[PATCH v2 1/3] um: virt-pci: fix 32-bit compile

[PATCH v2 1/3] um: virt-pci: fix 32-bit compile

[Johannes Berg]

From: Johannes Berg <johannes.berg@intel.com>

There were a few 32-bit compile warnings that of course
turned into errors with -Werror, fix the 32-bit build.

Fixes: 68f5d3f3b654 ("um: add PCI over virtio emulation driver")
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
---
 arch/um/drivers/virt-pci.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/arch/um/drivers/virt-pci.c b/arch/um/drivers/virt-pci.c
index c08066633023..0ab58016db22 100644
--- a/arch/um/drivers/virt-pci.c
+++ b/arch/um/drivers/virt-pci.c
@@ -181,15 +181,15 @@ static unsigned long um_pci_cfgspace_read(void *priv, unsigned int offset,
 	/* buf->data is maximum size - we may only use parts of it */
 	struct um_pci_message_buffer *buf;
 	u8 *data;
-	unsigned long ret = ~0ULL;
+	unsigned long ret = ULONG_MAX;
 
 	if (!dev)
-		return ~0ULL;
+		return ULONG_MAX;
 
 	buf = get_cpu_var(um_pci_msg_bufs);
 	data = buf->data;
 
-	memset(data, 0xff, sizeof(data));
+	memset(buf->data, 0xff, sizeof(buf->data));
 
 	switch (size) {
 	case 1:
@@ -304,7 +304,7 @@ static unsigned long um_pci_bar_read(void *priv, unsigned int offset,
 	/* buf->data is maximum size - we may only use parts of it */
 	struct um_pci_message_buffer *buf;
 	u8 *data;
-	unsigned long ret = ~0ULL;
+	unsigned long ret = ULONG_MAX;
 
 	buf = get_cpu_var(um_pci_msg_bufs);
 	data = buf->data;
-- 
2.31.1


_______________________________________________
linux-um mailing list
linux-um@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-um

[PATCH v2 2/3] lib/logic_iomem: fix 32-bit build

[Johannes Berg]

From: Johannes Berg <johannes.berg@intel.com>

On a 32-bit build, the (unsigned long long) casts throw warnings
(or errors) due to being to a different integer size. Cast to
uintptr_t first (with the __force for sparse) and then further
to get the consistent print on 32 and 64-bit.

Fixes: ca2e334232b6 ("lib: add iomem emulation (logic_iomem)")
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
---
 lib/logic_iomem.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/lib/logic_iomem.c b/lib/logic_iomem.c
index 9bdfde0c0f86..54fa601f3300 100644
--- a/lib/logic_iomem.c
+++ b/lib/logic_iomem.c
@@ -79,7 +79,7 @@ static void __iomem *real_ioremap(phys_addr_t offset, size_t size)
 static void real_iounmap(void __iomem *addr)
 {
 	WARN(1, "invalid iounmap for addr 0x%llx\n",
-	     (unsigned long long __force)addr);
+	     (unsigned long long)(uintptr_t __force)addr);
 }
 #endif /* CONFIG_LOGIC_IOMEM_FALLBACK */
 
@@ -173,7 +173,7 @@ EXPORT_SYMBOL(iounmap);
 static u##sz real_raw_read ## op(const volatile void __iomem *addr)	\
 {									\
 	WARN(1, "Invalid read" #op " at address %llx\n",		\
-	     (unsigned long long __force)addr);				\
+	     (unsigned long long)(uintptr_t __force)addr);		\
 	return (u ## sz)~0ULL;						\
 }									\
 									\
@@ -181,7 +181,8 @@ static void real_raw_write ## op(u ## sz val,				\
 				 volatile void __iomem *addr)		\
 {									\
 	WARN(1, "Invalid writeq" #op " of 0x%llx at address %llx\n",	\
-	     (unsigned long long)val, (unsigned long long __force)addr);\
+	     (unsigned long long)val,					\
+	     (unsigned long long)(uintptr_t __force)addr);\
 }									\
 
 MAKE_FALLBACK(b, 8);
@@ -194,14 +195,14 @@ MAKE_FALLBACK(q, 64);
 static void real_memset_io(volatile void __iomem *addr, int value, size_t size)
 {
 	WARN(1, "Invalid memset_io at address 0x%llx\n",
-	     (unsigned long long __force)addr);
+	     (unsigned long long)(uintptr_t __force)addr);
 }
 
 static void real_memcpy_fromio(void *buffer, const volatile void __iomem *addr,
 			       size_t size)
 {
 	WARN(1, "Invalid memcpy_fromio at address 0x%llx\n",
-	     (unsigned long long __force)addr);
+	     (unsigned long long)(uintptr_t __force)addr);
 
 	memset(buffer, 0xff, size);
 }
@@ -210,7 +211,7 @@ static void real_memcpy_toio(volatile void __iomem *addr, const void *buffer,
 			     size_t size)
 {
 	WARN(1, "Invalid memcpy_toio at address 0x%llx\n",
-	     (unsigned long long __force)addr);
+	     (unsigned long long)(uintptr_t __force)addr);
 }
 #endif /* CONFIG_LOGIC_IOMEM_FALLBACK */
 
