Thread (48 messages) 48 messages, 8 authors, 2019-11-12

Re: [PATCH 11/21] asm-generic: don't provide ioremap for CONFIG_MMU

From: Arnd Bergmann <arnd@arndb.de>
Date: 2019-11-11 10:31:28
Also in: linux-arch, linux-arm-kernel, linux-m68k, linux-mips, linux-riscv, linux-s390, linux-sh, lkml, sparclinux

On Wed, Nov 6, 2019 at 7:16 PM Geert Uytterhoeven [off-list ref] wrote:
Hi Palmer,

On Wed, Nov 6, 2019 at 7:11 PM Palmer Dabbelt [off-list ref] wrote:
quoted
It looks like the difference in prototype between the architectures is between

    void __iomem *ioremap(resource_size_t, size_t)
    void __iomem *ioremap(phys_addr_t, size_t)
    void __iomem *ioremap(phys_addr_t, unsigned long)
    void __iomem *ioremap(unsigned long, unsigned long)

shouldn't they all just be that first one?  In other words, wouldn't it be
better to always provide the generic ioremap prototype and unify the ports
instead?
Agreed. But I'd go for the second one.
Right, phys_addr_t is the correct type here, resource_size_t is just a generic
type that is at least as long as any resource, and usually the same as
phys_addr_t, which is supposed to be used for physical addresses.

      Arnd
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help