [raw]
Subject: Re: [PATCH v4 1/2] lib: devres: add a helper function for ioremap_uc

On Tue, 2019-10-15 at 02:46 +0800, kbuild test robot wrote:
> -O ~/bin/make.cross
> chmod +x ~/bin/make.cross
> # save the attached .config to linux build tree
> GCC_VERSION=7.4.0 make.cross ARCH=sparc64

Oops, I'm not sure how would we best fix this. Clearly the patch is not
intended for sparc64. Maybe adding devm_ioremap_uc is rather not safe
right now.

Although, We could declare dummies for these architectures like it has
been for powerpc.

I just noticed another driver having this issue, and fixed with direct
calls to ioremap_uc().

3cc2dac5be3f2: drivers/video/fbdev/atyfb: Replace MTRR UC hole with
strong UC

Tuowen

2019-10-15 07:46:06

by Andy Shevchenko

[permalink] [raw]
Subject: Re: [PATCH v4 1/2] lib: devres: add a helper function for ioremap_uc

On Mon, Oct 14, 2019 at 01:15:53PM -0600, Tuowen Zhao wrote:
> On Tue, 2019-10-15 at 02:46 +0800, kbuild test robot wrote:
> > -O ~/bin/make.cross
> > chmod +x ~/bin/make.cross
> > # save the attached .config to linux build tree
> > GCC_VERSION=7.4.0 make.cross ARCH=sparc64
>
> Oops, I'm not sure how would we best fix this. Clearly the patch is not
> intended for sparc64. Maybe adding devm_ioremap_uc is rather not safe
> right now.

It seems you need a preparatory patch to satisfy sparc64.

> Although, We could declare dummies for these architectures like it has
> been for powerpc.
>
> I just noticed another driver having this issue, and fixed with direct
> calls to ioremap_uc().
>
> 3cc2dac5be3f2: drivers/video/fbdev/atyfb: Replace MTRR UC hole with
> strong UC

That's why I asked Luis to have a look at this :)

--
With Best Regards,
Andy Shevchenko


2019-10-16 16:01:01

by Luis Chamberlain

[permalink] [raw]
Subject: Re: [PATCH v4 1/2] lib: devres: add a helper function for ioremap_uc

On Tue, Oct 15, 2019 at 10:44:34AM +0300, Andy Shevchenko wrote:
> On Mon, Oct 14, 2019 at 01:15:53PM -0600, Tuowen Zhao wrote:
> > On Tue, 2019-10-15 at 02:46 +0800, kbuild test robot wrote:
> > > -O ~/bin/make.cross
> > > chmod +x ~/bin/make.cross
> > > # save the attached .config to linux build tree
> > > GCC_VERSION=7.4.0 make.cross ARCH=sparc64
> >
> > Oops, I'm not sure how would we best fix this. Clearly the patch is not
> > intended for sparc64. Maybe adding devm_ioremap_uc is rather not safe
> > right now.
>
> It seems you need a preparatory patch to satisfy sparc64.

Indeed, can you add that? If you are not comfortable the way to leave
behind lazy architectures is the HAS_FOO feature and then have your
driver require that or depend on the archs that support this. This
allows non-lazy architecturess to move forward with life.

Luis

2019-10-17 12:50:05

by Tuowen Zhao

[permalink] [raw]
Subject: Re: [PATCH v4 1/2] lib: devres: add a helper function for ioremap_uc

On Wed, 2019-10-16 at 12:56 +0000, Luis Chamberlain wrote:
> Indeed, can you add that? If you are not comfortable the way to leave
> behind lazy architectures is the HAS_FOO feature and then have your
> driver require that or depend on the archs that support this. This
> allows non-lazy architecturess to move forward with life.
>
> Luis

Upon close examination, the issue seems easy to fix. Going to submit a
new set shortly.

sparc64 and hexagon don't have ioremap_uc defined. Hexagon also doesn't
have ioremap_wc but didn't report an issue before so devm_ioremap won't
have an problem.

Interestingly tho, majority of the archs include <asm-generic/io.h>,
thus having prototypes. These two archs and a few others don't. I'm
wondering if including the prototypes is actually the recommended
practice.

Tuowen

2019-12-04 11:33:43

by Guenter Roeck

[permalink] [raw]
Subject: Re: [PATCH v4 1/2] lib: devres: add a helper function for ioremap_uc

On Mon, Oct 14, 2019 at 09:33:43AM -0600, Tuowen Zhao wrote:
> Implement a resource managed strongly uncachable ioremap function.
>
> Cc: <[email protected]>

Really ?

> Tested-by: AceLan Kao <[email protected]>
> Signed-off-by: Tuowen Zhao <[email protected]>
> Acked-by: Mika Westerberg <[email protected]>
> Acked-by: Andy Shevchenko <[email protected]>
> Acked-by: Luis Chamberlain <[email protected]>
> Acked-for-MFD-by: Lee Jones <[email protected]>

This patch results in hard build failures when building hexagon images.

lib/devres.c:44:3: error: implicit declaration of function 'ioremap_uc'

Guenter