[PATCH net-next v9 00/10] lib: packing: introduce and use (un)pack_fields

[PATCH net-next v9 00/10] lib: packing: introduce and use (un)pack_fields

[Jacob Keller]

This series improves the packing library with a new API for packing or
unpacking a large number of fields at once with minimal code footprint. The
API is then used to replace bespoke packing logic in the ice driver,
preparing it to handle unpacking in the future. Finally, the ice driver has
a few other cleanups related to the packing logic.

The pack_fields and unpack_fields functions have the following improvements
over the existing pack() and unpack() API:

 1. Packing or unpacking a large number of fields takes significantly less
    code. This significantly reduces the .text size for an increase in the
    .data size which is much smaller.

 2. The unpacked data can be stored in sizes smaller than u64 variables.
    This reduces the storage requirement both for runtime data structures,
    and for the rodata defining the fields. This scales with the number of
    fields used.

 3. Most of the error checking is done at compile time, rather than
    runtime, via CHECK_PACKED_FIELD macros.

The actual packing and unpacking code still uses the u64 size
variables. However, these are converted to the appropriate field sizes when
storing or reading the data from the buffer.

This version now uses significantly improved macro checks, thanks to the
work of Vladimir. We now only need 300 lines of macro for the generated
checks. In addition, each new check only requires 4 lines of code for its
macro implementation and 1 extra line in the CHECK_PACKED_FIELDS macro.
This is significantly better than previous versions which required ~2700
lines.

The CHECK_PACKED_FIELDS macro uses __builtin_choose_expr to select the
appropriately sized CHECK_PACKED_FIELDS_N macro. This enables directly
adding CHECK_PACKED_FIELDS calls into the pack_fields and unpack_fields
macros. Drivers no longer need to call the CHECK_PACKED_FIELDS_N macros
directly, and we do not need to modify Kbuild or introduce multiple CONFIG
options.

The code for the CHECK_PACKED_FIELDS_(0..50) and CHECK_PACKED_FIELDS itself
can be generated from the C program in scripts/gen_packed_field_checks.c.
This little C program may be used in the future to update the checks to
more sizes if a driver with more than 50 fields appears in the future.
The total amount of required code is now much smaller, and we don't
anticipate needing to increase the size very often. Thus, it makes sense to
simply commit the result directly instead of attempting to modify Kbuild to
automatically generate it.

This version uses the 5-argument format of pack_fields and unpack_fields,
with the size of the packed buffer passed as one of the arguments. We do
enforce that the compiler can tell its a constant using
__builtin_constant_p(), ensuring that the size checks are handled at
compile time. We could reduce these to 4 arguments and require that the
passed in pbuf be of a type which has the appropriate size. I opted against
that because it makes the API less flexible and a bit less natural to use
in existing code.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
Changes in v9:
- Use BUILD_BUG_ON_MSG to provide more useful and detailed error messages,
  including the field array name, associated field index values, and the
  actual rule being violated. This improves the usability of the resulting
  error messages, especially for users unfamiliar with the API
  requirements.
- New implementation of CHECK_PACKED_FIELD and CHECK_PACKED_FIELD_OVERLAP,
  taking the reference of the total array field directly. This allows
  tail-calling the CHECK_PACKED_FIELD_OVERLAP from within
  CHECK_PACKED_FIELD, significantly reducing the number of lines required
  to implement all the macros.
- Drop the ARRAY_SIZE checks from the CHECK_PACKED_FIELDS_* macros. These
  were only necessary when users called the macros directly. Now that we
  always use __builtin_choose_expr to determine which one to call, this is
  a waste of the CPU cycles.
- Implement each CHECK_PACKED_FIELD_N recursively, by calling the previous
  CHECK_PACKED_FIELD_* macro. This means each additional macro now only
  needs 4 lines of code, instead of scaling linearly with the size of N.
  This is possible now that we no longer directly check the ARRAY_SIZE in
  each macro.
- Use do {} while(0) for implementing the multiline checks in each
  CHECK_PACKED_FIELD_* macro, instead of statement expressions. This helps
  with GCC giving up on processing due to the multiple layers of statement
  expressions when evaluating the CHECK_PACKED_FIELD_* macros.
- Link to v8: https://lore.kernel.org/r/20241203-packing-pack-fields-and-ice-implementation-v8-0-2ed68edfe583@intel.com

Changes in v8:
- Add my missing SOB on one of the patches
- Remove include/linux/packing_types.h and put the generated code directly
  into include/linux/packing.h
- Split documentation to its own patch, and use the proposed documentation
  from Vladimir
- Link to v7: https://lore.kernel.org/r/20241202-packing-pack-fields-and-ice-implementation-v7-0-ed22e38e6c65@intel.com

Changes in v7:
- Dropped the RFC tag for submission to net-next
- Link to v6: https://lore.kernel.org/r/20241118-packing-pack-fields-and-ice-implementation-v6-0-6af8b658a6c3@intel.com

Changes in v6:
- Revert to macro checks similar to v2.
- Add a __builtin_choose_expr() based macro to automatically select the
  appropriate size macro.
- Keep the pbuflen check separate from the main loop check, similar to v5.
- Link to v5: https://lore.kernel.org/r/20241111-packing-pack-fields-and-ice-implementation-v5-0-80c07349e6b7@intel.com

Changes in v5:
- Fix printf format specifier for the sym->st_size
- Link to v4: https://lore.kernel.org/r/20241108-packing-pack-fields-and-ice-implementation-v4-0-81a9f42c30e5@intel.com

Changes in v4:
- Move the buffer size checks to (un)pack_fields() macros.
- Enforce use of a sized type of the packed buffer, removing the now
  unnecessary pbuflen argument of (un)pack_fields().
- Drop exporting the buffer size to modpost.
- Simplify modpost implementation to directly check each symbol in the
  handle_packed_field_symbol() function. This removes the need for a hash,
  and is ultimately much simpler now that modpost doesn't need the size of
  the target buffer.
- Fix the width check to correctly calculate the width and compare it
  properly.
- Refactor modpost messages to consistently report the module name first,
  the symbol name second, and the field number 3rd.
- Correctly implement overlap checks in the modpost, rather than only
  checking field ordering.
- Link to v3: https://lore.kernel.org/r/20241107-packing-pack-fields-and-ice-implementation-v3-0-27c566ac2436@intel.com

Changes in v3:
- Replace macro-based C pre-processor checks with checks implemented in
  modpost.
- Move structure definitions into  <linux/packing_types.h> to enable reuse
  within modpost.
- Add DECLARE_PACKED_FIELDS_S and DECLARE_PACKED_FIELDS_M to enable
  automatically generating the buffer size constants and the section
  attributes.
- Add additional unit tests for the pack_fields and unpack_fields APIs.
- Update documentation with an explanation of the new API as well as some
  example code.
- Link to v2: https://lore.kernel.org/r/20241025-packing-pack-fields-and-ice-implementation-v2-0-734776c88e40@intel.com

Changes in v2:
- Add my missing sign-off to the first patch
- Update the descriptions for a few patches
- Only generate CHECK_PACKED_FIELDS_N when another module selects it
- Add a new patch introducing wrapper structures for the packed Tx and Rx
  queue context, suggested by Vladimir.
- Drop the now unnecessary macros in ice, thanks to the new types
- Link to v1: https://lore.kernel.org/r/20241011-packing-pack-fields-and-ice-implementation-v1-0-d9b1f7500740@intel.com

---
Jacob Keller (7):
      lib: packing: document recently added APIs
      ice: remove int_q_state from ice_tlan_ctx
      ice: use structures to keep track of queue context size
      ice: use <linux/packing.h> for Tx and Rx queue context data
      ice: reduce size of queue context fields
      ice: move prefetch enable to ice_setup_rx_ctx
      ice: cleanup Rx queue context programming functions

Vladimir Oltean (3):
      lib: packing: create __pack() and __unpack() variants without error checking
      lib: packing: demote truncation error in pack() to a warning in __pack()
      lib: packing: add pack_fields() and unpack_fields()

 Makefile                                        |   4 +
 drivers/net/ethernet/intel/ice/ice_adminq_cmd.h |  11 +-
 drivers/net/ethernet/intel/ice/ice_common.h     |   5 +-
 drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h  |  49 +--
 include/linux/packing.h                         | 427 ++++++++++++++++++++++++
 drivers/net/dsa/sja1105/sja1105_static_config.c |   8 +-
 drivers/net/ethernet/intel/ice/ice_base.c       |   6 +-
 drivers/net/ethernet/intel/ice/ice_common.c     | 293 ++++------------
 lib/packing.c                                   | 285 ++++++++++++----
 lib/packing_test.c                              |  61 ++++
 scripts/gen_packed_field_checks.c               |  37 ++
 Documentation/core-api/packing.rst              | 118 ++++++-
 MAINTAINERS                                     |   1 +
 drivers/net/ethernet/intel/Kconfig              |   1 +
 scripts/Makefile                                |   2 +-
 15 files changed, 948 insertions(+), 360 deletions(-)
---
base-commit: bb18265c3aba92b91a1355609769f3e967b65dee
change-id: 20241004-packing-pack-fields-and-ice-implementation-b17c7ce8e373

Best regards,
-- 
Jacob Keller <jacob.e.keller@intel.com>

[PATCH net-next v9 01/10] lib: packing: create __pack() and __unpack() variants without error checking

[Jacob Keller]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

A future variant of the API, which works on arrays of packed_field
structures, will make most of these checks redundant. The idea will be
that we want to perform sanity checks at compile time, not once
for every function call.

Introduce new variants of pack() and unpack(), which elide the sanity
checks, assuming that the input was pre-sanitized.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
 lib/packing.c | 142 ++++++++++++++++++++++++++++++++--------------------------
 1 file changed, 78 insertions(+), 64 deletions(-)

diff --git a/lib/packing.c b/lib/packing.c
index 793942745e34fde1810010e303742e6484861bc8..f237b8af99f5fa8e839c38126769c50b2bfe6361 100644
--- a/lib/packing.c
+++ b/lib/packing.c
@@ -51,64 +51,20 @@ static size_t calculate_box_addr(size_t box, size_t len, u8 quirks)
 	return offset_of_group + offset_in_group;
 }
 
-/**
- * pack - Pack u64 number into bitfield of buffer.
- *
- * @pbuf: Pointer to a buffer holding the packed value.
- * @uval: CPU-readable unpacked value to pack.
- * @startbit: The index (in logical notation, compensated for quirks) where
- *	      the packed value starts within pbuf. Must be larger than, or
- *	      equal to, endbit.
- * @endbit: The index (in logical notation, compensated for quirks) where
- *	    the packed value ends within pbuf. Must be smaller than, or equal
- *	    to, startbit.
- * @pbuflen: The length in bytes of the packed buffer pointed to by @pbuf.
- * @quirks: A bit mask of QUIRK_LITTLE_ENDIAN, QUIRK_LSW32_IS_FIRST and
- *	    QUIRK_MSB_ON_THE_RIGHT.
- *
- * Return: 0 on success, EINVAL or ERANGE if called incorrectly. Assuming
- *	   correct usage, return code may be discarded. The @pbuf memory will
- *	   be modified on success.
- */
-int pack(void *pbuf, u64 uval, size_t startbit, size_t endbit, size_t pbuflen,
-	 u8 quirks)
+static void __pack(void *pbuf, u64 uval, size_t startbit, size_t endbit,
+		   size_t pbuflen, u8 quirks)
 {
 	/* Logical byte indices corresponding to the
 	 * start and end of the field.
 	 */
-	int plogical_first_u8, plogical_last_u8, box;
-	/* width of the field to access in the pbuf */
-	u64 value_width;
-
-	/* startbit is expected to be larger than endbit, and both are
-	 * expected to be within the logically addressable range of the buffer.
-	 */
-	if (unlikely(startbit < endbit || startbit >= BITS_PER_BYTE * pbuflen))
-		/* Invalid function call */
-		return -EINVAL;
-
-	value_width = startbit - endbit + 1;
-	if (unlikely(value_width > 64))
-		return -ERANGE;
-
-	/* Check if "uval" fits in "value_width" bits.
-	 * If value_width is 64, the check will fail, but any
-	 * 64-bit uval will surely fit.
-	 */
-	if (unlikely(value_width < 64 && uval >= (1ull << value_width)))
-		/* Cannot store "uval" inside "value_width" bits.
-		 * Truncating "uval" is most certainly not desirable,
-		 * so simply erroring out is appropriate.
-		 */
-		return -ERANGE;
+	int plogical_first_u8 = startbit / BITS_PER_BYTE;
+	int plogical_last_u8 = endbit / BITS_PER_BYTE;
+	int box;
 
 	/* Iterate through an idealistic view of the pbuf as an u64 with
 	 * no quirks, u8 by u8 (aligned at u8 boundaries), from high to low
 	 * logical bit significance. "box" denotes the current logical u8.
 	 */
-	plogical_first_u8 = startbit / BITS_PER_BYTE;
-	plogical_last_u8  = endbit / BITS_PER_BYTE;
-
 	for (box = plogical_first_u8; box >= plogical_last_u8; box--) {
 		/* Bit indices into the currently accessed 8-bit box */
 		size_t box_start_bit, box_end_bit, box_addr;
@@ -163,15 +119,13 @@ int pack(void *pbuf, u64 uval, size_t startbit, size_t endbit, size_t pbuflen,
 		((u8 *)pbuf)[box_addr] &= ~box_mask;
 		((u8 *)pbuf)[box_addr] |= pval;
 	}
-	return 0;
 }
-EXPORT_SYMBOL(pack);
 
 /**
- * unpack - Unpack u64 number from packed buffer.
+ * pack - Pack u64 number into bitfield of buffer.
  *
  * @pbuf: Pointer to a buffer holding the packed value.
- * @uval: Pointer to an u64 holding the unpacked value.
+ * @uval: CPU-readable unpacked value to pack.
  * @startbit: The index (in logical notation, compensated for quirks) where
  *	      the packed value starts within pbuf. Must be larger than, or
  *	      equal to, endbit.
@@ -183,16 +137,12 @@ EXPORT_SYMBOL(pack);
  *	    QUIRK_MSB_ON_THE_RIGHT.
  *
  * Return: 0 on success, EINVAL or ERANGE if called incorrectly. Assuming
- *	   correct usage, return code may be discarded. The @uval will be
- *	   modified on success.
+ *	   correct usage, return code may be discarded. The @pbuf memory will
+ *	   be modified on success.
  */
-int unpack(const void *pbuf, u64 *uval, size_t startbit, size_t endbit,
-	   size_t pbuflen, u8 quirks)
+int pack(void *pbuf, u64 uval, size_t startbit, size_t endbit, size_t pbuflen,
+	 u8 quirks)
 {
-	/* Logical byte indices corresponding to the
-	 * start and end of the field.
-	 */
-	int plogical_first_u8, plogical_last_u8, box;
 	/* width of the field to access in the pbuf */
 	u64 value_width;
 
@@ -207,6 +157,33 @@ int unpack(const void *pbuf, u64 *uval, size_t startbit, size_t endbit,
 	if (unlikely(value_width > 64))
 		return -ERANGE;
 
+	/* Check if "uval" fits in "value_width" bits.
+	 * If value_width is 64, the check will fail, but any
+	 * 64-bit uval will surely fit.
+	 */
+	if (value_width < 64 && uval >= (1ull << value_width))
+		/* Cannot store "uval" inside "value_width" bits.
+		 * Truncating "uval" is most certainly not desirable,
+		 * so simply erroring out is appropriate.
+		 */
+		return -ERANGE;
+
+	__pack(pbuf, uval, startbit, endbit, pbuflen, quirks);
+
+	return 0;
+}
+EXPORT_SYMBOL(pack);
+
+static void __unpack(const void *pbuf, u64 *uval, size_t startbit, size_t endbit,
+		     size_t pbuflen, u8 quirks)
+{
+	/* Logical byte indices corresponding to the
+	 * start and end of the field.
+	 */
+	int plogical_first_u8 = startbit / BITS_PER_BYTE;
+	int plogical_last_u8 = endbit / BITS_PER_BYTE;
+	int box;
+
 	/* Initialize parameter */
 	*uval = 0;
 
@@ -214,9 +191,6 @@ int unpack(const void *pbuf, u64 *uval, size_t startbit, size_t endbit,
 	 * no quirks, u8 by u8 (aligned at u8 boundaries), from high to low
 	 * logical bit significance. "box" denotes the current logical u8.
 	 */
-	plogical_first_u8 = startbit / BITS_PER_BYTE;
-	plogical_last_u8  = endbit / BITS_PER_BYTE;
-
 	for (box = plogical_first_u8; box >= plogical_last_u8; box--) {
 		/* Bit indices into the currently accessed 8-bit box */
 		size_t box_start_bit, box_end_bit, box_addr;
@@ -271,6 +245,46 @@ int unpack(const void *pbuf, u64 *uval, size_t startbit, size_t endbit,
 		*uval &= ~proj_mask;
 		*uval |= pval;
 	}
+}
+
+/**
+ * unpack - Unpack u64 number from packed buffer.
+ *
+ * @pbuf: Pointer to a buffer holding the packed value.
+ * @uval: Pointer to an u64 holding the unpacked value.
+ * @startbit: The index (in logical notation, compensated for quirks) where
+ *	      the packed value starts within pbuf. Must be larger than, or
+ *	      equal to, endbit.
+ * @endbit: The index (in logical notation, compensated for quirks) where
+ *	    the packed value ends within pbuf. Must be smaller than, or equal
+ *	    to, startbit.
+ * @pbuflen: The length in bytes of the packed buffer pointed to by @pbuf.
+ * @quirks: A bit mask of QUIRK_LITTLE_ENDIAN, QUIRK_LSW32_IS_FIRST and
+ *	    QUIRK_MSB_ON_THE_RIGHT.
+ *
+ * Return: 0 on success, EINVAL or ERANGE if called incorrectly. Assuming
+ *	   correct usage, return code may be discarded. The @uval will be
+ *	   modified on success.
+ */
+int unpack(const void *pbuf, u64 *uval, size_t startbit, size_t endbit,
+	   size_t pbuflen, u8 quirks)
+{
+	/* width of the field to access in the pbuf */
+	u64 value_width;
+
+	/* startbit is expected to be larger than endbit, and both are
+	 * expected to be within the logically addressable range of the buffer.
+	 */
+	if (startbit < endbit || startbit >= BITS_PER_BYTE * pbuflen)
+		/* Invalid function call */
+		return -EINVAL;
+
+	value_width = startbit - endbit + 1;
+	if (value_width > 64)
+		return -ERANGE;
+
+	__unpack(pbuf, uval, startbit, endbit, pbuflen, quirks);
+
 	return 0;
 }
 EXPORT_SYMBOL(unpack);

-- 
2.47.0.265.g4ca455297942

[PATCH net-next v9 02/10] lib: packing: demote truncation error in pack() to a warning in __pack()

[Jacob Keller]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

Most of the sanity checks in pack() and unpack() can be covered at
compile time. There is only one exception, and that is truncation of the
uval during a pack() operation.

We'd like the error-less __pack() to catch that condition as well. But
at the same time, it is currently the responsibility of consumer drivers
(currently just sja1105) to print anything at all when this error
occurs, and then discard the return code.

We can just print a loud warning in the library code and continue with
the truncated __pack() operation. In practice, having the warning is
very important, see commit 24deec6b9e4a ("net: dsa: sja1105: disallow
C45 transactions on the BASE-TX MDIO bus") where the bug was caught
exactly by noticing this print.

Add the first print to the packing library, and at the same time remove
the print for the same condition from the sja1105 driver, to avoid
double printing.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/dsa/sja1105/sja1105_static_config.c |  8 ++------
 lib/packing.c                                   | 26 ++++++++++---------------
 2 files changed, 12 insertions(+), 22 deletions(-)

diff --git a/drivers/net/dsa/sja1105/sja1105_static_config.c b/drivers/net/dsa/sja1105/sja1105_static_config.c
index baba204ad62f6b507a6ccf3337248dd02b777249..3d790f8c6f4dab3640ede014345ef469fefb7085 100644
--- a/drivers/net/dsa/sja1105/sja1105_static_config.c
+++ b/drivers/net/dsa/sja1105/sja1105_static_config.c
@@ -26,12 +26,8 @@ void sja1105_pack(void *buf, const u64 *val, int start, int end, size_t len)
 		pr_err("Start bit (%d) expected to be larger than end (%d)\n",
 		       start, end);
 	} else if (rc == -ERANGE) {
-		if ((start - end + 1) > 64)
-			pr_err("Field %d-%d too large for 64 bits!\n",
-			       start, end);
-		else
-			pr_err("Cannot store %llx inside bits %d-%d (would truncate)\n",
-			       *val, start, end);
+		pr_err("Field %d-%d too large for 64 bits!\n",
+		       start, end);
 	}
 	dump_stack();
 }
diff --git a/lib/packing.c b/lib/packing.c
index f237b8af99f5fa8e839c38126769c50b2bfe6361..09a2d195b9433b61c86f3b63ff019ab319c83e97 100644
--- a/lib/packing.c
+++ b/lib/packing.c
@@ -59,8 +59,17 @@ static void __pack(void *pbuf, u64 uval, size_t startbit, size_t endbit,
 	 */
 	int plogical_first_u8 = startbit / BITS_PER_BYTE;
 	int plogical_last_u8 = endbit / BITS_PER_BYTE;
+	int value_width = startbit - endbit + 1;
 	int box;
 
+	/* Check if "uval" fits in "value_width" bits.
+	 * The test only works for value_width < 64, but in the latter case,
+	 * any 64-bit uval will surely fit.
+	 */
+	WARN(value_width < 64 && uval >= (1ull << value_width),
+	     "Cannot store 0x%llx inside bits %zu-%zu - will truncate\n",
+	     uval, startbit, endbit);
+
 	/* Iterate through an idealistic view of the pbuf as an u64 with
 	 * no quirks, u8 by u8 (aligned at u8 boundaries), from high to low
 	 * logical bit significance. "box" denotes the current logical u8.
@@ -143,9 +152,6 @@ static void __pack(void *pbuf, u64 uval, size_t startbit, size_t endbit,
 int pack(void *pbuf, u64 uval, size_t startbit, size_t endbit, size_t pbuflen,
 	 u8 quirks)
 {
-	/* width of the field to access in the pbuf */
-	u64 value_width;
-
 	/* startbit is expected to be larger than endbit, and both are
 	 * expected to be within the logically addressable range of the buffer.
 	 */
@@ -153,19 +159,7 @@ int pack(void *pbuf, u64 uval, size_t startbit, size_t endbit, size_t pbuflen,
 		/* Invalid function call */
 		return -EINVAL;
 
-	value_width = startbit - endbit + 1;
-	if (unlikely(value_width > 64))
-		return -ERANGE;
-
-	/* Check if "uval" fits in "value_width" bits.
-	 * If value_width is 64, the check will fail, but any
-	 * 64-bit uval will surely fit.
-	 */
-	if (value_width < 64 && uval >= (1ull << value_width))
-		/* Cannot store "uval" inside "value_width" bits.
-		 * Truncating "uval" is most certainly not desirable,
-		 * so simply erroring out is appropriate.
-		 */
+	if (unlikely(startbit - endbit >= 64))
 		return -ERANGE;
 
 	__pack(pbuf, uval, startbit, endbit, pbuflen, quirks);

-- 
2.47.0.265.g4ca455297942

[PATCH net-next v9 03/10] lib: packing: add pack_fields() and unpack_fields()

[Jacob Keller]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

This is new API which caters to the following requirements:

- Pack or unpack a large number of fields to/from a buffer with a small
  code footprint. The current alternative is to open-code a large number
  of calls to pack() and unpack(), or to use packing() to reduce that
  number to half. But packing() is not const-correct.

- Use unpacked numbers stored in variables smaller than u64. This
  reduces the rodata footprint of the stored field arrays.

- Perform error checking at compile time, rather than runtime, and return
  void from the API functions. Because the C preprocessor can't generate
  variable length code (loops), this is a bit tricky to do with macros.

  To handle this, implement macros which sanity check the packed field
  definitions based on their size. Finally, a single macro with a chain of
  __builtin_choose_expr() is used to select the appropriate macros. We
  enforce the use of ascending or descending order to avoid O(N^2) scaling
  when checking for overlap. Note that the macros are written with care to
  ensure that the compilers can correctly evaluate the resulting code at
  compile time. In particular, care was taken with avoiding too many nested
  statement expressions. Nested statement expressions trip up some
  compilers, especially when passing down variables created in previous
  statement expressions.

  There are two key design choices intended to keep the overall macro code
  size small. First, the definition of each CHECK_PACKED_FIELDS_N macro is
  implemented recursively, by calling the N-1 macro. This avoids needing
  the code to repeat multiple times.

  Second, the CHECK_PACKED_FIELD macro enforces that the fields in the
  array are sorted in order. This allows checking for overlap only with
  neighboring fields, rather than the general overlap case where each field
  would need to be checked against other fields.

  The overlap checks use the first two fields to determine the order of the
  remaining fields, thus allowing either ascending or descending order.
  This enables drivers the flexibility to keep the fields ordered in which
  ever order most naturally fits their hardware design and its associated
  documentation.

  The CHECK_PACKED_FIELDS macro is directly called from within pack_fields
  and unpack_fields, ensuring that all drivers using the API receive the
  benefits of the compile-time checks. Users do not need to directly call
  any of the macros directly.

  The CHECK_PACKED_FIELDS and its helper macros CHECK_PACKED_FIELDS_(0..50)
  are generated using a simple C program in scripts/gen_packed_field_checks.c
  This program can be compiled on demand and executed to generate the macro
  code in include/linux/packing.h. This will aid in the event that a driver
  needs more than 50 fields. The generator can be updated with a new size,
  and used to update the packing.h header file. In practice, the ice driver
  will need to support 27 fields, and the sja1105 driver will need to
  support 40 fields. This on-demand generation avoids the need to modify
  Kbuild. We do not anticipate the maximum number of fields to grow very
  often.

- Reduced rodata footprint for the storage of the packed field arrays.
  To that end, we have struct packed_field_s (small) and packed_field_m
  (medium). More can be added as needed (unlikely for now). On these
  types, the same generic pack_fields() and unpack_fields() API can be
  used, thanks to the new C11 _Generic() selection feature, which can
  call pack_fields_s() or pack_fields_m(), depending on the type of the
  "fields" array - a simplistic form of polymorphism. It is evaluated at
  compile time which function will actually be called.

Over time, packing() is expected to be completely replaced either with
pack() or with pack_fields().

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Co-developed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
 Makefile                          |   4 +
 include/linux/packing.h           | 427 ++++++++++++++++++++++++++++++++++++++
 lib/packing.c                     | 145 +++++++++++++
 lib/packing_test.c                |  61 ++++++
 scripts/gen_packed_field_checks.c |  37 ++++
 MAINTAINERS                       |   1 +
 scripts/Makefile                  |   2 +-
 7 files changed, 676 insertions(+), 1 deletion(-)

diff --git a/Makefile b/Makefile
index 8129de0b214f5b73a3b1cca0798041d74270836b..58496942a7d13c6a53e4210d83deb2cc2033d00a 100644
--- a/Makefile
+++ b/Makefile
@@ -1315,6 +1315,10 @@ PHONY += scripts_unifdef
 scripts_unifdef: scripts_basic
 	$(Q)$(MAKE) $(build)=scripts scripts/unifdef
 
+PHONY += scripts_gen_packed_field_checks
+scripts_gen_packed_field_checks: scripts_basic
+	$(Q)$(MAKE) $(build)=scripts scripts/gen_packed_field_checks
+
 # ---------------------------------------------------------------------------
 # Install
 
diff --git a/include/linux/packing.h b/include/linux/packing.h
index 5d36dcd06f60420325473dae3a0e9ac37d03da4b..d9454ea7fdcdec142796f5bbec1fccb7e74d5fbb 100644
--- a/include/linux/packing.h
+++ b/include/linux/packing.h
@@ -8,6 +8,85 @@
 #include <linux/types.h>
 #include <linux/bitops.h>
 
+#define GEN_PACKED_FIELD_MEMBERS(__type) \
+	__type startbit; \
+	__type endbit; \
+	__type offset; \
+	__type size
+
+/* Small packed field. Use with bit offsets < 256, buffers < 32B and
+ * unpacked structures < 256B.
+ */
+struct packed_field_s {
+	GEN_PACKED_FIELD_MEMBERS(u8);
+};
+
+/* Medium packed field. Use with bit offsets < 65536, buffers < 8KB and
+ * unpacked structures < 64KB.
+ */
+struct packed_field_m {
+	GEN_PACKED_FIELD_MEMBERS(u16);
+};
+
+#define PACKED_FIELD(start, end, struct_name, struct_field) \
+{ \
+	(start), \
+	(end), \
+	offsetof(struct_name, struct_field), \
+	sizeof_field(struct_name, struct_field), \
+}
+
+#define CHECK_PACKED_FIELD_OVERLAP(fields, index1, index2) ({ \
+	typeof(&(fields)[0]) __f = (fields); \
+	typeof(__f[0]) _f1 = __f[index1]; typeof(__f[0]) _f2 = __f[index2]; \
+	const bool _ascending = __f[0].startbit < __f[1].startbit; \
+	BUILD_BUG_ON_MSG(_ascending && _f1.startbit >= _f2.startbit, \
+			 __stringify(fields) " field " __stringify(index2) \
+			 " breaks ascending order"); \
+	BUILD_BUG_ON_MSG(!_ascending && _f1.startbit <= _f2.startbit, \
+			 __stringify(fields) " field " __stringify(index2) \
+			 " breaks descending order"); \
+	BUILD_BUG_ON_MSG(max(_f1.endbit, _f2.endbit) <= \
+			 min(_f1.startbit, _f2.startbit), \
+			 __stringify(fields) " field " __stringify(index2) \
+			 " overlaps with previous field"); \
+})
+
+#define CHECK_PACKED_FIELD(fields, index) ({ \
+	typeof(&(fields)[0]) _f = (fields); \
+	typeof(_f[0]) __f = _f[index]; \
+	BUILD_BUG_ON_MSG(__f.startbit < __f.endbit, \
+			 __stringify(fields) " field " __stringify(index) \
+			 " start bit must not be smaller than end bit"); \
+	BUILD_BUG_ON_MSG(__f.size != 1 && __f.size != 2 && \
+			 __f.size != 4 && __f.size != 8, \
+			 __stringify(fields) " field " __stringify(index) \
+			" has unsupported unpacked storage size"); \
+	BUILD_BUG_ON_MSG(__f.startbit - __f.endbit >= BITS_PER_BYTE * __f.size, \
+			 __stringify(fields) " field " __stringify(index) \
+			 " exceeds unpacked storage size"); \
+	__builtin_choose_expr(index != 0, \
+			      CHECK_PACKED_FIELD_OVERLAP(fields, index - 1, index), \
+			      1); \
+})
+
+/* Note that the packed fields may be either in ascending or descending order.
+ * Thus, we must check that both the first and last field wit within the
+ * packed buffer size.
+ */
+#define CHECK_PACKED_FIELDS_SIZE(fields, pbuflen) ({ \
+	typeof(&(fields)[0]) _f = (fields); \
+	typeof(pbuflen) _len = (pbuflen); \
+	const size_t num_fields = ARRAY_SIZE(fields); \
+	BUILD_BUG_ON_MSG(!__builtin_constant_p(_len), \
+			 __stringify(fields) " pbuflen " __stringify(pbuflen) \
+			 " must be a compile time constant"); \
+	BUILD_BUG_ON_MSG(_f[0].startbit >= BITS_PER_BYTE * _len, \
+			 __stringify(fields) " first field exceeds packed buffer size"); \
+	BUILD_BUG_ON_MSG(_f[num_fields - 1].startbit >= BITS_PER_BYTE * _len, \
+			 __stringify(fields) " last field exceeds packed buffer size"); \
+})
+
 #define QUIRK_MSB_ON_THE_RIGHT	BIT(0)
 #define QUIRK_LITTLE_ENDIAN	BIT(1)
 #define QUIRK_LSW32_IS_FIRST	BIT(2)
@@ -26,4 +105,352 @@ int pack(void *pbuf, u64 uval, size_t startbit, size_t endbit, size_t pbuflen,
 int unpack(const void *pbuf, u64 *uval, size_t startbit, size_t endbit,
 	   size_t pbuflen, u8 quirks);
 
+void pack_fields_s(void *pbuf, size_t pbuflen, const void *ustruct,
+		   const struct packed_field_s *fields, size_t num_fields,
+		   u8 quirks);
+
+void pack_fields_m(void *pbuf, size_t pbuflen, const void *ustruct,
+		   const struct packed_field_m *fields, size_t num_fields,
+		   u8 quirks);
+
+void unpack_fields_s(const void *pbuf, size_t pbuflen, void *ustruct,
+		     const struct packed_field_s *fields, size_t num_fields,
+		     u8 quirks);
+
+void unpack_fields_m(const void *pbuf, size_t pbuflen, void *ustruct,
+		     const struct packed_field_m *fields, size_t num_fields,
+		     u8 quirks);
+
+/* Do not hand-edit the following packed field check macros!
+ *
+ * They are generated using scripts/gen_packed_field_checks.c, which may be
+ * built via "make scripts_gen_packed_field_checks". If larger macro sizes are
+ * needed in the future, please use this program to re-generate the macros and
+ * insert them here.
+ */
+
+#define CHECK_PACKED_FIELDS_1(fields) \
+	CHECK_PACKED_FIELD(fields, 0)
+
+#define CHECK_PACKED_FIELDS_2(fields) do { \
+	CHECK_PACKED_FIELDS_1(fields); \
+	CHECK_PACKED_FIELD(fields, 1); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_3(fields) do { \
+	CHECK_PACKED_FIELDS_2(fields); \
+	CHECK_PACKED_FIELD(fields, 2); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_4(fields) do { \
+	CHECK_PACKED_FIELDS_3(fields); \
+	CHECK_PACKED_FIELD(fields, 3); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_5(fields) do { \
+	CHECK_PACKED_FIELDS_4(fields); \
+	CHECK_PACKED_FIELD(fields, 4); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_6(fields) do { \
+	CHECK_PACKED_FIELDS_5(fields); \
+	CHECK_PACKED_FIELD(fields, 5); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_7(fields) do { \
+	CHECK_PACKED_FIELDS_6(fields); \
+	CHECK_PACKED_FIELD(fields, 6); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_8(fields) do { \
+	CHECK_PACKED_FIELDS_7(fields); \
+	CHECK_PACKED_FIELD(fields, 7); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_9(fields) do { \
+	CHECK_PACKED_FIELDS_8(fields); \
+	CHECK_PACKED_FIELD(fields, 8); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_10(fields) do { \
+	CHECK_PACKED_FIELDS_9(fields); \
+	CHECK_PACKED_FIELD(fields, 9); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_11(fields) do { \
+	CHECK_PACKED_FIELDS_10(fields); \
+	CHECK_PACKED_FIELD(fields, 10); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_12(fields) do { \
+	CHECK_PACKED_FIELDS_11(fields); \
+	CHECK_PACKED_FIELD(fields, 11); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_13(fields) do { \
+	CHECK_PACKED_FIELDS_12(fields); \
+	CHECK_PACKED_FIELD(fields, 12); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_14(fields) do { \
+	CHECK_PACKED_FIELDS_13(fields); \
+	CHECK_PACKED_FIELD(fields, 13); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_15(fields) do { \
+	CHECK_PACKED_FIELDS_14(fields); \
+	CHECK_PACKED_FIELD(fields, 14); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_16(fields) do { \
+	CHECK_PACKED_FIELDS_15(fields); \
+	CHECK_PACKED_FIELD(fields, 15); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_17(fields) do { \
+	CHECK_PACKED_FIELDS_16(fields); \
+	CHECK_PACKED_FIELD(fields, 16); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_18(fields) do { \
+	CHECK_PACKED_FIELDS_17(fields); \
+	CHECK_PACKED_FIELD(fields, 17); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_19(fields) do { \
+	CHECK_PACKED_FIELDS_18(fields); \
+	CHECK_PACKED_FIELD(fields, 18); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_20(fields) do { \
+	CHECK_PACKED_FIELDS_19(fields); \
+	CHECK_PACKED_FIELD(fields, 19); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_21(fields) do { \
+	CHECK_PACKED_FIELDS_20(fields); \
+	CHECK_PACKED_FIELD(fields, 20); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_22(fields) do { \
+	CHECK_PACKED_FIELDS_21(fields); \
+	CHECK_PACKED_FIELD(fields, 21); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_23(fields) do { \
+	CHECK_PACKED_FIELDS_22(fields); \
+	CHECK_PACKED_FIELD(fields, 22); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_24(fields) do { \
+	CHECK_PACKED_FIELDS_23(fields); \
+	CHECK_PACKED_FIELD(fields, 23); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_25(fields) do { \
+	CHECK_PACKED_FIELDS_24(fields); \
+	CHECK_PACKED_FIELD(fields, 24); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_26(fields) do { \
+	CHECK_PACKED_FIELDS_25(fields); \
+	CHECK_PACKED_FIELD(fields, 25); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_27(fields) do { \
+	CHECK_PACKED_FIELDS_26(fields); \
+	CHECK_PACKED_FIELD(fields, 26); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_28(fields) do { \
+	CHECK_PACKED_FIELDS_27(fields); \
+	CHECK_PACKED_FIELD(fields, 27); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_29(fields) do { \
+	CHECK_PACKED_FIELDS_28(fields); \
+	CHECK_PACKED_FIELD(fields, 28); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_30(fields) do { \
+	CHECK_PACKED_FIELDS_29(fields); \
+	CHECK_PACKED_FIELD(fields, 29); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_31(fields) do { \
+	CHECK_PACKED_FIELDS_30(fields); \
+	CHECK_PACKED_FIELD(fields, 30); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_32(fields) do { \
+	CHECK_PACKED_FIELDS_31(fields); \
+	CHECK_PACKED_FIELD(fields, 31); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_33(fields) do { \
+	CHECK_PACKED_FIELDS_32(fields); \
+	CHECK_PACKED_FIELD(fields, 32); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_34(fields) do { \
+	CHECK_PACKED_FIELDS_33(fields); \
+	CHECK_PACKED_FIELD(fields, 33); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_35(fields) do { \
+	CHECK_PACKED_FIELDS_34(fields); \
+	CHECK_PACKED_FIELD(fields, 34); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_36(fields) do { \
+	CHECK_PACKED_FIELDS_35(fields); \
+	CHECK_PACKED_FIELD(fields, 35); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_37(fields) do { \
+	CHECK_PACKED_FIELDS_36(fields); \
+	CHECK_PACKED_FIELD(fields, 36); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_38(fields) do { \
+	CHECK_PACKED_FIELDS_37(fields); \
+	CHECK_PACKED_FIELD(fields, 37); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_39(fields) do { \
+	CHECK_PACKED_FIELDS_38(fields); \
+	CHECK_PACKED_FIELD(fields, 38); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_40(fields) do { \
+	CHECK_PACKED_FIELDS_39(fields); \
+	CHECK_PACKED_FIELD(fields, 39); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_41(fields) do { \
+	CHECK_PACKED_FIELDS_40(fields); \
+	CHECK_PACKED_FIELD(fields, 40); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_42(fields) do { \
+	CHECK_PACKED_FIELDS_41(fields); \
+	CHECK_PACKED_FIELD(fields, 41); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_43(fields) do { \
+	CHECK_PACKED_FIELDS_42(fields); \
+	CHECK_PACKED_FIELD(fields, 42); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_44(fields) do { \
+	CHECK_PACKED_FIELDS_43(fields); \
+	CHECK_PACKED_FIELD(fields, 43); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_45(fields) do { \
+	CHECK_PACKED_FIELDS_44(fields); \
+	CHECK_PACKED_FIELD(fields, 44); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_46(fields) do { \
+	CHECK_PACKED_FIELDS_45(fields); \
+	CHECK_PACKED_FIELD(fields, 45); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_47(fields) do { \
+	CHECK_PACKED_FIELDS_46(fields); \
+	CHECK_PACKED_FIELD(fields, 46); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_48(fields) do { \
+	CHECK_PACKED_FIELDS_47(fields); \
+	CHECK_PACKED_FIELD(fields, 47); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_49(fields) do { \
+	CHECK_PACKED_FIELDS_48(fields); \
+	CHECK_PACKED_FIELD(fields, 48); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS_50(fields) do { \
+	CHECK_PACKED_FIELDS_49(fields); \
+	CHECK_PACKED_FIELD(fields, 49); \
+} while (0)
+
+#define CHECK_PACKED_FIELDS(fields) \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 1, ({ CHECK_PACKED_FIELDS_1(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 2, ({ CHECK_PACKED_FIELDS_2(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 3, ({ CHECK_PACKED_FIELDS_3(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 4, ({ CHECK_PACKED_FIELDS_4(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 5, ({ CHECK_PACKED_FIELDS_5(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 6, ({ CHECK_PACKED_FIELDS_6(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 7, ({ CHECK_PACKED_FIELDS_7(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 8, ({ CHECK_PACKED_FIELDS_8(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 9, ({ CHECK_PACKED_FIELDS_9(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 10, ({ CHECK_PACKED_FIELDS_10(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 11, ({ CHECK_PACKED_FIELDS_11(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 12, ({ CHECK_PACKED_FIELDS_12(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 13, ({ CHECK_PACKED_FIELDS_13(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 14, ({ CHECK_PACKED_FIELDS_14(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 15, ({ CHECK_PACKED_FIELDS_15(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 16, ({ CHECK_PACKED_FIELDS_16(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 17, ({ CHECK_PACKED_FIELDS_17(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 18, ({ CHECK_PACKED_FIELDS_18(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 19, ({ CHECK_PACKED_FIELDS_19(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 20, ({ CHECK_PACKED_FIELDS_20(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 21, ({ CHECK_PACKED_FIELDS_21(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 22, ({ CHECK_PACKED_FIELDS_22(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 23, ({ CHECK_PACKED_FIELDS_23(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 24, ({ CHECK_PACKED_FIELDS_24(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 25, ({ CHECK_PACKED_FIELDS_25(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 26, ({ CHECK_PACKED_FIELDS_26(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 27, ({ CHECK_PACKED_FIELDS_27(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 28, ({ CHECK_PACKED_FIELDS_28(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 29, ({ CHECK_PACKED_FIELDS_29(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 30, ({ CHECK_PACKED_FIELDS_30(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 31, ({ CHECK_PACKED_FIELDS_31(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 32, ({ CHECK_PACKED_FIELDS_32(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 33, ({ CHECK_PACKED_FIELDS_33(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 34, ({ CHECK_PACKED_FIELDS_34(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 35, ({ CHECK_PACKED_FIELDS_35(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 36, ({ CHECK_PACKED_FIELDS_36(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 37, ({ CHECK_PACKED_FIELDS_37(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 38, ({ CHECK_PACKED_FIELDS_38(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 39, ({ CHECK_PACKED_FIELDS_39(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 40, ({ CHECK_PACKED_FIELDS_40(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 41, ({ CHECK_PACKED_FIELDS_41(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 42, ({ CHECK_PACKED_FIELDS_42(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 43, ({ CHECK_PACKED_FIELDS_43(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 44, ({ CHECK_PACKED_FIELDS_44(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 45, ({ CHECK_PACKED_FIELDS_45(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 46, ({ CHECK_PACKED_FIELDS_46(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 47, ({ CHECK_PACKED_FIELDS_47(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 48, ({ CHECK_PACKED_FIELDS_48(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 49, ({ CHECK_PACKED_FIELDS_49(fields); }), \
+	__builtin_choose_expr(ARRAY_SIZE(fields) == 50, ({ CHECK_PACKED_FIELDS_50(fields); }), \
+	({ BUILD_BUG_ON_MSG(1, "CHECK_PACKED_FIELDS() must be regenerated to support array sizes larger than 50."); }) \
+))))))))))))))))))))))))))))))))))))))))))))))))))
+
+/* End of generated content */
+
+#define pack_fields(pbuf, pbuflen, ustruct, fields, quirks) \
+	({ \
+		CHECK_PACKED_FIELDS(fields); \
+		CHECK_PACKED_FIELDS_SIZE((fields), (pbuflen)); \
+		_Generic((fields), \
+			 const struct packed_field_s * : pack_fields_s, \
+			 const struct packed_field_m * : pack_fields_m \
+			)((pbuf), (pbuflen), (ustruct), (fields), ARRAY_SIZE(fields), (quirks)); \
+	})
+
+#define unpack_fields(pbuf, pbuflen, ustruct, fields, quirks) \
+	({ \
+		CHECK_PACKED_FIELDS(fields); \
+		CHECK_PACKED_FIELDS_SIZE((fields), (pbuflen)); \
+		_Generic((fields), \
+			 const struct packed_field_s * : unpack_fields_s, \
+			 const struct packed_field_m * : unpack_fields_m \
+			)((pbuf), (pbuflen), (ustruct), (fields), ARRAY_SIZE(fields), (quirks)); \
+	})
+
 #endif
diff --git a/lib/packing.c b/lib/packing.c
index 09a2d195b9433b61c86f3b63ff019ab319c83e97..45164f73fe5bf9f2c547eb22016af7e44fed9eb0 100644
--- a/lib/packing.c
+++ b/lib/packing.c
@@ -5,10 +5,37 @@
 #include <linux/packing.h>
 #include <linux/module.h>
 #include <linux/bitops.h>
+#include <linux/bits.h>
 #include <linux/errno.h>
 #include <linux/types.h>
 #include <linux/bitrev.h>
 
+#define __pack_fields(pbuf, pbuflen, ustruct, fields, num_fields, quirks)	\
+	({									\
+		for (size_t i = 0; i < (num_fields); i++) {			\
+			typeof(&(fields)[0]) field = &(fields)[i];		\
+			u64 uval;						\
+										\
+			uval = ustruct_field_to_u64(ustruct, field->offset, field->size); \
+										\
+			__pack(pbuf, uval, field->startbit, field->endbit,	\
+			       pbuflen, quirks);				\
+		}								\
+	})
+
+#define __unpack_fields(pbuf, pbuflen, ustruct, fields, num_fields, quirks)	\
+	({									\
+		for (size_t i = 0; i < (num_fields); i++) {			\
+			typeof(&(fields)[0]) field = &fields[i];		\
+			u64 uval;						\
+										\
+			__unpack(pbuf, &uval, field->startbit, field->endbit,	\
+				 pbuflen, quirks);				\
+										\
+			u64_to_ustruct_field(ustruct, field->offset, field->size, uval); \
+		}								\
+	})
+
 /**
  * calculate_box_addr - Determine physical location of byte in buffer
  * @box: Index of byte within buffer seen as a logical big-endian big number
@@ -322,4 +349,122 @@ int packing(void *pbuf, u64 *uval, int startbit, int endbit, size_t pbuflen,
 }
 EXPORT_SYMBOL(packing);
 
+static u64 ustruct_field_to_u64(const void *ustruct, size_t field_offset,
+				size_t field_size)
+{
+	switch (field_size) {
+	case 1:
+		return *((u8 *)(ustruct + field_offset));
+	case 2:
+		return *((u16 *)(ustruct + field_offset));
+	case 4:
+		return *((u32 *)(ustruct + field_offset));
+	default:
+		return *((u64 *)(ustruct + field_offset));
+	}
+}
+
+static void u64_to_ustruct_field(void *ustruct, size_t field_offset,
+				 size_t field_size, u64 uval)
+{
+	switch (field_size) {
+	case 1:
+		*((u8 *)(ustruct + field_offset)) = uval;
+		break;
+	case 2:
+		*((u16 *)(ustruct + field_offset)) = uval;
+		break;
+	case 4:
+		*((u32 *)(ustruct + field_offset)) = uval;
+		break;
+	default:
+		*((u64 *)(ustruct + field_offset)) = uval;
+		break;
+	}
+}
+
+/**
+ * pack_fields_s - Pack array of small fields
+ *
+ * @pbuf: Pointer to a buffer holding the packed value.
+ * @pbuflen: The length in bytes of the packed buffer pointed to by @pbuf.
+ * @ustruct: Pointer to CPU-readable structure holding the unpacked value.
+ *	     It is expected (but not checked) that this has the same data type
+ *	     as all struct packed_field_s definitions.
+ * @fields: Array of small packed fields definition. They must not overlap.
+ * @num_fields: Length of @fields array.
+ * @quirks: A bit mask of QUIRK_LITTLE_ENDIAN, QUIRK_LSW32_IS_FIRST and
+ *	    QUIRK_MSB_ON_THE_RIGHT.
+ */
+void pack_fields_s(void *pbuf, size_t pbuflen, const void *ustruct,
+		   const struct packed_field_s *fields, size_t num_fields,
+		   u8 quirks)
+{
+	__pack_fields(pbuf, pbuflen, ustruct, fields, num_fields, quirks);
+}
+EXPORT_SYMBOL(pack_fields_s);
+
+/**
+ * pack_fields_m - Pack array of medium fields
+ *
+ * @pbuf: Pointer to a buffer holding the packed value.
+ * @pbuflen: The length in bytes of the packed buffer pointed to by @pbuf.
+ * @ustruct: Pointer to CPU-readable structure holding the unpacked value.
+ *	     It is expected (but not checked) that this has the same data type
+ *	     as all struct packed_field_s definitions.
+ * @fields: Array of medium packed fields definition. They must not overlap.
+ * @num_fields: Length of @fields array.
+ * @quirks: A bit mask of QUIRK_LITTLE_ENDIAN, QUIRK_LSW32_IS_FIRST and
+ *	    QUIRK_MSB_ON_THE_RIGHT.
+ */
+void pack_fields_m(void *pbuf, size_t pbuflen, const void *ustruct,
+		   const struct packed_field_m *fields, size_t num_fields,
+		   u8 quirks)
+{
+	__pack_fields(pbuf, pbuflen, ustruct, fields, num_fields, quirks);
+}
+EXPORT_SYMBOL(pack_fields_m);
+
+/**
+ * unpack_fields_s - Unpack array of small fields
+ *
+ * @pbuf: Pointer to a buffer holding the packed value.
+ * @pbuflen: The length in bytes of the packed buffer pointed to by @pbuf.
+ * @ustruct: Pointer to CPU-readable structure holding the unpacked value.
+ *	     It is expected (but not checked) that this has the same data type
+ *	     as all struct packed_field_s definitions.
+ * @fields: Array of small packed fields definition. They must not overlap.
+ * @num_fields: Length of @fields array.
+ * @quirks: A bit mask of QUIRK_LITTLE_ENDIAN, QUIRK_LSW32_IS_FIRST and
+ *	    QUIRK_MSB_ON_THE_RIGHT.
+ */
+void unpack_fields_s(const void *pbuf, size_t pbuflen, void *ustruct,
+		     const struct packed_field_s *fields, size_t num_fields,
+		     u8 quirks)
+{
+	__unpack_fields(pbuf, pbuflen, ustruct, fields, num_fields, quirks);
+}
+EXPORT_SYMBOL(unpack_fields_s);
+
+/**
+ * unpack_fields_m - Unpack array of medium fields
+ *
+ * @pbuf: Pointer to a buffer holding the packed value.
+ * @pbuflen: The length in bytes of the packed buffer pointed to by @pbuf.
+ * @ustruct: Pointer to CPU-readable structure holding the unpacked value.
+ *	     It is expected (but not checked) that this has the same data type
+ *	     as all struct packed_field_s definitions.
+ * @fields: Array of medium packed fields definition. They must not overlap.
+ * @num_fields: Length of @fields array.
+ * @quirks: A bit mask of QUIRK_LITTLE_ENDIAN, QUIRK_LSW32_IS_FIRST and
+ *	    QUIRK_MSB_ON_THE_RIGHT.
+ */
+void unpack_fields_m(const void *pbuf, size_t pbuflen, void *ustruct,
+		     const struct packed_field_m *fields, size_t num_fields,
+		     u8 quirks)
+{
+	__unpack_fields(pbuf, pbuflen, ustruct, fields, num_fields, quirks);
+}
+EXPORT_SYMBOL(unpack_fields_m);
+
 MODULE_DESCRIPTION("Generic bitfield packing and unpacking");
diff --git a/lib/packing_test.c b/lib/packing_test.c
index b38ea43c03fd83639f18a6f3e2a42eae36118c45..3b4167ce56bf65fa4d66cb55d3215aecc33f64c4 100644
--- a/lib/packing_test.c
+++ b/lib/packing_test.c
@@ -396,9 +396,70 @@ static void packing_test_unpack(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, uval, params->uval);
 }
 
+#define PACKED_BUF_SIZE 8
+
+typedef struct __packed { u8 buf[PACKED_BUF_SIZE]; } packed_buf_t;
+
+struct test_data {
+	u32 field3;
+	u16 field2;
+	u16 field4;
+	u16 field6;
+	u8 field1;
+	u8 field5;
+};
+
+static const struct packed_field_s test_fields[] = {
+	PACKED_FIELD(63, 61, struct test_data, field1),
+	PACKED_FIELD(60, 52, struct test_data, field2),
+	PACKED_FIELD(51, 28, struct test_data, field3),
+	PACKED_FIELD(27, 14, struct test_data, field4),
+	PACKED_FIELD(13, 9, struct test_data, field5),
+	PACKED_FIELD(8, 0, struct test_data, field6),
+};
+
+static void packing_test_pack_fields(struct kunit *test)
+{
+	const struct test_data data = {
+		.field1 = 0x2,
+		.field2 = 0x100,
+		.field3 = 0xF00050,
+		.field4 = 0x7D3,
+		.field5 = 0x9,
+		.field6 = 0x10B,
+	};
+	packed_buf_t expect = {
+		.buf = { 0x50, 0x0F, 0x00, 0x05, 0x01, 0xF4, 0xD3, 0x0B },
+	};
+	packed_buf_t buf = {};
+
+	pack_fields(&buf, sizeof(buf), &data, test_fields, 0);
+
+	KUNIT_EXPECT_MEMEQ(test, &expect, &buf, sizeof(buf));
+}
+
+static void packing_test_unpack_fields(struct kunit *test)
+{
+	const packed_buf_t buf = {
+		.buf = { 0x17, 0x28, 0x10, 0x19, 0x3D, 0xA9, 0x07, 0x9C },
+	};
+	struct test_data data = {};
+
+	unpack_fields(&buf, sizeof(buf), &data, test_fields, 0);
+
+	KUNIT_EXPECT_EQ(test, 0, data.field1);
+	KUNIT_EXPECT_EQ(test, 0x172, data.field2);
+	KUNIT_EXPECT_EQ(test, 0x810193, data.field3);
+	KUNIT_EXPECT_EQ(test, 0x36A4, data.field4);
+	KUNIT_EXPECT_EQ(test, 0x3, data.field5);
+	KUNIT_EXPECT_EQ(test, 0x19C, data.field6);
+}
+
 static struct kunit_case packing_test_cases[] = {
 	KUNIT_CASE_PARAM(packing_test_pack, packing_gen_params),
 	KUNIT_CASE_PARAM(packing_test_unpack, packing_gen_params),
+	KUNIT_CASE(packing_test_pack_fields),
+	KUNIT_CASE(packing_test_unpack_fields),
 	{},
 };
 
diff --git a/scripts/gen_packed_field_checks.c b/scripts/gen_packed_field_checks.c
new file mode 100644
index 0000000000000000000000000000000000000000..4e0b6e1bf4be23d9d2cc8c0cb1f13fdcb6fd07a3
--- /dev/null
+++ b/scripts/gen_packed_field_checks.c
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (c) 2024, Intel Corporation
+#include <stdbool.h>
+#include <stdio.h>
+
+#define MAX_PACKED_FIELD_SIZE 50
+
+int main(int argc, char **argv)
+{
+	/* The first macro doesn't need a do/while loop */
+	printf("#define CHECK_PACKED_FIELDS_1(fields) \\\n");
+	printf("\tCHECK_PACKED_FIELD(fields, 0)\n\n");
+
+	/* Remaining macros require a do/while loop, and are implemented
+	 * recursively by calling the previous iteration's macro.
+	 */
+	for (int i = 2; i <= MAX_PACKED_FIELD_SIZE; i++) {
+		printf("#define CHECK_PACKED_FIELDS_%d(fields) do { \\\n", i);
+		printf("\tCHECK_PACKED_FIELDS_%d(fields); \\\n", i - 1);
+		printf("\tCHECK_PACKED_FIELD(fields, %d); \\\n", i - 1);
+		printf("} while (0)\n\n");
+	}
+
+	printf("#define CHECK_PACKED_FIELDS(fields) \\\n");
+
+	for (int i = 1; i <= MAX_PACKED_FIELD_SIZE; i++)
+		printf("\t__builtin_choose_expr(ARRAY_SIZE(fields) == %d, ({ CHECK_PACKED_FIELDS_%d(fields); }), \\\n",
+		       i, i);
+
+	printf("\t({ BUILD_BUG_ON_MSG(1, \"CHECK_PACKED_FIELDS() must be regenerated to support array sizes larger than %d.\"); }) \\\n",
+	       MAX_PACKED_FIELD_SIZE);
+
+	for (int i = 1; i <= MAX_PACKED_FIELD_SIZE; i++)
+		printf(")");
+
+	printf("\n");
+}
diff --git a/MAINTAINERS b/MAINTAINERS
index 0456a33ef65792bacb5d305a6384d245844fb743..397dfdab2d92a969d367dcf77207d387cda451e3 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -17561,6 +17561,7 @@ F:	Documentation/core-api/packing.rst
 F:	include/linux/packing.h
 F:	lib/packing.c
 F:	lib/packing_test.c
+F:	scripts/gen_packed_field_checks.c
 
 PADATA PARALLEL EXECUTION MECHANISM
 M:	Steffen Klassert <steffen.klassert@secunet.com>
diff --git a/scripts/Makefile b/scripts/Makefile
index 6bcda4b9d054021b185488841cd36c6e0fb86d0c..546e8175e1c4c8209e67a7f92f7d1e795a030988 100644
--- a/scripts/Makefile
+++ b/scripts/Makefile
@@ -47,7 +47,7 @@ HOSTCFLAGS_sorttable.o += -DMCOUNT_SORT_ENABLED
 endif
 
 # The following programs are only built on demand
-hostprogs += unifdef
+hostprogs += unifdef gen_packed_field_checks
 
 # The module linker script is preprocessed on demand
 targets += module.lds

-- 
2.47.0.265.g4ca455297942

[PATCH net-next v9 04/10] lib: packing: document recently added APIs

[Jacob Keller]

Extend the documentation for the packing library, covering the intended use
for the recently added APIs. This includes the pack() and unpack() macros,
as well as the pack_fields() and unpack_fields() macros.

Add a note that the packing() API is now deprecated in favor of pack() and
unpack().

For the pack_fields() and unpack_fields() APIs, explain the rationale for
when a driver may want to select this API. Provide an example which shows
how to define the fields and call the pack_fields() and unpack_fields()
macros.

Co-developed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
 Documentation/core-api/packing.rst | 118 +++++++++++++++++++++++++++++++++++--
 1 file changed, 113 insertions(+), 5 deletions(-)

diff --git a/Documentation/core-api/packing.rst b/Documentation/core-api/packing.rst
index 821691f23c541cee27995bb1d77e23ff04f82433..a0c2d90a44861042263e8da1ea329245681e2a82 100644
--- a/Documentation/core-api/packing.rst
+++ b/Documentation/core-api/packing.rst
@@ -227,11 +227,119 @@ Intended use
 
 Drivers that opt to use this API first need to identify which of the above 3
 quirk combinations (for a total of 8) match what the hardware documentation
-describes. Then they should wrap the packing() function, creating a new
-xxx_packing() that calls it using the proper QUIRK_* one-hot bits set.
+describes.
+
+There are 3 supported usage patterns, detailed below.
+
+packing()
+^^^^^^^^^
+
+This API function is deprecated.
 
 The packing() function returns an int-encoded error code, which protects the
 programmer against incorrect API use.  The errors are not expected to occur
-during runtime, therefore it is reasonable for xxx_packing() to return void
-and simply swallow those errors. Optionally it can dump stack or print the
-error description.
+during runtime, therefore it is reasonable to wrap packing() into a custom
+function which returns void and swallows those errors. Optionally it can
+dump stack or print the error description.
+
+.. code-block:: c
+
+  void my_packing(void *buf, u64 *val, int startbit, int endbit,
+                  size_t len, enum packing_op op)
+  {
+          int err;
+
+          /* Adjust quirks accordingly */
+          err = packing(buf, val, startbit, endbit, len, op, QUIRK_LSW32_IS_FIRST);
+          if (likely(!err))
+                  return;
+
+          if (err == -EINVAL) {
+                  pr_err("Start bit (%d) expected to be larger than end (%d)\n",
+                         startbit, endbit);
+          } else if (err == -ERANGE) {
+                  if ((startbit - endbit + 1) > 64)
+                          pr_err("Field %d-%d too large for 64 bits!\n",
+                                 startbit, endbit);
+                  else
+                          pr_err("Cannot store %llx inside bits %d-%d (would truncate)\n",
+                                 *val, startbit, endbit);
+          }
+          dump_stack();
+  }
+
+pack() and unpack()
+^^^^^^^^^^^^^^^^^^^
+
+These are const-correct variants of packing(), and eliminate the last "enum
+packing_op op" argument.
+
+Calling pack(...) is equivalent, and preferred, to calling packing(..., PACK).
+
+Calling unpack(...) is equivalent, and preferred, to calling packing(..., UNPACK).
+
+pack_fields() and unpack_fields()
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The library exposes optimized functions for the scenario where there are many
+fields represented in a buffer, and it encourages consumer drivers to avoid
+repetitive calls to pack() and unpack() for each field, but instead use
+pack_fields() and unpack_fields(), which reduces the code footprint.
+
+These APIs use field definitions in arrays of ``struct packed_field_s`` (small)
+or ``struct packed_field_m`` (medium), allowing consumer drivers to minimize
+the size of these arrays according to their custom requirements.
+
+The pack_fields() and unpack_fields() API functions are actually macros which
+automatically select the appropriate function at compile time, based on the
+type of the fields array passed in.
+
+An additional benefit over pack() and unpack() is that sanity checks on the
+field definitions are handled at compile time with ``BUILD_BUG_ON`` rather
+than only when the offending code is executed. These functions return void and
+wrapping them to handle unexpected errors is not necessary.
+
+It is recommended, but not required, that you wrap your packed buffer into a
+structured type with a fixed size. This generally makes it easier for the
+compiler to enforce that the correct size buffer is used.
+
+Here is an example of how to use the fields APIs:
+
+.. code-block:: c
+
+   /* Ordering inside the unpacked structure is flexible and can be different
+    * from the packed buffer. Here, it is optimized to reduce padding.
+    */
+   struct data {
+        u64 field3;
+        u32 field4;
+        u16 field1;
+        u8 field2;
+   };
+
+   #define SIZE 13
+
+   typdef struct __packed { u8 buf[SIZE]; } packed_buf_t;
+
+   static const struct packed_field_s fields[] = {
+           PACKED_FIELD(100, 90, struct data, field1),
+           PACKED_FIELD(90, 87, struct data, field2),
+           PACKED_FIELD(86, 30, struct data, field3),
+           PACKED_FIELD(29, 0, struct data, field4),
+   };
+
+   void unpack_your_data(const packed_buf_t *buf, struct data *unpacked)
+   {
+           BUILD_BUG_ON(sizeof(*buf) != SIZE;
+
+           unpack_fields(buf, sizeof(*buf), unpacked, fields,
+                         QUIRK_LITTLE_ENDIAN);
+   }
+
+   void pack_your_data(const struct data *unpacked, packed_buf_t *buf)
+   {
+           BUILD_BUG_ON(sizeof(*buf) != SIZE;
+
+           pack_fields(buf, sizeof(*buf), unpacked, fields,
+                       QUIRK_LITTLE_ENDIAN);
+   }

-- 
2.47.0.265.g4ca455297942

[PATCH net-next v9 05/10] ice: remove int_q_state from ice_tlan_ctx

[Jacob Keller]

The int_q_state field of the ice_tlan_ctx structure represents the internal
queue state. However, we never actually need to assign this or read this
during normal operation. In fact, trying to unpack it would not be possible
as it is larger than a u64. Remove this field from the ice_tlan_ctx
structure, and remove its packing field from the ice_tlan_ctx_info array.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h | 1 -
 drivers/net/ethernet/intel/ice/ice_common.c    | 1 -
 2 files changed, 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
index 611577ebc29d8250c8cce85f58f3477ff3b51a66..0e8ed8c226e68988664d64c1fd3297cee32af020 100644
--- a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
+++ b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
@@ -590,7 +590,6 @@ struct ice_tlan_ctx {
 	u8 drop_ena;
 	u8 cache_prof_idx;
 	u8 pkt_shaper_prof_idx;
-	u8 int_q_state;	/* width not needed - internal - DO NOT WRITE!!! */
 };
 
 #endif /* _ICE_LAN_TX_RX_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index b22e71dc59d4e4ec0efea96e5afd812859a98bdd..0f5a80269a7be0a302d4229a42bb8bbfc500905a 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -1467,7 +1467,6 @@ const struct ice_ctx_ele ice_tlan_ctx_info[] = {
 	ICE_CTX_STORE(ice_tlan_ctx, drop_ena,			1,	165),
 	ICE_CTX_STORE(ice_tlan_ctx, cache_prof_idx,		2,	166),
 	ICE_CTX_STORE(ice_tlan_ctx, pkt_shaper_prof_idx,	3,	168),
-	ICE_CTX_STORE(ice_tlan_ctx, int_q_state,		122,	171),
 	{ 0 }
 };
 

-- 
2.47.0.265.g4ca455297942

[PATCH net-next v9 06/10] ice: use structures to keep track of queue context size

[Jacob Keller]

The ice Tx and Rx queue context are currently stored as arrays of bytes
with defined size (ICE_RXQ_CTX_SZ and ICE_TXQ_CTX_SZ). The packed queue
context is often passed to other functions as a simple u8 * pointer, which
does not allow tracking the size. This makes the queue context API easy to
misuse, as you can pass an arbitrary u8 array or pointer.

Introduce wrapper typedefs which use a __packed structure that has the
proper fixed size for the Tx and Rx context buffers. This enables the
compiler to track the size of the value and ensures that passing the wrong
buffer size will be detected by the compiler.

The existing APIs do not benefit much from this change, however the
wrapping structures will be used to simplify the arguments of new packing
functions based on the recently introduced pack_fields API.

Co-developed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_adminq_cmd.h | 11 +++++++++--
 drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h  |  2 --
 drivers/net/ethernet/intel/ice/ice_base.c       |  2 +-
 drivers/net/ethernet/intel/ice/ice_common.c     | 24 +++++++++++-------------
 4 files changed, 21 insertions(+), 18 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_adminq_cmd.h b/drivers/net/ethernet/intel/ice/ice_adminq_cmd.h
index 1489a8ceec51df890f481f7fdc04b1845ca85255..3bf05b135b3557e5867ef037d02621a59dc251d4 100644
--- a/drivers/net/ethernet/intel/ice/ice_adminq_cmd.h
+++ b/drivers/net/ethernet/intel/ice/ice_adminq_cmd.h
@@ -12,6 +12,13 @@
 #define ICE_AQC_TOPO_MAX_LEVEL_NUM	0x9
 #define ICE_AQ_SET_MAC_FRAME_SIZE_MAX	9728
 
+#define ICE_RXQ_CTX_SIZE_DWORDS		8
+#define ICE_RXQ_CTX_SZ			(ICE_RXQ_CTX_SIZE_DWORDS * sizeof(u32))
+#define ICE_TXQ_CTX_SZ			22
+
+typedef struct __packed { u8 buf[ICE_RXQ_CTX_SZ]; } ice_rxq_ctx_buf_t;
+typedef struct __packed { u8 buf[ICE_TXQ_CTX_SZ]; } ice_txq_ctx_buf_t;
+
 struct ice_aqc_generic {
 	__le32 param0;
 	__le32 param1;
@@ -2084,10 +2091,10 @@ struct ice_aqc_add_txqs_perq {
 	__le16 txq_id;
 	u8 rsvd[2];
 	__le32 q_teid;
-	u8 txq_ctx[22];
+	ice_txq_ctx_buf_t txq_ctx;
 	u8 rsvd2[2];
 	struct ice_aqc_txsched_elem info;
-};
+} __packed;
 
 /* The format of the command buffer for Add Tx LAN Queues (0x0C30)
  * is an array of the following structs. Please note that the length of
diff --git a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
index 0e8ed8c226e68988664d64c1fd3297cee32af020..a76e5b0e7861e39e59013637cb176f67d1f7ef15 100644
--- a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
+++ b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
@@ -371,8 +371,6 @@ enum ice_rx_flex_desc_status_error_1_bits {
 	ICE_RX_FLEX_DESC_STATUS1_LAST /* this entry must be last!!! */
 };
 
-#define ICE_RXQ_CTX_SIZE_DWORDS		8
-#define ICE_RXQ_CTX_SZ			(ICE_RXQ_CTX_SIZE_DWORDS * sizeof(u32))
 #define ICE_TX_CMPLTNQ_CTX_SIZE_DWORDS	22
 #define ICE_TX_DRBELL_Q_CTX_SIZE_DWORDS	5
 #define GLTCLAN_CQ_CNTX(i, CQ)		(GLTCLAN_CQ_CNTX0(CQ) + ((i) * 0x0800))
diff --git a/drivers/net/ethernet/intel/ice/ice_base.c b/drivers/net/ethernet/intel/ice/ice_base.c
index 82a9cd4ec7aec90febdc7ab31cf8d707314cbd1f..e7aaa06241210e764b4cb627031310e4fd5b6520 100644
--- a/drivers/net/ethernet/intel/ice/ice_base.c
+++ b/drivers/net/ethernet/intel/ice/ice_base.c
@@ -910,7 +910,7 @@ ice_vsi_cfg_txq(struct ice_vsi *vsi, struct ice_tx_ring *ring,
 	ice_setup_tx_ctx(ring, &tlan_ctx, pf_q);
 	/* copy context contents into the qg_buf */
 	qg_buf->txqs[0].txq_id = cpu_to_le16(pf_q);
-	ice_set_ctx(hw, (u8 *)&tlan_ctx, qg_buf->txqs[0].txq_ctx,
+	ice_set_ctx(hw, (u8 *)&tlan_ctx, (u8 *)&qg_buf->txqs[0].txq_ctx,
 		    ice_tlan_ctx_info);
 
 	/* init queue specific tail reg. It is referred as
diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index 0f5a80269a7be0a302d4229a42bb8bbfc500905a..48d95cb49864ad73769487d2b95b2e8306156cf9 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -1359,29 +1359,27 @@ int ice_reset(struct ice_hw *hw, enum ice_reset_req req)
 /**
  * ice_copy_rxq_ctx_to_hw
  * @hw: pointer to the hardware structure
- * @ice_rxq_ctx: pointer to the rxq context
+ * @rxq_ctx: pointer to the packed Rx queue context
  * @rxq_index: the index of the Rx queue
  *
  * Copies rxq context from dense structure to HW register space
  */
-static int
-ice_copy_rxq_ctx_to_hw(struct ice_hw *hw, u8 *ice_rxq_ctx, u32 rxq_index)
+static int ice_copy_rxq_ctx_to_hw(struct ice_hw *hw,
+				  const ice_rxq_ctx_buf_t *rxq_ctx,
+				  u32 rxq_index)
 {
 	u8 i;
 
-	if (!ice_rxq_ctx)
-		return -EINVAL;
-
 	if (rxq_index > QRX_CTRL_MAX_INDEX)
 		return -EINVAL;
 
 	/* Copy each dword separately to HW */
 	for (i = 0; i < ICE_RXQ_CTX_SIZE_DWORDS; i++) {
-		wr32(hw, QRX_CONTEXT(i, rxq_index),
-		     *((u32 *)(ice_rxq_ctx + (i * sizeof(u32)))));
+		u32 ctx = ((const u32 *)rxq_ctx)[i];
 
-		ice_debug(hw, ICE_DBG_QCTX, "qrxdata[%d]: %08X\n", i,
-			  *((u32 *)(ice_rxq_ctx + (i * sizeof(u32)))));
+		wr32(hw, QRX_CONTEXT(i, rxq_index), ctx);
+
+		ice_debug(hw, ICE_DBG_QCTX, "qrxdata[%d]: %08X\n", i, ctx);
 	}
 
 	return 0;
@@ -1426,15 +1424,15 @@ static const struct ice_ctx_ele ice_rlan_ctx_info[] = {
 int ice_write_rxq_ctx(struct ice_hw *hw, struct ice_rlan_ctx *rlan_ctx,
 		      u32 rxq_index)
 {
-	u8 ctx_buf[ICE_RXQ_CTX_SZ] = { 0 };
+	ice_rxq_ctx_buf_t buf = {};
 
 	if (!rlan_ctx)
 		return -EINVAL;
 
 	rlan_ctx->prefena = 1;
 
-	ice_set_ctx(hw, (u8 *)rlan_ctx, ctx_buf, ice_rlan_ctx_info);
-	return ice_copy_rxq_ctx_to_hw(hw, ctx_buf, rxq_index);
+	ice_set_ctx(hw, (u8 *)rlan_ctx, (u8 *)&buf, ice_rlan_ctx_info);
+	return ice_copy_rxq_ctx_to_hw(hw, &buf, rxq_index);
 }
 
 /* LAN Tx Queue Context */

-- 
2.47.0.265.g4ca455297942

[PATCH net-next v9 07/10] ice: use <linux/packing.h> for Tx and Rx queue context data

[Jacob Keller]

The ice driver needs to write the Tx and Rx queue context when programming
Tx and Rx queues. This is currently done using some bespoke custom logic
via the ice_set_ctx() and its helper functions, along with bit position
definitions in the ice_tlan_ctx_info and ice_rlan_ctx_info structures.

This logic does work, but is problematic for several reasons:

1) ice_set_ctx requires a helper function for each byte size being packed,
   as it uses a separate function to pack u8, u16, u32, and u64 fields.
   This requires 4 functions which contain near-duplicate logic with the
   types changed out.

2) The logic in the ice_pack_ctx_word, ice_pack_ctx_dword, and
   ice_pack_ctx_qword does not handle values which straddle alignment
   boundaries very well. This requires that several fields in the
   ice_tlan_ctx_info and ice_rlan_ctx_info be a size larger than their bit
   size should require.

3) Future support for live migration will require adding unpacking
   functions to take the packed hardware context and unpack it into the
   ice_rlan_ctx and ice_tlan_ctx structures. Implementing this would
   require implementing ice_get_ctx, and its associated helper functions,
   which essentially doubles the amount of code required.

The Linux kernel has had a packing library that can handle this logic since
commit 554aae35007e ("lib: Add support for generic packing operations").
The library was recently extended with support for packing or unpacking an
array of fields, with a similar structure as the ice_ctx_ele structure.

Replace the ice-specific ice_set_ctx() logic with the recently added
pack_fields and packed_field_s infrastructure from <linux/packing.h>

For API simplicity, the Tx and Rx queue context are programmed using
separate ice_pack_txq_ctx() and ice_pack_rxq_ctx(). This avoids needing to
export the packed_field_s arrays. The functions can pointers to the
appropriate ice_txq_ctx_buf_t and ice_rxq_ctx_buf_t types, ensuring that
only buffers of the appropriate size are passed.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_common.h    |   5 +-
 drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h |  14 --
 drivers/net/ethernet/intel/ice/ice_base.c      |   3 +-
 drivers/net/ethernet/intel/ice/ice_common.c    | 243 ++++---------------------
 drivers/net/ethernet/intel/Kconfig             |   1 +
 5 files changed, 42 insertions(+), 224 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_common.h b/drivers/net/ethernet/intel/ice/ice_common.h
index 27208a60cece51f49d928a09ec6e9dcbb27fb9b1..a68bea3934e35a270a9809239fdf2f5cdce7b6bd 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.h
+++ b/drivers/net/ethernet/intel/ice/ice_common.h
@@ -92,9 +92,8 @@ ice_aq_set_rss_key(struct ice_hw *hw, u16 vsi_handle,
 bool ice_check_sq_alive(struct ice_hw *hw, struct ice_ctl_q_info *cq);
 int ice_aq_q_shutdown(struct ice_hw *hw, bool unloading);
 void ice_fill_dflt_direct_cmd_desc(struct ice_aq_desc *desc, u16 opcode);
-extern const struct ice_ctx_ele ice_tlan_ctx_info[];
-int ice_set_ctx(struct ice_hw *hw, u8 *src_ctx, u8 *dest_ctx,
-		const struct ice_ctx_ele *ce_info);
+
+void ice_pack_txq_ctx(const struct ice_tlan_ctx *ctx, ice_txq_ctx_buf_t *buf);
 
 extern struct mutex ice_global_cfg_lock_sw;
 
diff --git a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
index a76e5b0e7861e39e59013637cb176f67d1f7ef15..31d4a445d640df21c2aa007ffbd4f2310da264ad 100644
--- a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
+++ b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
@@ -408,20 +408,6 @@ struct ice_rlan_ctx {
 	u8 prefena;	/* NOTE: normally must be set to 1 at init */
 };
 
-struct ice_ctx_ele {
-	u16 offset;
-	u16 size_of;
-	u16 width;
-	u16 lsb;
-};
-
-#define ICE_CTX_STORE(_struct, _ele, _width, _lsb) {	\
-	.offset = offsetof(struct _struct, _ele),	\
-	.size_of = sizeof_field(struct _struct, _ele),	\
-	.width = _width,				\
-	.lsb = _lsb,					\
-}
-
 /* for hsplit_0 field of Rx RLAN context */
 enum ice_rlan_ctx_rx_hsplit_0 {
 	ICE_RLAN_RX_HSPLIT_0_NO_SPLIT		= 0,
diff --git a/drivers/net/ethernet/intel/ice/ice_base.c b/drivers/net/ethernet/intel/ice/ice_base.c
index e7aaa06241210e764b4cb627031310e4fd5b6520..5fe7b5a100202e6f0c33c617c604d45f9487b1f4 100644
--- a/drivers/net/ethernet/intel/ice/ice_base.c
+++ b/drivers/net/ethernet/intel/ice/ice_base.c
@@ -910,8 +910,7 @@ ice_vsi_cfg_txq(struct ice_vsi *vsi, struct ice_tx_ring *ring,
 	ice_setup_tx_ctx(ring, &tlan_ctx, pf_q);
 	/* copy context contents into the qg_buf */
 	qg_buf->txqs[0].txq_id = cpu_to_le16(pf_q);
-	ice_set_ctx(hw, (u8 *)&tlan_ctx, (u8 *)&qg_buf->txqs[0].txq_ctx,
-		    ice_tlan_ctx_info);
+	ice_pack_txq_ctx(&tlan_ctx, &qg_buf->txqs[0].txq_ctx);
 
 	/* init queue specific tail reg. It is referred as
 	 * transmit comm scheduler queue doorbell.
diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index 48d95cb49864ad73769487d2b95b2e8306156cf9..1b013c9c937826633db8cbe29d8e1dc310c7b6f0 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -6,6 +6,7 @@
 #include "ice_adminq_cmd.h"
 #include "ice_flow.h"
 #include "ice_ptp_hw.h"
+#include <linux/packing.h>
 
 #define ICE_PF_RESET_WAIT_COUNT	300
 #define ICE_MAX_NETLIST_SIZE	10
@@ -1385,9 +1386,12 @@ static int ice_copy_rxq_ctx_to_hw(struct ice_hw *hw,
 	return 0;
 }
 
+#define ICE_CTX_STORE(struct_name, struct_field, width, lsb) \
+	PACKED_FIELD((lsb) + (width) - 1, (lsb), struct struct_name, struct_field)
+
 /* LAN Rx Queue Context */
-static const struct ice_ctx_ele ice_rlan_ctx_info[] = {
-	/* Field		Width	LSB */
+static const struct packed_field_s ice_rlan_ctx_fields[] = {
+				 /* Field		Width	LSB */
 	ICE_CTX_STORE(ice_rlan_ctx, head,		13,	0),
 	ICE_CTX_STORE(ice_rlan_ctx, cpuid,		8,	13),
 	ICE_CTX_STORE(ice_rlan_ctx, base,		57,	32),
@@ -1408,9 +1412,23 @@ static const struct ice_ctx_ele ice_rlan_ctx_info[] = {
 	ICE_CTX_STORE(ice_rlan_ctx, tphhead_ena,	1,	196),
 	ICE_CTX_STORE(ice_rlan_ctx, lrxqthresh,		3,	198),
 	ICE_CTX_STORE(ice_rlan_ctx, prefena,		1,	201),
-	{ 0 }
 };
 
+/**
+ * ice_pack_rxq_ctx - Pack Rx queue context into a HW buffer
+ * @ctx: the Rx queue context to pack
+ * @buf: the HW buffer to pack into
+ *
+ * Pack the Rx queue context from the CPU-friendly unpacked buffer into its
+ * bit-packed HW layout.
+ */
+static void ice_pack_rxq_ctx(const struct ice_rlan_ctx *ctx,
+			     ice_rxq_ctx_buf_t *buf)
+{
+	pack_fields(buf, sizeof(*buf), ctx, ice_rlan_ctx_fields,
+		    QUIRK_LITTLE_ENDIAN | QUIRK_LSW32_IS_FIRST);
+}
+
 /**
  * ice_write_rxq_ctx
  * @hw: pointer to the hardware structure
@@ -1431,12 +1449,13 @@ int ice_write_rxq_ctx(struct ice_hw *hw, struct ice_rlan_ctx *rlan_ctx,
 
 	rlan_ctx->prefena = 1;
 
-	ice_set_ctx(hw, (u8 *)rlan_ctx, (u8 *)&buf, ice_rlan_ctx_info);
+	ice_pack_rxq_ctx(rlan_ctx, &buf);
+
 	return ice_copy_rxq_ctx_to_hw(hw, &buf, rxq_index);
 }
 
 /* LAN Tx Queue Context */
-const struct ice_ctx_ele ice_tlan_ctx_info[] = {
+static const struct packed_field_s ice_tlan_ctx_fields[] = {
 				    /* Field			Width	LSB */
 	ICE_CTX_STORE(ice_tlan_ctx, base,			57,	0),
 	ICE_CTX_STORE(ice_tlan_ctx, port_num,			3,	57),
@@ -1465,9 +1484,22 @@ const struct ice_ctx_ele ice_tlan_ctx_info[] = {
 	ICE_CTX_STORE(ice_tlan_ctx, drop_ena,			1,	165),
 	ICE_CTX_STORE(ice_tlan_ctx, cache_prof_idx,		2,	166),
 	ICE_CTX_STORE(ice_tlan_ctx, pkt_shaper_prof_idx,	3,	168),
-	{ 0 }
 };
 
+/**
+ * ice_pack_txq_ctx - Pack Tx queue context into a HW buffer
+ * @ctx: the Tx queue context to pack
+ * @buf: the HW buffer to pack into
+ *
+ * Pack the Tx queue context from the CPU-friendly unpacked buffer into its
+ * bit-packed HW layout.
+ */
+void ice_pack_txq_ctx(const struct ice_tlan_ctx *ctx, ice_txq_ctx_buf_t *buf)
+{
+	pack_fields(buf, sizeof(*buf), ctx, ice_tlan_ctx_fields,
+		    QUIRK_LITTLE_ENDIAN | QUIRK_LSW32_IS_FIRST);
+}
+
 /* Sideband Queue command wrappers */
 
 /**
@@ -4545,205 +4577,6 @@ ice_aq_add_rdma_qsets(struct ice_hw *hw, u8 num_qset_grps,
 
