[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <b2e202f7-173f-9a26-2858-47768db46d2e@intel.com>
Date: Fri, 31 May 2019 13:53:05 -0700
From: Dave Hansen <dave.hansen@...el.com>
To: Qian Cai <cai@....pw>, mingo@...hat.com, tglx@...utronix.de,
bp@...en8.de
Cc: dave.hansen@...ux.intel.com, luto@...nel.org, peterz@...radead.org,
hpa@...or.com, x86@...nel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] x86/mm: fix an unused variable "tsk" warning
On 5/31/19 1:38 PM, Qian Cai wrote:
> --- a/arch/x86/mm/fault.c
> +++ b/arch/x86/mm/fault.c
> @@ -1015,7 +1015,9 @@ static inline bool bad_area_access_from_pkeys(unsigned long error_code,
> do_sigbus(struct pt_regs *regs, unsigned long error_code, unsigned long address,
> vm_fault_t fault)
> {
> +#ifdef CONFIG_MEMORY_FAILURE
> struct task_struct *tsk = current;
> +#endif
>
> /* Kernel mode? Handle exceptions or die: */
> if (!(error_code & X86_PF_USER)) {
Ick. I'd much rather see this:
#ifdef CONFIG_MEMORY_FAILURE
if (fault & (VM_FAULT_HWPOISON|VM_FAULT_HWPOISON_LARGE)) {
unsigned lsb = 0;
become:
if (IS_ENABLED(CONFIG_MEMORY_FAILURE) &&
(fault & (VM_FAULT_HWPOISON|VM_FAULT_HWPOISON_LARGE))) {
unsigned lsb = 0;
than add another #ifdef.
BTW, which tree are you doing this against? Current -linus uses tsk in
the last line of do_sigbus():
force_sig_fault(SIGBUS, BUS_ADRERR, ..., tsk);
Powered by blists - more mailing lists