Re: [PATCH v3 3/7] arm64: HWCAP: encapsulate elf_hwcap

From
Suzuki K Poulose <suzuki.poulose@arm.com>
Date
2019-04-02 15:33:09
Also in
linux-arm-kernel
Reply to
Andrew Murray, 2019-04-02
Hi,

On 02/04/2019 16:06, Andrew Murray wrote:
On Tue, Apr 02, 2019 at 03:58:21PM +0100, Dave Martin wrote:
quoted
On Mon, Apr 01, 2019 at 11:45:11AM +0100, Andrew Murray wrote:
quoted
The introduction of AT_HWCAP2 introduced accessors which ensure that
hwcap features are set and tested appropriately.

Let's now mandate access to elf_hwcap via these accessors by making
elf_hwcap static within cpufeature.c.
Looks reasonable except for a couple of minor nits below.

I had wondered whether putting these accessors out of line would affect
any hot paths, but I can't see these used from anything that looks like
a hot path.  So we're probably fine.

cpus_have_const_cap() is preferred for places where this matters,
anyway.
Btw, thats for cpu_hwcaps, which is completely different from elf_hwcaps.
quoted
[...]
quoted
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 986ceeacd19f..84ca52fa75e5 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -35,8 +35,7 @@
  #include <asm/traps.h>
  #include <asm/virt.h>
  
-unsigned long elf_hwcap __read_mostly;
-EXPORT_SYMBOL_GPL(elf_hwcap);
+static unsigned long elf_hwcap __read_mostly;
Now that this doesn't correspond directly to ELF_HWCAP any more and we
hide it, can we rename it to avoid confusion?

Maybe "kernel_hwcap"?
Yes this seems reasonable.
nit:

As mentioned above we have "cpu_hwcaps" for the features only internally
by the kernel. Naming it "kernel_hwcap" kind of looses the hint that the
major purpose is for userspace consumption and could easily confuse with
the poorly named "cpu_hwcaps" which should have been called kernel_hwcaps.

How about "user_hwcaps" ? Or preferrably something closer to that.

Cheers
Suzuki
Thread (23 messages) 23 messages, 3 authors, last reply 2019-04-03
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help