 /* End of FW Admin Queue command wrappers */
 
-/**
- * ice_pack_ctx_byte - write a byte to a packed context structure
- * @src_ctx: unpacked source context structure
- * @dest_ctx: packed destination context data
- * @ce_info: context element description
- */
-static void ice_pack_ctx_byte(u8 *src_ctx, u8 *dest_ctx,
-			      const struct ice_ctx_ele *ce_info)
-{
-	u8 src_byte, dest_byte, mask;
-	u8 *from, *dest;
-	u16 shift_width;
-
-	/* copy from the next struct field */
-	from = src_ctx + ce_info->offset;
-
-	/* prepare the bits and mask */
-	shift_width = ce_info->lsb % 8;
-	mask = GENMASK(ce_info->width - 1 + shift_width, shift_width);
-
-	src_byte = *from;
-	src_byte <<= shift_width;
-	src_byte &= mask;
-
-	/* get the current bits from the target bit string */
-	dest = dest_ctx + (ce_info->lsb / 8);
-
-	memcpy(&dest_byte, dest, sizeof(dest_byte));
-
-	dest_byte &= ~mask;	/* get the bits not changing */
-	dest_byte |= src_byte;	/* add in the new bits */
-
-	/* put it all back */
-	memcpy(dest, &dest_byte, sizeof(dest_byte));
-}
-
-/**
- * ice_pack_ctx_word - write a word to a packed context structure
- * @src_ctx: unpacked source context structure
- * @dest_ctx: packed destination context data
- * @ce_info: context element description
- */
-static void ice_pack_ctx_word(u8 *src_ctx, u8 *dest_ctx,
-			      const struct ice_ctx_ele *ce_info)
-{
-	u16 src_word, mask;
-	__le16 dest_word;
-	u8 *from, *dest;
-	u16 shift_width;
-
-	/* copy from the next struct field */
-	from = src_ctx + ce_info->offset;
-
-	/* prepare the bits and mask */
-	shift_width = ce_info->lsb % 8;
-	mask = GENMASK(ce_info->width - 1 + shift_width, shift_width);
-
-	/* don't swizzle the bits until after the mask because the mask bits
-	 * will be in a different bit position on big endian machines
-	 */
-	src_word = *(u16 *)from;
-	src_word <<= shift_width;
-	src_word &= mask;
-
-	/* get the current bits from the target bit string */
-	dest = dest_ctx + (ce_info->lsb / 8);
-
-	memcpy(&dest_word, dest, sizeof(dest_word));
-
-	dest_word &= ~(cpu_to_le16(mask));	/* get the bits not changing */
-	dest_word |= cpu_to_le16(src_word);	/* add in the new bits */
-
-	/* put it all back */
-	memcpy(dest, &dest_word, sizeof(dest_word));
-}
-
-/**
- * ice_pack_ctx_dword - write a dword to a packed context structure
- * @src_ctx: unpacked source context structure
- * @dest_ctx: packed destination context data
- * @ce_info: context element description
- */
-static void ice_pack_ctx_dword(u8 *src_ctx, u8 *dest_ctx,
-			       const struct ice_ctx_ele *ce_info)
-{
-	u32 src_dword, mask;
-	__le32 dest_dword;
-	u8 *from, *dest;
-	u16 shift_width;
-
-	/* copy from the next struct field */
-	from = src_ctx + ce_info->offset;
-
-	/* prepare the bits and mask */
-	shift_width = ce_info->lsb % 8;
-	mask = GENMASK(ce_info->width - 1 + shift_width, shift_width);
-
-	/* don't swizzle the bits until after the mask because the mask bits
-	 * will be in a different bit position on big endian machines
-	 */
-	src_dword = *(u32 *)from;
-	src_dword <<= shift_width;
-	src_dword &= mask;
-
-	/* get the current bits from the target bit string */
-	dest = dest_ctx + (ce_info->lsb / 8);
-
-	memcpy(&dest_dword, dest, sizeof(dest_dword));
-
-	dest_dword &= ~(cpu_to_le32(mask));	/* get the bits not changing */
-	dest_dword |= cpu_to_le32(src_dword);	/* add in the new bits */
-
-	/* put it all back */
-	memcpy(dest, &dest_dword, sizeof(dest_dword));
-}
-
-/**
- * ice_pack_ctx_qword - write a qword to a packed context structure
- * @src_ctx: unpacked source context structure
- * @dest_ctx: packed destination context data
- * @ce_info: context element description
- */
-static void ice_pack_ctx_qword(u8 *src_ctx, u8 *dest_ctx,
-			       const struct ice_ctx_ele *ce_info)
-{
-	u64 src_qword, mask;
-	__le64 dest_qword;
-	u8 *from, *dest;
-	u16 shift_width;
-
-	/* copy from the next struct field */
-	from = src_ctx + ce_info->offset;
-
-	/* prepare the bits and mask */
-	shift_width = ce_info->lsb % 8;
-	mask = GENMASK_ULL(ce_info->width - 1 + shift_width, shift_width);
-
-	/* don't swizzle the bits until after the mask because the mask bits
-	 * will be in a different bit position on big endian machines
-	 */
-	src_qword = *(u64 *)from;
-	src_qword <<= shift_width;
-	src_qword &= mask;
-
-	/* get the current bits from the target bit string */
-	dest = dest_ctx + (ce_info->lsb / 8);
-
-	memcpy(&dest_qword, dest, sizeof(dest_qword));
-
-	dest_qword &= ~(cpu_to_le64(mask));	/* get the bits not changing */
-	dest_qword |= cpu_to_le64(src_qword);	/* add in the new bits */
-
-	/* put it all back */
-	memcpy(dest, &dest_qword, sizeof(dest_qword));
-}
-
-/**
- * ice_set_ctx - set context bits in packed structure
- * @hw: pointer to the hardware structure
- * @src_ctx:  pointer to a generic non-packed context structure
- * @dest_ctx: pointer to memory for the packed structure
- * @ce_info: List of Rx context elements
- */
-int ice_set_ctx(struct ice_hw *hw, u8 *src_ctx, u8 *dest_ctx,
-		const struct ice_ctx_ele *ce_info)
-{
-	int f;
-
-	for (f = 0; ce_info[f].width; f++) {
-		/* We have to deal with each element of the FW response
-		 * using the correct size so that we are correct regardless
-		 * of the endianness of the machine.
-		 */
-		if (ce_info[f].width > (ce_info[f].size_of * BITS_PER_BYTE)) {
-			ice_debug(hw, ICE_DBG_QCTX, "Field %d width of %d bits larger than size of %d byte(s) ... skipping write\n",
-				  f, ce_info[f].width, ce_info[f].size_of);
-			continue;
-		}
-		switch (ce_info[f].size_of) {
-		case sizeof(u8):
-			ice_pack_ctx_byte(src_ctx, dest_ctx, &ce_info[f]);
-			break;
-		case sizeof(u16):
-			ice_pack_ctx_word(src_ctx, dest_ctx, &ce_info[f]);
-			break;
-		case sizeof(u32):
-			ice_pack_ctx_dword(src_ctx, dest_ctx, &ce_info[f]);
-			break;
-		case sizeof(u64):
-			ice_pack_ctx_qword(src_ctx, dest_ctx, &ce_info[f]);
-			break;
-		default:
-			return -EINVAL;
-		}
-	}
-
-	return 0;
-}
-
 /**
  * ice_get_lan_q_ctx - get the LAN queue context for the given VSI and TC
  * @hw: pointer to the HW struct
diff --git a/drivers/net/ethernet/intel/Kconfig b/drivers/net/ethernet/intel/Kconfig
index 20bc40eec487a3449ff1ac7432b2b029cb2b1155..24ec9a4f1ffa88ef550d4a552f6cd7a2f84dd83a 100644
--- a/drivers/net/ethernet/intel/Kconfig
+++ b/drivers/net/ethernet/intel/Kconfig
@@ -292,6 +292,7 @@ config ICE
 	select DIMLIB
 	select LIBIE
 	select NET_DEVLINK
+	select PACKING
 	select PLDMFW
 	select DPLL
 	help

-- 
2.47.0.265.g4ca455297942

[PATCH net-next v9 08/10] ice: reduce size of queue context fields

[Jacob Keller]

The ice_rlan_ctx and ice_tlan_ctx structures have some fields which are
intentionally sized larger than necessary relative to the packed sizes the
data must fit into. This was done because the original ice_set_ctx()
function and its helpers did not correctly handle packing when the packed
bits straddled a byte. This is no longer the case with the use of the
<linux/packing.h> implementation.

Save some bytes in these structures by sizing the variables to the number
of bytes the actual bitpacked fields fit into.

There are a couple of gaps left in the structure, which is a result of the
fields being in the order they appear in the packed bit layout, but where
alignment forces some extra gaps. We could fix this, saving ~8 bytes from
each structure. However, these structures are not used heavily, and the
resulting savings is minimal:

$ bloat-o-meter ice-before-reorder.ko ice-after-reorder.ko
add/remove: 0/0 grow/shrink: 1/1 up/down: 26/-70 (-44)
Function                                     old     new   delta
ice_vsi_cfg_txq                             1873    1899     +26
ice_setup_rx_ctx.constprop                  1529    1459     -70
Total: Before=1459555, After=1459511, chg -0.00%

Thus, the fields are left in the same order as the packed bit layout,
despite the gaps this causes.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h | 32 ++++++++------------------
 1 file changed, 10 insertions(+), 22 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
index 31d4a445d640df21c2aa007ffbd4f2310da264ad..1479b45738af15bf6e00aed24b2c6a3f91675f4d 100644
--- a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
+++ b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
@@ -375,23 +375,17 @@ enum ice_rx_flex_desc_status_error_1_bits {
 #define ICE_TX_DRBELL_Q_CTX_SIZE_DWORDS	5
 #define GLTCLAN_CQ_CNTX(i, CQ)		(GLTCLAN_CQ_CNTX0(CQ) + ((i) * 0x0800))
 
-/* RLAN Rx queue context data
- *
- * The sizes of the variables may be larger than needed due to crossing byte
- * boundaries. If we do not have the width of the variable set to the correct
- * size then we could end up shifting bits off the top of the variable when the
- * variable is at the top of a byte and crosses over into the next byte.
- */
+/* RLAN Rx queue context data */
 struct ice_rlan_ctx {
 	u16 head;
-	u16 cpuid; /* bigger than needed, see above for reason */
+	u8 cpuid;
 #define ICE_RLAN_BASE_S 7
 	u64 base;
 	u16 qlen;
 #define ICE_RLAN_CTX_DBUF_S 7
-	u16 dbuf; /* bigger than needed, see above for reason */
+	u8 dbuf;
 #define ICE_RLAN_CTX_HBUF_S 6
-	u16 hbuf; /* bigger than needed, see above for reason */
+	u8 hbuf;
 	u8 dtype;
 	u8 dsize;
 	u8 crcstrip;
@@ -399,12 +393,12 @@ struct ice_rlan_ctx {
 	u8 hsplit_0;
 	u8 hsplit_1;
 	u8 showiv;
-	u32 rxmax; /* bigger than needed, see above for reason */
+	u16 rxmax;
 	u8 tphrdesc_ena;
 	u8 tphwdesc_ena;
 	u8 tphdata_ena;
 	u8 tphhead_ena;
-	u16 lrxqthresh; /* bigger than needed, see above for reason */
+	u8 lrxqthresh;
 	u8 prefena;	/* NOTE: normally must be set to 1 at init */
 };
 
@@ -535,18 +529,12 @@ enum ice_tx_ctx_desc_eipt_offload {
 #define ICE_LAN_TXQ_MAX_QGRPS	127
 #define ICE_LAN_TXQ_MAX_QDIS	1023
 
-/* Tx queue context data
- *
- * The sizes of the variables may be larger than needed due to crossing byte
- * boundaries. If we do not have the width of the variable set to the correct
- * size then we could end up shifting bits off the top of the variable when the
- * variable is at the top of a byte and crosses over into the next byte.
- */
+/* Tx queue context data */
 struct ice_tlan_ctx {
 #define ICE_TLAN_CTX_BASE_S	7
 	u64 base;		/* base is defined in 128-byte units */
 	u8 port_num;
-	u16 cgd_num;		/* bigger than needed, see above for reason */
+	u8 cgd_num;
 	u8 pf_num;
 	u16 vmvf_num;
 	u8 vmvf_type;
@@ -557,7 +545,7 @@ struct ice_tlan_ctx {
 	u8 tsyn_ena;
 	u8 internal_usage_flag;
 	u8 alt_vlan;
-	u16 cpuid;		/* bigger than needed, see above for reason */
+	u8 cpuid;
 	u8 wb_mode;
 	u8 tphrd_desc;
 	u8 tphrd;
@@ -566,7 +554,7 @@ struct ice_tlan_ctx {
 	u16 qnum_in_func;
 	u8 itr_notification_mode;
 	u8 adjust_prof_id;
-	u32 qlen;		/* bigger than needed, see above for reason */
+	u16 qlen;
 	u8 quanta_prof_idx;
 	u8 tso_ena;
 	u16 tso_qnum;

-- 
2.47.0.265.g4ca455297942

[PATCH net-next v9 09/10] ice: move prefetch enable to ice_setup_rx_ctx

[Jacob Keller]

The ice_write_rxq_ctx() function is responsible for programming the Rx
Queue context into hardware. It receives the configuration in unpacked form
via the ice_rlan_ctx structure.

This function unconditionally modifies the context to set the prefetch
enable bit. This was done by commit c31a5c25bb19 ("ice: Always set prefena
when configuring an Rx queue"). Setting this bit makes sense, since
prefetching descriptors is almost always the preferred behavior.

However, the ice_write_rxq_ctx() function is not the place that actually
defines the queue context. We initialize the Rx Queue context in
ice_setup_rx_ctx(). It is surprising to have the Rx queue context changed
by a function who's responsibility is to program the given context to
hardware.

Following the principle of least surprise, move the setting of the prefetch
enable bit out of ice_write_rxq_ctx() and into the ice_setup_rx_ctx().

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_base.c   | 3 +++
 drivers/net/ethernet/intel/ice/ice_common.c | 9 +++------
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_base.c b/drivers/net/ethernet/intel/ice/ice_base.c
index 5fe7b5a100202e6f0c33c617c604d45f9487b1f4..b2af8e3586f7620d372f2055e337485d102d3cbc 100644
--- a/drivers/net/ethernet/intel/ice/ice_base.c
+++ b/drivers/net/ethernet/intel/ice/ice_base.c
@@ -454,6 +454,9 @@ static int ice_setup_rx_ctx(struct ice_rx_ring *ring)
 	/* Rx queue threshold in units of 64 */
 	rlan_ctx.lrxqthresh = 1;
 
+	/* Enable descriptor prefetch */
+	rlan_ctx.prefena = 1;
+
 	/* PF acts as uplink for switchdev; set flex descriptor with src_vsi
 	 * metadata and flags to allow redirecting to PR netdev
 	 */
diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index 1b013c9c937826633db8cbe29d8e1dc310c7b6f0..379040593d975342eaa2a3032938683b419f4f60 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -1430,14 +1430,13 @@ static void ice_pack_rxq_ctx(const struct ice_rlan_ctx *ctx,
 }
 
 /**
- * ice_write_rxq_ctx
+ * ice_write_rxq_ctx - Write Rx Queue context to hardware
  * @hw: pointer to the hardware structure
  * @rlan_ctx: pointer to the rxq context
  * @rxq_index: the index of the Rx queue
  *
- * Converts rxq context from sparse to dense structure and then writes
- * it to HW register space and enables the hardware to prefetch descriptors
- * instead of only fetching them on demand
+ * Pack the sparse Rx Queue context into dense hardware format and write it
+ * into the HW register space.
  */
 int ice_write_rxq_ctx(struct ice_hw *hw, struct ice_rlan_ctx *rlan_ctx,
 		      u32 rxq_index)
@@ -1447,8 +1446,6 @@ int ice_write_rxq_ctx(struct ice_hw *hw, struct ice_rlan_ctx *rlan_ctx,
 	if (!rlan_ctx)
 		return -EINVAL;
 
-	rlan_ctx->prefena = 1;
-
 	ice_pack_rxq_ctx(rlan_ctx, &buf);
 
 	return ice_copy_rxq_ctx_to_hw(hw, &buf, rxq_index);

-- 
2.47.0.265.g4ca455297942

[PATCH net-next v9 10/10] ice: cleanup Rx queue context programming functions

[Jacob Keller]

The ice_copy_rxq_ctx_to_hw() and ice_write_rxq_ctx() functions perform some
defensive checks which are typically frowned upon by kernel style
guidelines.

In particular, NULL checks on buffers which point to the stack are
discouraged, especially when the functions are static and only called once.
Checks of this sort only serve to hide potential programming error, as we
will not produce the normal crash dump on a NULL access.

In addition, ice_copy_rxq_ctx_to_hw() cannot fail in another way, so could
be made void.

Future support for VF Live Migration will need to introduce an inverse
function for reading Rx queue context from HW registers to unpack it, as
well as functions to pack and unpack Tx queue context from HW.

Rather than copying these style issues into the new functions, lets first
cleanup the existing code.

For the ice_copy_rxq_ctx_to_hw() function:

 * Move the Rx queue index check out of this function.
 * Convert the function to a void return.
 * Use a simple int variable instead of a u8 for the for loop index, and
   initialize it inside the for loop.
 * Update the function description to better align with kernel doc style.

For the ice_write_rxq_ctx() function:

 * Move the Rx queue index check into this function.
 * Update the function description with a Returns: to align with kernel doc
   style.