-- 
2.31.1


_______________________________________________
linux-um mailing list
linux-um@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-um

[PATCH v2 3/3] lib/logic_iomem: fix operation on 32-bit

[Johannes Berg]

From: Johannes Berg <johannes.berg@intel.com>

On 32-bit, the first entry might be at 0/NULL, but that's
strange and leads to issues, e.g. where we check "if (ret)".
Use a IOREMAP_BIAS/IOREMAP_MASK of 0x80000000UL to avoid
this. This then requires reducing the number of areas (via
MAX_AREAS), but we still have 128 areas, which is enough.

Fixes: ca2e334232b6 ("lib: add iomem emulation (logic_iomem)")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
---
 lib/logic_iomem.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/lib/logic_iomem.c b/lib/logic_iomem.c
index 54fa601f3300..549b22d4bcde 100644
--- a/lib/logic_iomem.c
+++ b/lib/logic_iomem.c
@@ -21,15 +21,15 @@ struct logic_iomem_area {
 
 #define AREA_SHIFT	24
 #define MAX_AREA_SIZE	(1 << AREA_SHIFT)
-#define MAX_AREAS	((1ULL<<32) / MAX_AREA_SIZE)
+#define MAX_AREAS	((1U << 31) / MAX_AREA_SIZE)
 #define AREA_BITS	((MAX_AREAS - 1) << AREA_SHIFT)
 #define AREA_MASK	(MAX_AREA_SIZE - 1)
 #ifdef CONFIG_64BIT
 #define IOREMAP_BIAS	0xDEAD000000000000UL
 #define IOREMAP_MASK	0xFFFFFFFF00000000UL
 #else
-#define IOREMAP_BIAS	0
-#define IOREMAP_MASK	0
+#define IOREMAP_BIAS	0x80000000UL
+#define IOREMAP_MASK	0x80000000UL
 #endif
 
 static DEFINE_MUTEX(regions_mtx);
-- 
2.31.1


_______________________________________________
linux-um mailing list
linux-um@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-um

Re: [PATCH v2 1/3] um: virt-pci: fix 32-bit compile

[Geert Uytterhoeven]

Hi Johannes,

On Wed, Sep 15, 2021 at 8:30 PM Johannes Berg <johannes@sipsolutions.net> wrote:
> From: Johannes Berg <johannes.berg@intel.com>
>
> There were a few 32-bit compile warnings that of course
> turned into errors with -Werror, fix the 32-bit build.
>
> Fixes: 68f5d3f3b654 ("um: add PCI over virtio emulation driver")
> Reported-by: Al Viro <viro@zeniv.linux.org.uk>
> Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Thanks for your patch!

> --- a/arch/um/drivers/virt-pci.c
> +++ b/arch/um/drivers/virt-pci.c
> @@ -181,15 +181,15 @@ static unsigned long um_pci_cfgspace_read(void *priv, unsigned int offset,
>         /* buf->data is maximum size - we may only use parts of it */
>         struct um_pci_message_buffer *buf;
>         u8 *data;
> -       unsigned long ret = ~0ULL;
> +       unsigned long ret = ULONG_MAX;

Why not just drop the last "L"? ;-)

>
>         if (!dev)
> -               return ~0ULL;
> +               return ULONG_MAX;
>
>         buf = get_cpu_var(um_pci_msg_bufs);
>         data = buf->data;
>
> -       memset(data, 0xff, sizeof(data));
> +       memset(buf->data, 0xff, sizeof(buf->data));

The first change is not needed.
The second change is a genuine bug fix, also on 64-bit, which should be
submitted separately, and backported to stable.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH v2 1/3] um: virt-pci: fix 32-bit compile

[Johannes Berg]

On Wed, 2021-09-15 at 20:42 +0200, Geert Uytterhoeven wrote:
> 
> > -       unsigned long ret = ~0ULL;
> > +       unsigned long ret = ULONG_MAX;
> 
> Why not just drop the last "L"? ;-)
> 
> > 
> >         if (!dev)
> > -               return ~0ULL;
> > +               return ULONG_MAX;
> 
> The first change is not needed.

True, but it seemed inconsistent without that.


> > -       memset(data, 0xff, sizeof(data));
> > +       memset(buf->data, 0xff, sizeof(buf->data));

> The second change is a genuine bug fix, also on 64-bit, which should be
> submitted separately, and backported to stable.

It's kind of a bug, I agree. However:

On 64-bit platforms, the pointer size is 8 bytes, and thus all the data
is memset to 0xff as needed.

On 32-bit platforms, we cannot do 64-bit reads (there's no ioread64 on
32-bit), and thus we can only ever need four 0xff bytes (for a failing
32-bit read), matching the pointer size, so it ends up being correct as
well.

johannes