[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20171130182956.uwc63g6i6yobxcqy@lakrids.cambridge.arm.com>
Date: Thu, 30 Nov 2017 18:29:56 +0000
From: Mark Rutland <mark.rutland@....com>
To: Will Deacon <will.deacon@....com>
Cc: linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
catalin.marinas@....com, ard.biesheuvel@...aro.org,
sboyd@...eaurora.org, dave.hansen@...ux.intel.com,
keescook@...omium.org, msalter@...hat.com, labbott@...hat.com,
tglx@...utronix.de
Subject: Re: [PATCH v2 11/18] arm64: mm: Map entry trampoline into trampoline
and kernel page tables
Hi Will,
On Thu, Nov 30, 2017 at 04:39:39PM +0000, Will Deacon wrote:
> diff --git a/arch/arm64/include/asm/fixmap.h b/arch/arm64/include/asm/fixmap.h
> index 4052ec39e8db..8119b49be98d 100644
> --- a/arch/arm64/include/asm/fixmap.h
> +++ b/arch/arm64/include/asm/fixmap.h
> @@ -58,6 +58,10 @@ enum fixed_addresses {
> FIX_APEI_GHES_NMI,
> #endif /* CONFIG_ACPI_APEI_GHES */
>
> +#ifdef CONFIG_UNMAP_KERNEL_AT_EL0
> + FIX_ENTRY_TRAMP_TEXT,
> +#define TRAMP_VALIAS (__fix_to_virt(FIX_ENTRY_TRAMP_TEXT))
> +#endif /* CONFIG_UNMAP_KERNEL_AT_EL0 */
> __end_of_permanent_fixed_addresses,
Defining TRAMP_VALIAS here is a little surprising, especially given we
reuse the name in asm-offsets:
> + DEFINE(TRAMP_VALIAS, TRAMP_VALIAS);
Can't we have asm-offsets do:
DEFINE(TRAMP_VALIAS, __fix_to_virt(FIX_ENTRY_TRAMP_TEXT));
... and rely on the asm-offsets TRAMP_VALIAS definition everywhere?
[...]
> +#ifdef CONFIG_UNMAP_KERNEL_AT_EL0
> +static int __init map_entry_trampoline(void)
> +{
> + extern char __entry_tramp_text_start[];
> +
> + pgprot_t prot = rodata_enabled ? PAGE_KERNEL_ROX : PAGE_KERNEL_EXEC;
> + phys_addr_t pa_start = __pa_symbol(__entry_tramp_text_start);
> +
> + /* The trampoline is always mapped and can therefore be global */
> + pgprot_val(prot) &= ~PTE_NG;
> +
> + /* Map only the text into the trampoline page table */
> + memset((char *)tramp_pg_dir, 0, PGD_SIZE);
The (char *) cast can go; memset() takes a void pointer and we don't do
similar casts for other memset instances.
> + __create_pgd_mapping(tramp_pg_dir, pa_start, TRAMP_VALIAS, PAGE_SIZE,
> + prot, pgd_pgtable_alloc, 0);
> +
> + /* ...as well as the kernel page table */
This might be clearer as:
/* map the text in the kernel page table, too */
Otherwise, this looks good to me.
Thanks,
Mark.
Powered by blists - more mailing lists