These changes align the existing write functions to current kernel
style, and will align with the style of the new functions added when we
implement live migration in a future series.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_common.c | 28 +++++++++++-----------------
 1 file changed, 11 insertions(+), 17 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index 379040593d975342eaa2a3032938683b419f4f60..6c6862beab6a961ce5c0bc34e9c5794ed8cda865 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -1358,32 +1358,23 @@ int ice_reset(struct ice_hw *hw, enum ice_reset_req req)
 }
 
 /**
- * ice_copy_rxq_ctx_to_hw
+ * ice_copy_rxq_ctx_to_hw - Copy packed Rx queue context to HW registers
  * @hw: pointer to the hardware structure
  * @rxq_ctx: pointer to the packed Rx queue context
  * @rxq_index: the index of the Rx queue
- *
- * Copies rxq context from dense structure to HW register space
  */
-static int ice_copy_rxq_ctx_to_hw(struct ice_hw *hw,
-				  const ice_rxq_ctx_buf_t *rxq_ctx,
-				  u32 rxq_index)
+static void ice_copy_rxq_ctx_to_hw(struct ice_hw *hw,
+				   const ice_rxq_ctx_buf_t *rxq_ctx,
+				   u32 rxq_index)
 {
-	u8 i;
-
-	if (rxq_index > QRX_CTRL_MAX_INDEX)
-		return -EINVAL;
-
 	/* Copy each dword separately to HW */
-	for (i = 0; i < ICE_RXQ_CTX_SIZE_DWORDS; i++) {
+	for (int i = 0; i < ICE_RXQ_CTX_SIZE_DWORDS; i++) {
 		u32 ctx = ((const u32 *)rxq_ctx)[i];
 
 		wr32(hw, QRX_CONTEXT(i, rxq_index), ctx);
 
 		ice_debug(hw, ICE_DBG_QCTX, "qrxdata[%d]: %08X\n", i, ctx);
 	}
-
-	return 0;
 }
 
 #define ICE_CTX_STORE(struct_name, struct_field, width, lsb) \
@@ -1432,23 +1423,26 @@ static void ice_pack_rxq_ctx(const struct ice_rlan_ctx *ctx,
 /**
  * ice_write_rxq_ctx - Write Rx Queue context to hardware
  * @hw: pointer to the hardware structure
- * @rlan_ctx: pointer to the rxq context
+ * @rlan_ctx: pointer to the unpacked Rx queue context
  * @rxq_index: the index of the Rx queue
  *
  * Pack the sparse Rx Queue context into dense hardware format and write it
  * into the HW register space.
+ *
+ * Return: 0 on success, or -EINVAL if the Rx queue index is invalid.
  */
 int ice_write_rxq_ctx(struct ice_hw *hw, struct ice_rlan_ctx *rlan_ctx,
 		      u32 rxq_index)
 {
 	ice_rxq_ctx_buf_t buf = {};
 
-	if (!rlan_ctx)
+	if (rxq_index > QRX_CTRL_MAX_INDEX)
 		return -EINVAL;
 
 	ice_pack_rxq_ctx(rlan_ctx, &buf);
+	ice_copy_rxq_ctx_to_hw(hw, &buf, rxq_index);
 
-	return ice_copy_rxq_ctx_to_hw(hw, &buf, rxq_index);
+	return 0;
 }
 
 /* LAN Tx Queue Context */

-- 
2.47.0.265.g4ca455297942

Re: [PATCH net-next v9 03/10] lib: packing: add pack_fields() and unpack_fields()

[Vladimir Oltean]

On Wed, Dec 04, 2024 at 05:22:49PM -0800, Jacob Keller wrote:
> From: Vladimir Oltean <vladimir.oltean@nxp.com>
> 
> This is new API which caters to the following requirements:
> 
> - Pack or unpack a large number of fields to/from a buffer with a small
>   code footprint. The current alternative is to open-code a large number
>   of calls to pack() and unpack(), or to use packing() to reduce that
>   number to half. But packing() is not const-correct.
> 
> - Use unpacked numbers stored in variables smaller than u64. This
>   reduces the rodata footprint of the stored field arrays.
> 
> - Perform error checking at compile time, rather than runtime, and return
>   void from the API functions. Because the C preprocessor can't generate
>   variable length code (loops), this is a bit tricky to do with macros.
> 
>   To handle this, implement macros which sanity check the packed field
>   definitions based on their size. Finally, a single macro with a chain of
>   __builtin_choose_expr() is used to select the appropriate macros. We
>   enforce the use of ascending or descending order to avoid O(N^2) scaling
>   when checking for overlap. Note that the macros are written with care to
>   ensure that the compilers can correctly evaluate the resulting code at
>   compile time. In particular, care was taken with avoiding too many nested
>   statement expressions. Nested statement expressions trip up some
>   compilers, especially when passing down variables created in previous
>   statement expressions.
> 
>   There are two key design choices intended to keep the overall macro code
>   size small. First, the definition of each CHECK_PACKED_FIELDS_N macro is
>   implemented recursively, by calling the N-1 macro. This avoids needing
>   the code to repeat multiple times.
> 
>   Second, the CHECK_PACKED_FIELD macro enforces that the fields in the
>   array are sorted in order. This allows checking for overlap only with
>   neighboring fields, rather than the general overlap case where each field
>   would need to be checked against other fields.
> 
>   The overlap checks use the first two fields to determine the order of the
>   remaining fields, thus allowing either ascending or descending order.
>   This enables drivers the flexibility to keep the fields ordered in which
>   ever order most naturally fits their hardware design and its associated
>   documentation.
> 
>   The CHECK_PACKED_FIELDS macro is directly called from within pack_fields
>   and unpack_fields, ensuring that all drivers using the API receive the
>   benefits of the compile-time checks. Users do not need to directly call
>   any of the macros directly.
> 
>   The CHECK_PACKED_FIELDS and its helper macros CHECK_PACKED_FIELDS_(0..50)
>   are generated using a simple C program in scripts/gen_packed_field_checks.c
>   This program can be compiled on demand and executed to generate the macro
>   code in include/linux/packing.h. This will aid in the event that a driver
>   needs more than 50 fields. The generator can be updated with a new size,
>   and used to update the packing.h header file. In practice, the ice driver
>   will need to support 27 fields, and the sja1105 driver will need to
>   support 40 fields. This on-demand generation avoids the need to modify
>   Kbuild. We do not anticipate the maximum number of fields to grow very
>   often.
> 
> - Reduced rodata footprint for the storage of the packed field arrays.
>   To that end, we have struct packed_field_s (small) and packed_field_m
>   (medium). More can be added as needed (unlikely for now). On these
>   types, the same generic pack_fields() and unpack_fields() API can be
>   used, thanks to the new C11 _Generic() selection feature, which can
>   call pack_fields_s() or pack_fields_m(), depending on the type of the
>   "fields" array - a simplistic form of polymorphism. It is evaluated at
>   compile time which function will actually be called.
> 
> Over time, packing() is expected to be completely replaced either with
> pack() or with pack_fields().
> 
> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
> Co-developed-by: Jacob Keller <jacob.e.keller@intel.com>
> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
> ---

Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com> # KUnit

Re: [PATCH net-next v9 00/10] lib: packing: introduce and use (un)pack_fields

[Vladimir Oltean]

On Wed, Dec 04, 2024 at 05:22:46PM -0800, Jacob Keller wrote:
> This series improves the packing library with a new API for packing or
> unpacking a large number of fields at once with minimal code footprint. The
> API is then used to replace bespoke packing logic in the ice driver,
> preparing it to handle unpacking in the future. Finally, the ice driver has
> a few other cleanups related to the packing logic.
> 
> The pack_fields and unpack_fields functions have the following improvements
> over the existing pack() and unpack() API:
> 
>  1. Packing or unpacking a large number of fields takes significantly less
>     code. This significantly reduces the .text size for an increase in the
>     .data size which is much smaller.
> 
>  2. The unpacked data can be stored in sizes smaller than u64 variables.
>     This reduces the storage requirement both for runtime data structures,
>     and for the rodata defining the fields. This scales with the number of
>     fields used.
> 
>  3. Most of the error checking is done at compile time, rather than
>     runtime, via CHECK_PACKED_FIELD macros.
> 
> The actual packing and unpacking code still uses the u64 size
> variables. However, these are converted to the appropriate field sizes when
> storing or reading the data from the buffer.

For the set:

Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>

Thank you!

Re: [PATCH net-next v9 03/10] lib: packing: add pack_fields() and unpack_fields()

[Jakub Kicinski]

On Wed, 04 Dec 2024 17:22:49 -0800 Jacob Keller wrote:
> +PHONY += scripts_gen_packed_field_checks
> +scripts_gen_packed_field_checks: scripts_basic
> +	$(Q)$(MAKE) $(build)=scripts scripts/gen_packed_field_checks

You need to add this binary to .gitignore, one more round :(
The rest LGTM

> +/* Small packed field. Use with bit offsets < 256, buffers < 32B and
> + * unpacked structures < 256B.
> + */
> +struct packed_field_s {
> +	GEN_PACKED_FIELD_MEMBERS(u8);
> +};
> +
> +/* Medium packed field. Use with bit offsets < 65536, buffers < 8KB and
> + * unpacked structures < 64KB.
> + */
> +struct packed_field_m {
> +	GEN_PACKED_FIELD_MEMBERS(u16);
> +};

Random thought - would it be more intuitive to use the same size
suffixes as readX() / writeX()? b = byte, w = u16, l = u32, q = 64? 
If you're immediate reaction isn't "of course!" -- ignore me.

Re: [PATCH net-next v9 03/10] lib: packing: add pack_fields() and unpack_fields()

[Jacob Keller]

On 12/9/2024 2:18 PM, Jakub Kicinski wrote:
> On Wed, 04 Dec 2024 17:22:49 -0800 Jacob Keller wrote:
>> +PHONY += scripts_gen_packed_field_checks
>> +scripts_gen_packed_field_checks: scripts_basic
>> +	$(Q)$(MAKE) $(build)=scripts scripts/gen_packed_field_checks
> 
> You need to add this binary to .gitignore, one more round :(
> The rest LGTM
> 

At least its an easy enough fix.

>> +/* Small packed field. Use with bit offsets < 256, buffers < 32B and
>> + * unpacked structures < 256B.
>> + */
>> +struct packed_field_s {
>> +	GEN_PACKED_FIELD_MEMBERS(u8);
>> +};
>> +
>> +/* Medium packed field. Use with bit offsets < 65536, buffers < 8KB and
>> + * unpacked structures < 64KB.
>> + */
>> +struct packed_field_m {
>> +	GEN_PACKED_FIELD_MEMBERS(u16);
>> +};
> 
> Random thought - would it be more intuitive to use the same size
> suffixes as readX() / writeX()? b = byte, w = u16, l = u32, q = 64? 
> If you're immediate reaction isn't "of course!" -- ignore me.

Its fine with me, but Vladimir was the one to change them from numbers
(packed_field_8 to packed_field_s and packed_field_16 to packed_field_m).

@Vladimir, thoughts on using the byte/word suffixes over "small/medium"?

I'll work on preparing v10 with the git ignore fix, but will wait a bit
before sending to get feedback here.

Thanks,
Jake

Re: [PATCH net-next v9 03/10] lib: packing: add pack_fields() and unpack_fields()

[Vladimir Oltean]

On Mon, Dec 09, 2024 at 03:05:54PM -0800, Jacob Keller wrote:
> On 12/9/2024 2:18 PM, Jakub Kicinski wrote:
> > On Wed, 04 Dec 2024 17:22:49 -0800 Jacob Keller wrote:
> >> +/* Small packed field. Use with bit offsets < 256, buffers < 32B and
> >> + * unpacked structures < 256B.
> >> + */
> >> +struct packed_field_s {
> >> +	GEN_PACKED_FIELD_MEMBERS(u8);
> >> +};
> >> +
> >> +/* Medium packed field. Use with bit offsets < 65536, buffers < 8KB and
> >> + * unpacked structures < 64KB.
> >> + */
> >> +struct packed_field_m {
> >> +	GEN_PACKED_FIELD_MEMBERS(u16);
> >> +};
> > 
> > Random thought - would it be more intuitive to use the same size
> > suffixes as readX() / writeX()? b = byte, w = u16, l = u32, q = 64? 
> > If you're immediate reaction isn't "of course!" -- ignore me.
> 
> Its fine with me, but Vladimir was the one to change them from numbers
> (packed_field_8 to packed_field_s and packed_field_16 to packed_field_m).

That was to avoid confusion with the numbers in CHECK_PACKED_FIELDS_8(),
which meant something completely different (array length).

> @Vladimir, thoughts on using the byte/word suffixes over "small/medium"?
> 
> I'll work on preparing v10 with the git ignore fix, but will wait a bit
> before sending to get feedback here.

If you both think it is more intuitive to have struct packed_field_b,
packed_field_w etc, then so be it, it's just a name. I'm not too
attached to the current scheme either, and I do agree that "small" and
"medium" have burger connotations :(

Re: [PATCH net-next v9 03/10] lib: packing: add pack_fields() and unpack_fields()

[Jacob Keller]

On 12/10/2024 2:59 AM, Vladimir Oltean wrote:
> On Mon, Dec 09, 2024 at 03:05:54PM -0800, Jacob Keller wrote:
>> On 12/9/2024 2:18 PM, Jakub Kicinski wrote:
>>> On Wed, 04 Dec 2024 17:22:49 -0800 Jacob Keller wrote:
>>>> +/* Small packed field. Use with bit offsets < 256, buffers < 32B and
>>>> + * unpacked structures < 256B.
>>>> + */
>>>> +struct packed_field_s {
>>>> +	GEN_PACKED_FIELD_MEMBERS(u8);
>>>> +};
>>>> +
>>>> +/* Medium packed field. Use with bit offsets < 65536, buffers < 8KB and
>>>> + * unpacked structures < 64KB.
>>>> + */
>>>> +struct packed_field_m {
>>>> +	GEN_PACKED_FIELD_MEMBERS(u16);
>>>> +};
>>>
>>> Random thought - would it be more intuitive to use the same size
>>> suffixes as readX() / writeX()? b = byte, w = u16, l = u32, q = 64? 
>>> If you're immediate reaction isn't "of course!" -- ignore me.
>>
>> Its fine with me, but Vladimir was the one to change them from numbers
>> (packed_field_8 to packed_field_s and packed_field_16 to packed_field_m).
> 
> That was to avoid confusion with the numbers in CHECK_PACKED_FIELDS_8(),
> which meant something completely different (array length).
> 
>> @Vladimir, thoughts on using the byte/word suffixes over "small/medium"?
>>
>> I'll work on preparing v10 with the git ignore fix, but will wait a bit
>> before sending to get feedback here.
> 
> If you both think it is more intuitive to have struct packed_field_b,
> packed_field_w etc, then so be it, it's just a name. I'm not too
> attached to the current scheme either, and I do agree that "small" and
> "medium" have burger connotations :(

I opted to go with "packed_field_u8" and "packed_field_u16" since I
believe that makes it obvious these are different from the length of the
array itself.

I'll have v10 out soon!