[cocci] Searching inappropriate readl specifications with SmPL?

Markus Elfring <[email protected]> Sun, 7 Jun 2026 21:30:27 +0200
Newsgroups fr.inria.cocci
Message-ID <[email protected]>
Hello,

Felix Gu pointed another software correction out.

I assumed that something can be achieved also by the means of the semantic
patch language accordingly.


// See also:
// [PATCH] spi: rzv2h-rspi: fix incorrect readl() accessor for 8-bit RX pa=
th
// https://lore.kernel.org/linux-spi/20260605-rzv2h-rspi-v1-1-c1d5ed562249=
@gmail.com/

// https://elixir.bootlin.com/linux/v7.1-rc6/source/drivers/spi/spi-rzv2h-=
rspi.c#L140
@display@
identifier action;
typedef u8;
@@
*action(readl, u8)


Questionable test result (in combination with the software =E2=80=9CCoccin=
elle 1.3.1=E2=80=9D):
Markus_Elfring@Sonne:=E2=80=A6/Projekte/Linux/next-analyses> time /usr/bin=
/spatch =E2=80=A6/Projekte/Coccinelle/janitor/show_questionable_readl.cocc=
i drivers/spi/spi-rzv2h-rspi.c
=E2=80=A6
real    0m0,119s
user    0m0,103s
sys     0m0,014s


Why is an expected difference display not generated so far?

Regards,
Markus