| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <20090409232957.b97dd46e.akpm@linux-foundation.org>
Date: Thu, 9 Apr 2009 23:29:57 -0700
From: Andrew Morton <akpm@...ux-foundation.org>
To: liqin.chen@...plusct.com
Cc: linux-arch@...r.kernel.org, linux-kernel@...r.kernel.org,
torvalds@...ux-foundation.org, Arnd Bergmann <arnd@...db.de>,
Sam Ravnborg <sam@...nborg.org>,
Thomas Gleixner <tglx@...utronix.de>,
Kyle McMartin <kyle@...artin.ca>
Subject: Re: [PATCH 4/14] score - New architecure port to SunplusCT S+CORE
On Wed, 8 Apr 2009 15:23:28 +0800 liqin.chen@...plusct.com wrote:
> From: Chen Liqin <liqin.chen@...plusct.com>
>
> asm/kmap_types.h, asm/linkage.h, asm/local.h, asm/mman.h,
> asm/mmu_context.h,
> asm/mmu.h, asm/module.h, asm/msgbuf.h, asm/mutex.h, asm/page.h,
> asm/param.h,
> asm/pci.h, asm/percpu.h, asm/pgalloc.h, asm/pgtable-32.h,
> asm/pgtable-bits.h,
> asm/pgtable.h, asm/poll.h, asm/posix_types.h and asm/processor.h
> for the score architecture.
>
> ...
>
> --- linux-2.6-git.ori/arch/score/include/asm/kmap_types.h 1970-01-01
> 08:00:00.000000000 +0800
> +++ linux-2.6-git.new/arch/score/include/asm/kmap_types.h 2009-04-03
> 17:01:04.000000000 +0800
> @@ -0,0 +1,21 @@
> +#ifndef __SCORE_KMAP_TYPES_H
> +#define __SCORE_KMAP_TYPES_H
> +
> +enum km_type {
> + KM_BOUNCE_READ,
> + KM_SKB_SUNRPC_DATA,
> + KM_SKB_DATA_SOFTIRQ,
> + KM_USER0,
> + KM_USER1,
> + KM_BIO_SRC_IRQ,
> + KM_BIO_DST_IRQ,
> + KM_PTE0,
> + KM_PTE1,
> + KM_IRQ0,
> + KM_IRQ1,
> + KM_SOFTIRQ0,
> + KM_SOFTIRQ1,
> + KM_TYPE_NR
> +};
hm. Our nineteenth copy of kmap_types.h, all of them basically the
same.
That's not a problem for this patchset, but we suck.
>
> ...
>
> +static inline void
> +get_new_mmu_context(struct mm_struct *mm, unsigned long cpu)
> +{
> + unsigned long asid = asid_cache(0);
> +
> + if (!((asid = asid + ASID_INC) & ASID_MASK)) {
ick. Complex statements likethis are to be avoided, please.
unsigned long asid = asid_cache(0) + ASID_INC;
if (!(asid & ASID_MASK)) {
is nicer, no?
> + local_flush_tlb_all(); /* start new asid cycle */
> + if (!asid) /* fix version if needed
> */
> + asid = ASID_FIRST_VERSION;
> + }
> + cpu_context(0, mm) = asid_cache(0) = asid;
This style causes less concern, but some would even prefer
cpu_context(0, mm) = asid;
asid_cache(0) = asid;
C lets you perform lots of weird shortcuts and tricks, but we actively
avoid many of them in the kernel.
> +}
> +
> +/*
> + * Initialize the context related info for a new mm_struct
> + * instance.
> + */
> +static inline int
> +init_new_context(struct task_struct *tsk, struct mm_struct *mm)
> +{
> + cpu_context(0, mm) = 0;
More stylistic whining - this is one of my pet peeves. It's just a
meaningless absurdity to call a function and to then modify its return
value.
So cpu_context() _has_ to be a macro which evaluates to an lvalue.
Which wrecks any concept of encapsulation.
Macros-which-look-like-functions should be interchangesable with
functions, but this one isn't.
I mean, the above statement jusst isn't C. Whereas
set_cpu_context(0, mm0, 0);
_is_ C. See what I mean.
Anwyay. I'm not saying you should go off and change everything. But
it's daft.
> + return 0;
> +}
> +
> +static inline void switch_mm(struct mm_struct *prev, struct mm_struct
> *next,
> + struct task_struct *tsk)
> +{
> + unsigned long flags;
> +
> + local_irq_save(flags);
> + if ((cpu_context(0, next) ^ asid_cache(0)) & ASID_VERSION_MASK)
> + get_new_mmu_context(next, 0);
> + set_PEVN(cpu_context(0, next));
> + TLBMISS_HANDLER_SETUP_PGD(next->pgd);
Some macros are all-lower-case.
Some macros are part-lower-case and part-upper-case.
Some macros are all-upper-case.
Is there any sense behind it all?
> + local_irq_restore(flags);
> +}
> +
> +/*
> + * Destroy context related info for an mm_struct that is about
> + * to be put to rest.
> + */
> +static inline void destroy_context(struct mm_struct *mm)
> +{}
> +
> +#define deactivate_mm(tsk, mm) do {} while (0)
static inline void deactivate_mm(struct task_struct *task, struct mm_struct *mm)
is nicer. It isn't a macro, and it provides typechecking and sometimes
it can fix an unused-variable warning at callsites.
Probably you copied this from somewhere else. Sigh.
> +/*
> + * After we have set current->mm to a new value, this activates
> + * the context for the new mm so we see the new mappings.
> + */
> +static inline void
> +activate_mm(struct mm_struct *prev, struct mm_struct *next)
> +{
> + unsigned long flags;
> +
> + local_irq_save(flags);
> + get_new_mmu_context(next, 0);
> + set_PEVN(cpu_context(0, next));
> + TLBMISS_HANDLER_SETUP_PGD(next->pgd);
> + local_irq_restore(flags);
> +}
> +
> +#endif /* __SCORE_MMU_CONTEXT_H */
> diff -uprN -x linux-2.6-git.ori/Documentation/dontdiff
> linux-2.6-git.ori/arch/score/include/asm/mmu.h
> linux-2.6-git.new/arch/score/include/asm/mmu.h
> --- linux-2.6-git.ori/arch/score/include/asm/mmu.h 1970-01-01
> 08:00:00.000000000 +0800
> +++ linux-2.6-git.new/arch/score/include/asm/mmu.h 2009-04-08
> 10:51:34.000000000 +0800
> @@ -0,0 +1,6 @@
> +#ifndef __SCORE_MMU_H
> +#define __SCORE_MMU_H
> +
> +typedef unsigned long mm_context_t[NR_CPUS];
hm. Is this an SMP-capable architecture?
> +#endif /* __SCORE_MMU_H */
> diff -uprN -x linux-2.6-git.ori/Documentation/dontdiff
>
> ...
>
> +#ifdef CONFIG_CPU_SCORE7
> +#define MODULE_PROC_FAMILY "SCORE7"
> +#else
> +#error MODULE_PROC_FAMILY undefined for your processor configuration
> +#endif
Is the #error necessary? If it triggers, that indicates a bug in
Kconfig, no?
>
> ...
>
> +#define ARCH_PFN_OFFSET PFN_UP(PHYS_OFFSET)
> +
> +#define clear_page(page) memset((void *)(page), 0, PAGE_SIZE)
> +#define copy_page(to, from) memcpy((to), (from), PAGE_SIZE)
> +
> +#define clear_user_page(page, vaddr, pg) clear_page(page)
> +#define copy_user_page(to, from, vaddr, pg) copy_page(to, from)
Well that's nice and simple ;)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists