diff mbox series

[3/3] kbuild: introduce KERNEL_VERSION2 and LINUX_VERSION_CODE2

Message ID 20210206035033.2036180-3-sashal@kernel.org (mailing list archive)
State New
Headers show
Series [1/3] Revert "kbuild: give the SUBLEVEL more room in KERNEL_VERSION" | expand

Commit Message

Sasha Levin Feb. 6, 2021, 3:50 a.m. UTC
SUBLEVEL only has 8 bits of space, which means that we'll overflow it
once it reaches 256.

Few of the stable branches will imminently overflow SUBLEVEL while
there's no risk of overflowing VERSION.

Thus, give SUBLEVEL 8 more bits which will be stolen from VERSION, this
should create a better balance between the different version numbers we
use.

We can't however use the original KERNEL_VERSION and LINUX_VERSION_CODE
as userspace has created ABI dependency on their structure, and we risk
breaking this userspace by modifying the layout of the version integers.

Cc: stable@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 Makefile                                       | 8 +++++++-
 drivers/net/ethernet/mellanox/mlx5/core/main.c | 4 ++--
 drivers/usb/core/hcd.c                         | 4 ++--
 drivers/usb/gadget/udc/aspeed-vhub/hub.c       | 4 ++--
 include/linux/usb/composite.h                  | 4 ++--
 kernel/sys.c                                   | 2 +-
 tools/perf/tests/bpf-script-example.c          | 2 +-
 tools/perf/tests/bpf-script-test-kbuild.c      | 2 +-
 tools/perf/tests/bpf-script-test-prologue.c    | 2 +-
 9 files changed, 19 insertions(+), 13 deletions(-)

Comments

Sasha Levin Feb. 6, 2021, 4:01 a.m. UTC | #1
On Fri, Feb 05, 2021 at 10:50:33PM -0500, Sasha Levin wrote:
>SUBLEVEL only has 8 bits of space, which means that we'll overflow it
>once it reaches 256.
>
>Few of the stable branches will imminently overflow SUBLEVEL while
>there's no risk of overflowing VERSION.
>
>Thus, give SUBLEVEL 8 more bits which will be stolen from VERSION, this
>should create a better balance between the different version numbers we
>use.
>
>We can't however use the original KERNEL_VERSION and LINUX_VERSION_CODE
>as userspace has created ABI dependency on their structure, and we risk
>breaking this userspace by modifying the layout of the version integers.
>
>Cc: stable@kernel.org
>Signed-off-by: Sasha Levin <sashal@kernel.org>
>Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

I wanted to re-use an older commit but forgot to drop the two tags
above. The tags from Masahiro and Greg shouldn't be here, sorry about
that.
Masahiro Yamada Feb. 6, 2021, 12:45 p.m. UTC | #2
On Sat, Feb 6, 2021 at 12:50 PM Sasha Levin <sashal@kernel.org> wrote:
>
> SUBLEVEL only has 8 bits of space, which means that we'll overflow it
> once it reaches 256.
>
> Few of the stable branches will imminently overflow SUBLEVEL while
> there's no risk of overflowing VERSION.
>
> Thus, give SUBLEVEL 8 more bits which will be stolen from VERSION, this
> should create a better balance between the different version numbers we
> use.
>
> We can't however use the original KERNEL_VERSION and LINUX_VERSION_CODE
> as userspace has created ABI dependency on their structure, and we risk
> breaking this userspace by modifying the layout of the version integers.
>
> Cc: stable@kernel.org
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
>  Makefile                                       | 8 +++++++-
>  drivers/net/ethernet/mellanox/mlx5/core/main.c | 4 ++--
>  drivers/usb/core/hcd.c                         | 4 ++--
>  drivers/usb/gadget/udc/aspeed-vhub/hub.c       | 4 ++--
>  include/linux/usb/composite.h                  | 4 ++--
>  kernel/sys.c                                   | 2 +-
>  tools/perf/tests/bpf-script-example.c          | 2 +-
>  tools/perf/tests/bpf-script-test-kbuild.c      | 2 +-
>  tools/perf/tests/bpf-script-test-prologue.c    | 2 +-
>  9 files changed, 19 insertions(+), 13 deletions(-)
>
> diff --git a/Makefile b/Makefile
> index 157be50c691e5..2177c548e4c24 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -1266,7 +1266,13 @@ define filechk_version.h
>                 expr $(VERSION) \* 65536 + 0$(PATCHLEVEL) \* 256 + $(SUBLEVEL)); \
>         fi;                                                              \
>         echo '#define KERNEL_VERSION(a,b,c) (((a) << 16) + ((b) << 8) +  \
> -       ((c) > 255 ? 255 : (c)))'
> +       ((c) > 255 ? 255 : (c)))';                                       \
> +       echo \#define LINUX_VERSION_CODE2 $(shell                        \
> +       expr $(VERSION) \* 16777216 + 0$(PATCHLEVEL) \* 65536 + 0$(SUBLEVEL)); \


