Re: [nft PATCH v2 00/11] Fix netlink debug output on Big Endian

[Phil Sutter <phil@nwl.cc>]

On Fri, Nov 14, 2025 at 01:25:31AM +0100, Phil Sutter wrote:
> Make use of recent changes to libnftnl to make test suites pass on both
> Little and Big Endian systems.
> 
> Changes since v1:
> - First 12 patches accepted and pushed already, patches 27 and 28 as
>   well
> - Minimize changes to existing code: Drop patch 13 changing string-based
>   expressions to be defined as Big Endian as well as related patches 14,
>   17 and 18
> - Pull a fix for concatenated wildcard interface names upfront
> - Follow with set element sorting changes, with improved patch
>   descriptions and related test suite record changes folded into them
>   for clarification of practical effect
> - Review patch descriptions in general
> - Add special casing to patch 5 to avoid string-based values being
>   printed in reverse by libnftnl, also communicate any byteorder
>   conversions from __netlink_gen_concat_key() back to caller for the
>   same purpose
> 
> Patch 1 works around "funny" behaviour of GMP when partially exporting
> data and unwanted prefix-padding when exporting into an oversized
> buffer, all happening on Big Endian only.
> 
> Patches 2 and 3 deal with sorting of set elements. They are effective on
> Little Endian only, changing sort ordering to match that of Big Endian.
> 
> Patches 4 and 5 are preparation for the next two patches.
> 
> Patches 6 and 7 collect data for calls to newly introduced libnftnl API
> functions (in patch 8) to communicate byte order and component sizes in
> data regs to libnftnl.
> 
> Patch 10 contains the big payload records update, created with help from
> the script introduced in patch 9.
> 
> Patch 11 still contains the expr_print_debug() macro definition for use
> with printf-debugging.
> 
> Phil Sutter (11):
>   segtree: Fix range aggregation on Big Endian
>   mergesort: Fix sorting of string values
>   mergesort: Align concatenation sort order with Big Endian
>   intervals: Convert byte order implicitly
>   expression: Set range expression 'len' field
>   netlink: Introduce struct nft_data_linearize::byteorder
>   netlink: Introduce struct nft_data_linearize::sizes
>   netlink: Make use of nftnl_{expr,set_elem}_set_imm()
>   tests: py: tools: Add regen_payloads.sh
>   tests: py: Update payload records
>   utils: Introduce expr_print_debug()

Series applied.