RE: [PATCH v2] fs/proc: Optimize arrays defined by struct ctl_path
From: David Laight <hidden>
Date: 2022-02-24 11:05:44
Also in:
linux-fscrypt, linux-fsdevel, lkml
From: Meng Tang <redacted>
Sent: 24 February 2022 10:53 Previously, arrays defined by struct ctl_path is terminated with an empty one. For example, when we actually only register one ctl_path, we've gone from 8 bytes to 16 bytes.
...
quoted hunk ↗ jump to hunk
diff --git a/arch/csky/abiv1/alignment.c b/arch/csky/abiv1/alignment.c index 2df115d0e210..b9290a252d84 100644 --- a/arch/csky/abiv1/alignment.c +++ b/arch/csky/abiv1/alignment.c@@ -340,14 +340,14 @@ static struct ctl_table sysctl_table[2] = { {} }; -static struct ctl_path sysctl_path[2] = { - {.procname = "csky"}, - {} +static struct ctl_path sysctl_path[1] = { + {.procname = "csky"} }; +#define SYSCTL_PATH_NUM ARRAY_SIZE(sysctl_path) static int __init csky_alignment_init(void) { - (sysctl_path, sysctl_table); + register_sysctl_paths(sysctl_path, SYSCTL_PATH_NUM, sysctl_table); return 0; }
That is horribly error prone. What might work is to add the ctl_path_num parameter but leave in the check for NULL. Then add: #define register_sysctl_paths(p, t) register_sysctl_paths(p, ARRAY_SIZE(t), t) in the header file after the prototype. Put the function name in () in the definition to stop the macro expansion. David - Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)