Is this needed?
No one in this patch uses LINUX_VERSION_CODE2.

The in-kernel code can use
LINUX_VERSION_MAJOR/PATCHLEVEL/SUBLEVEL directly.


Userspace does not need LINUX_VERSION_CODE2 either.

Your comment in 2/3 states that
userspace only decodes MAJOR and PATCHLEVEL fields,
so LINUX_VERSION_CODE will still do well.




> index ab4b98b3165db..a56bf381335e9 100644
> --- a/tools/perf/tests/bpf-script-example.c
> +++ b/tools/perf/tests/bpf-script-example.c
> @@ -5,7 +5,7 @@
>   */
>  #ifndef LINUX_VERSION_CODE
>  # error Need LINUX_VERSION_CODE
> -# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x40200" into llvm section of ~/.perfconfig'
> +# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x4020000" into llvm section of ~/.perfconfig'


Is this change needed?




>  #endif
>  #define BPF_ANY 0
>  #define BPF_MAP_TYPE_ARRAY 2
> diff --git a/tools/perf/tests/bpf-script-test-kbuild.c b/tools/perf/tests/bpf-script-test-kbuild.c
> index 219673aa278fb..21663295d5b5a 100644
> --- a/tools/perf/tests/bpf-script-test-kbuild.c
> +++ b/tools/perf/tests/bpf-script-test-kbuild.c
> @@ -5,7 +5,7 @@
>   */
>  #ifndef LINUX_VERSION_CODE
>  # error Need LINUX_VERSION_CODE
> -# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x40200" into llvm section of ~/.perfconfig'
> +# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x4020000" into llvm section of ~/.perfconfig'
>  #endif
>  #define SEC(NAME) __attribute__((section(NAME), used))
>
> diff --git a/tools/perf/tests/bpf-script-test-prologue.c b/tools/perf/tests/bpf-script-test-prologue.c
> index bd83d364cf30d..8db19e70813cc 100644
> --- a/tools/perf/tests/bpf-script-test-prologue.c
> +++ b/tools/perf/tests/bpf-script-test-prologue.c
> @@ -5,7 +5,7 @@
>   */
>  #ifndef LINUX_VERSION_CODE
>  # error Need LINUX_VERSION_CODE
> -# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x40200" into llvm section of ~/.perfconfig'
> +# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x4020000" into llvm section of ~/.perfconfig'
>  #endif
>  #define SEC(NAME) __attribute__((section(NAME), used))
>
> --
> 2.27.0
>
Greg Kroah-Hartman Feb. 6, 2021, 12:53 p.m. UTC | #3
On Fri, Feb 05, 2021 at 10:50:33PM -0500, Sasha Levin wrote:
> SUBLEVEL only has 8 bits of space, which means that we'll overflow it
> once it reaches 256.
> 
> Few of the stable branches will imminently overflow SUBLEVEL while
> there's no risk of overflowing VERSION.
> 
> Thus, give SUBLEVEL 8 more bits which will be stolen from VERSION, this
> should create a better balance between the different version numbers we
> use.
> 
> We can't however use the original KERNEL_VERSION and LINUX_VERSION_CODE
> as userspace has created ABI dependency on their structure, and we risk
> breaking this userspace by modifying the layout of the version integers.
> 
> Cc: stable@kernel.org
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
> Signed-off-by: Sasha Levin <sashal@kernel.org>

Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Greg Kroah-Hartman Feb. 6, 2021, 12:54 p.m. UTC | #4
On Fri, Feb 05, 2021 at 10:50:33PM -0500, Sasha Levin wrote:
> SUBLEVEL only has 8 bits of space, which means that we'll overflow it
> once it reaches 256.
> 
> Few of the stable branches will imminently overflow SUBLEVEL while
> there's no risk of overflowing VERSION.
> 
> Thus, give SUBLEVEL 8 more bits which will be stolen from VERSION, this
> should create a better balance between the different version numbers we
> use.
> 
> We can't however use the original KERNEL_VERSION and LINUX_VERSION_CODE
> as userspace has created ABI dependency on their structure, and we risk
> breaking this userspace by modifying the layout of the version integers.
> 
> Cc: stable@kernel.org
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
>  Makefile                                       | 8 +++++++-
>  drivers/net/ethernet/mellanox/mlx5/core/main.c | 4 ++--
>  drivers/usb/core/hcd.c                         | 4 ++--
>  drivers/usb/gadget/udc/aspeed-vhub/hub.c       | 4 ++--
>  include/linux/usb/composite.h                  | 4 ++--
>  kernel/sys.c                                   | 2 +-
>  tools/perf/tests/bpf-script-example.c          | 2 +-
>  tools/perf/tests/bpf-script-test-kbuild.c      | 2 +-
>  tools/perf/tests/bpf-script-test-prologue.c    | 2 +-
>  9 files changed, 19 insertions(+), 13 deletions(-)
> 
> diff --git a/Makefile b/Makefile
> index 157be50c691e5..2177c548e4c24 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -1266,7 +1266,13 @@ define filechk_version.h
>  		expr $(VERSION) \* 65536 + 0$(PATCHLEVEL) \* 256 + $(SUBLEVEL)); \
>  	fi;                                                              \
>  	echo '#define KERNEL_VERSION(a,b,c) (((a) << 16) + ((b) << 8) +  \
> -	((c) > 255 ? 255 : (c)))'
> +	((c) > 255 ? 255 : (c)))';                                       \
> +	echo \#define LINUX_VERSION_CODE2 $(shell                        \

Wait, no, who will use LINUX_VERSION_CODE2()?  Why export something like
this to userspace that no one is asking for?

> +	expr $(VERSION) \* 16777216 + 0$(PATCHLEVEL) \* 65536 + 0$(SUBLEVEL)); \
> +	echo \#define LINUX_VERSION_MAJOR $(VERSION);                    \
> +	echo \#define LINUX_VERSION_PATCHLEVEL $(PATCHLEVEL);            \
> +	echo \#define LINUX_VERSION_SUBLEVEL $(SUBLEVEL);                \

These are good to have, and clean up kernel code.

thanks,

greg k-h
Sasha Levin Feb. 6, 2021, 2:31 p.m. UTC | #5
On Sat, Feb 06, 2021 at 09:45:44PM +0900, Masahiro Yamada wrote:
>On Sat, Feb 6, 2021 at 12:50 PM Sasha Levin <sashal@kernel.org> wrote:
>> -       ((c) > 255 ? 255 : (c)))'
>> +       ((c) > 255 ? 255 : (c)))';                                       \
>> +       echo \#define LINUX_VERSION_CODE2 $(shell                        \
>> +       expr $(VERSION) \* 16777216 + 0$(PATCHLEVEL) \* 65536 + 0$(SUBLEVEL)); \
>
>
>Is this needed?
>No one in this patch uses LINUX_VERSION_CODE2.
>
>The in-kernel code can use
>LINUX_VERSION_MAJOR/PATCHLEVEL/SUBLEVEL directly.
>
>
>Userspace does not need LINUX_VERSION_CODE2 either.
>
>Your comment in 2/3 states that
>userspace only decodes MAJOR and PATCHLEVEL fields,
>so LINUX_VERSION_CODE will still do well.

Yes, that's fair. I wanted to have that to make it easy for out of tree
code to handle in-kernel ABI changes but as we don't care about such
code there's no strict need for LINUX_VERSION_CODE2. I'll drop it.

>> index ab4b98b3165db..a56bf381335e9 100644
>> --- a/tools/perf/tests/bpf-script-example.c
>> +++ b/tools/perf/tests/bpf-script-example.c
>> @@ -5,7 +5,7 @@
>>   */
>>  #ifndef LINUX_VERSION_CODE
>>  # error Need LINUX_VERSION_CODE
>> -# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x40200" into llvm section of ~/.perfconfig'
>> +# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x4020000" into llvm section of ~/.perfconfig'
>
>
>Is this change needed?

Good point, no. I'll resend.
diff mbox series

Patch

diff --git a/Makefile b/Makefile
index 157be50c691e5..2177c548e4c24 100644
--- a/Makefile
+++ b/Makefile
@@ -1266,7 +1266,13 @@  define filechk_version.h
 		expr $(VERSION) \* 65536 + 0$(PATCHLEVEL) \* 256 + $(SUBLEVEL)); \
 	fi;                                                              \
 	echo '#define KERNEL_VERSION(a,b,c) (((a) << 16) + ((b) << 8) +  \
-	((c) > 255 ? 255 : (c)))'
+	((c) > 255 ? 255 : (c)))';                                       \
+	echo \#define LINUX_VERSION_CODE2 $(shell                        \
+	expr $(VERSION) \* 16777216 + 0$(PATCHLEVEL) \* 65536 + 0$(SUBLEVEL)); \
+	echo \#define LINUX_VERSION_MAJOR $(VERSION);                    \
+	echo \#define LINUX_VERSION_PATCHLEVEL $(PATCHLEVEL);            \
+	echo \#define LINUX_VERSION_SUBLEVEL $(SUBLEVEL);                \
+	echo '#define KERNEL_VERSION2(a,b,c) (((a) << 24) + ((b) << 16) + (c))'
 endef
 
 $(version_h): FORCE
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index e4c9627485aa5..989f15d9aa7d4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -237,8 +237,8 @@  static void mlx5_set_driver_version(struct mlx5_core_dev *dev)
 	remaining_size = max_t(int, 0, driver_ver_sz - strlen(string));
 
 	snprintf(string + strlen(string), remaining_size, "%u.%u.%u",
-		 (u8)((LINUX_VERSION_CODE >> 16) & 0xff), (u8)((LINUX_VERSION_CODE >> 8) & 0xff),
-		 (u16)(LINUX_VERSION_CODE & 0xffff));
+		(u8)(LINUX_VERSION_MAJOR), (u8)(LINUX_VERSION_PATCHLEVEL),
+		(u16)(LINUX_VERSION_SUBLEVEL));
 
 	/*Send the command*/
 	MLX5_SET(set_driver_version_in, in, opcode,
diff --git a/drivers/usb/core/hcd.c b/drivers/usb/core/hcd.c
index ad5a0f405a75c..3f0381344221e 100644
--- a/drivers/usb/core/hcd.c
+++ b/drivers/usb/core/hcd.c
@@ -111,8 +111,8 @@  DECLARE_WAIT_QUEUE_HEAD(usb_kill_urb_queue);
  */
 
 /*-------------------------------------------------------------------------*/
-#define KERNEL_REL	bin2bcd(((LINUX_VERSION_CODE >> 16) & 0x0ff))
-#define KERNEL_VER	bin2bcd(((LINUX_VERSION_CODE >> 8) & 0x0ff))
+#define KERNEL_REL	bin2bcd(LINUX_VERSION_MAJOR)
+#define KERNEL_VER	bin2bcd(LINUX_VERSION_PATCHLEVEL)
 
 /* usb 3.1 root hub device descriptor */
 static const u8 usb31_rh_dev_descriptor[18] = {
diff --git a/drivers/usb/gadget/udc/aspeed-vhub/hub.c b/drivers/usb/gadget/udc/aspeed-vhub/hub.c
index bfd8e77788e29..5c7dea5e0ff16 100644
--- a/drivers/usb/gadget/udc/aspeed-vhub/hub.c
+++ b/drivers/usb/gadget/udc/aspeed-vhub/hub.c
@@ -46,8 +46,8 @@ 
  *    - Make vid/did overridable
  *    - make it look like usb1 if usb1 mode forced
  */
-#define KERNEL_REL	bin2bcd(((LINUX_VERSION_CODE >> 16) & 0x0ff))
-#define KERNEL_VER	bin2bcd(((LINUX_VERSION_CODE >> 8) & 0x0ff))
+#define KERNEL_REL	bin2bcd(LINUX_VERSION_MAJOR)
+#define KERNEL_VER	bin2bcd(LINUX_VERSION_PATCHLEVEL)
 
 enum {
 	AST_VHUB_STR_INDEX_MAX = 4,
diff --git a/include/linux/usb/composite.h b/include/linux/usb/composite.h
index 5646dad886e61..c71150f2c6390 100644
--- a/include/linux/usb/composite.h
+++ b/include/linux/usb/composite.h
@@ -575,8 +575,8 @@  static inline u16 get_default_bcdDevice(void)
 {
 	u16 bcdDevice;
 
-	bcdDevice = bin2bcd((LINUX_VERSION_CODE >> 16 & 0xff)) << 8;
-	bcdDevice |= bin2bcd((LINUX_VERSION_CODE >> 8 & 0xff));
+	bcdDevice = bin2bcd(LINUX_VERSION_MAJOR) << 8;
+	bcdDevice |= bin2bcd(LINUX_VERSION_PATCHLEVEL);
 	return bcdDevice;
 }
 
diff --git a/kernel/sys.c b/kernel/sys.c
index 8bb46e50f02d4..b09fe21e88ff5 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -1242,7 +1242,7 @@  static int override_release(char __user *release, size_t len)
 				break;
 			rest++;
 		}
-		v = ((LINUX_VERSION_CODE >> 8) & 0xff) + 60;
+		v = LINUX_VERSION_PATCHLEVEL + 60;
 		copy = clamp_t(size_t, len, 1, sizeof(buf));
 		copy = scnprintf(buf, copy, "2.6.%u%s", v, rest);
 		ret = copy_to_user(release, buf, copy + 1);
diff --git a/tools/perf/tests/bpf-script-example.c b/tools/perf/tests/bpf-script-example.c
index ab4b98b3165db..a56bf381335e9 100644
--- a/tools/perf/tests/bpf-script-example.c
+++ b/tools/perf/tests/bpf-script-example.c
@@ -5,7 +5,7 @@ 
  */
 #ifndef LINUX_VERSION_CODE
 # error Need LINUX_VERSION_CODE
-# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x40200" into llvm section of ~/.perfconfig'
+# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x4020000" into llvm section of ~/.perfconfig'
 #endif
 #define BPF_ANY 0
 #define BPF_MAP_TYPE_ARRAY 2
diff --git a/tools/perf/tests/bpf-script-test-kbuild.c b/tools/perf/tests/bpf-script-test-kbuild.c
index 219673aa278fb..21663295d5b5a 100644
--- a/tools/perf/tests/bpf-script-test-kbuild.c
+++ b/tools/perf/tests/bpf-script-test-kbuild.c
@@ -5,7 +5,7 @@ 
  */
 #ifndef LINUX_VERSION_CODE
 # error Need LINUX_VERSION_CODE
-# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x40200" into llvm section of ~/.perfconfig'
+# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x4020000" into llvm section of ~/.perfconfig'
 #endif
 #define SEC(NAME) __attribute__((section(NAME), used))
 
diff --git a/tools/perf/tests/bpf-script-test-prologue.c b/tools/perf/tests/bpf-script-test-prologue.c
index bd83d364cf30d..8db19e70813cc 100644
--- a/tools/perf/tests/bpf-script-test-prologue.c
+++ b/tools/perf/tests/bpf-script-test-prologue.c
@@ -5,7 +5,7 @@ 
  */
 #ifndef LINUX_VERSION_CODE
 # error Need LINUX_VERSION_CODE
-# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x40200" into llvm section of ~/.perfconfig'
+# error Example: for 4.2 kernel, put 'clang-opt="-DLINUX_VERSION_CODE=0x4020000" into llvm section of ~/.perfconfig'
 #endif
 #define SEC(NAME) __attribute__((section(NAME), used))