| 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
| ||
|
Date: Fri, 1 Jul 2022 19:51:58 +0300
From: Mike Rapoport <rppt@...nel.org>
To: "guanghui.fgh" <guanghuifeng@...ux.alibaba.com>
Cc: baolin.wang@...ux.alibaba.com, catalin.marinas@....com,
will@...nel.org, akpm@...ux-foundation.org, david@...hat.com,
jianyong.wu@....com, james.morse@....com, quic_qiancai@...cinc.com,
christophe.leroy@...roup.eu, jonathan@...ek.ca,
mark.rutland@....com, thunder.leizhen@...wei.com,
anshuman.khandual@....com, linux-arm-kernel@...ts.infradead.org,
linux-kernel@...r.kernel.org, geert+renesas@...der.be,
ardb@...nel.org, linux-mm@...ck.org, yaohongbo@...ux.alibaba.com,
alikernel-developer@...ux.alibaba.com
Subject: Re: [PATCH v3] arm64: mm: fix linear mapping mem access performance
degradation
On Fri, Jul 01, 2022 at 12:36:00PM +0800, guanghui.fgh wrote:
> Thanks.
>
> 在 2022/6/30 21:46, Mike Rapoport 写道:
> > Hi,
> >
> > On Thu, Jun 30, 2022 at 06:50:22PM +0800, Guanghui Feng wrote:
> > > The arm64 can build 2M/1G block/sectiion mapping. When using DMA/DMA32 zone
> > > (enable crashkernel, disable rodata full, disable kfence), the mem_map will
> > > use non block/section mapping(for crashkernel requires to shrink the region
> > > in page granularity). But it will degrade performance when doing larging
> > > continuous mem access in kernel(memcpy/memmove, etc).
> > >
> > > There are many changes and discussions:
> > > commit 031495635b46 ("arm64: Do not defer reserve_crashkernel() for
> > > platforms with no DMA memory zones")
> > > commit 0a30c53573b0 ("arm64: mm: Move reserve_crashkernel() into
> > > mem_init()")
> > > commit 2687275a5843 ("arm64: Force NO_BLOCK_MAPPINGS if crashkernel
> > > reservation is required")
> > >
> > > This patch changes mem_map to use block/section mapping with crashkernel.
> > > Firstly, do block/section mapping(normally 2M or 1G) for all avail mem at
> > > mem_map, reserve crashkernel memory. And then walking pagetable to split
> > > block/section mapping to non block/section mapping(normally 4K) [[[only]]]
> > > for crashkernel mem. So the linear mem mapping use block/section mapping
> > > as more as possible. We will reduce the cpu dTLB miss conspicuously, and
> > > accelerate mem access about 10-20% performance improvement.
> >
> > ...
> > > Signed-off-by: Guanghui Feng <guanghuifeng@...ux.alibaba.com>
> > > ---
> > > arch/arm64/include/asm/mmu.h | 1 +
> > > arch/arm64/mm/init.c | 8 +-
> > > arch/arm64/mm/mmu.c | 231 ++++++++++++++++++++++++++++++-------------
> > > 3 files changed, 168 insertions(+), 72 deletions(-)
> >
> > ...
> >
> > > diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
> > > index 626ec32..4b779cf 100644
> > > --- a/arch/arm64/mm/mmu.c
> > > +++ b/arch/arm64/mm/mmu.c
> > > @@ -42,6 +42,7 @@
> > > #define NO_BLOCK_MAPPINGS BIT(0)
> > > #define NO_CONT_MAPPINGS BIT(1)
> > > #define NO_EXEC_MAPPINGS BIT(2) /* assumes FEAT_HPDS is not used */
> > > +#define NO_SEC_REMAPPINGS BIT(3) /* rebuild with non block/sec mapping*/
> > > u64 idmap_t0sz = TCR_T0SZ(VA_BITS_MIN);
> > > u64 idmap_ptrs_per_pgd = PTRS_PER_PGD;
> > > @@ -156,11 +157,12 @@ static bool pgattr_change_is_safe(u64 old, u64 new)
> > > }
> > > static void init_pte(pmd_t *pmdp, unsigned long addr, unsigned long end,
> > > - phys_addr_t phys, pgprot_t prot)
> > > + phys_addr_t phys, pgprot_t prot, int flags)
> > > {
> > > pte_t *ptep;
> > > - ptep = pte_set_fixmap_offset(pmdp, addr);
> > > + ptep = (flags & NO_SEC_REMAPPINGS) ? pte_offset_kernel(pmdp, addr) :
> > > + pte_set_fixmap_offset(pmdp, addr);
> > > do {
> > > pte_t old_pte = READ_ONCE(*ptep);
> > > @@ -176,7 +178,8 @@ static void init_pte(pmd_t *pmdp, unsigned long addr, unsigned long end,
> > > phys += PAGE_SIZE;
> > > } while (ptep++, addr += PAGE_SIZE, addr != end);
> > > - pte_clear_fixmap();
> > > + if (!(flags & NO_SEC_REMAPPINGS))
> > > + pte_clear_fixmap();
> > > }
> > > static void alloc_init_cont_pte(pmd_t *pmdp, unsigned long addr,
> > > @@ -208,16 +211,59 @@ static void alloc_init_cont_pte(pmd_t *pmdp, unsigned long addr,
> > > next = pte_cont_addr_end(addr, end);
> > > /* use a contiguous mapping if the range is suitably aligned */
> > > - if ((((addr | next | phys) & ~CONT_PTE_MASK) == 0) &&
> > > + if (!(flags & NO_SEC_REMAPPINGS) &&
> > > + (((addr | next | phys) & ~CONT_PTE_MASK) == 0) &&
> > > (flags & NO_CONT_MAPPINGS) == 0)
> > > __prot = __pgprot(pgprot_val(prot) | PTE_CONT);
> > > - init_pte(pmdp, addr, next, phys, __prot);
> > > + init_pte(pmdp, addr, next, phys, __prot, flags);
> > > phys += next - addr;
> > > } while (addr = next, addr != end);
> > > }
> > > +static void init_pmd_remap(pud_t *pudp, unsigned long addr, unsigned long end,
> > > + phys_addr_t phys, pgprot_t prot,
> > > + phys_addr_t (*pgtable_alloc)(int), int flags)
> > > +{
> > > + unsigned long next;
> > > + pmd_t *pmdp;
> > > + phys_addr_t map_offset;
> > > + pmdval_t pmdval;
> > > +
> > > + pmdp = pmd_offset(pudp, addr);
> > > + do {
> > > + next = pmd_addr_end(addr, end);
> > > +
> > > + if (!pmd_none(*pmdp) && pmd_sect(*pmdp)) {
> > > + phys_addr_t pte_phys = pgtable_alloc(PAGE_SHIFT);
> > > + pmd_clear(pmdp);
> > > + pmdval = PMD_TYPE_TABLE | PMD_TABLE_UXN;
> > > + if (flags & NO_EXEC_MAPPINGS)
> > > + pmdval |= PMD_TABLE_PXN;
> > > + __pmd_populate(pmdp, pte_phys, pmdval);
> > > + flush_tlb_kernel_range(addr, addr + PAGE_SIZE);
> > > +
> > > + map_offset = addr - (addr & PMD_MASK);
> > > + if (map_offset)
> > > + alloc_init_cont_pte(pmdp, addr & PMD_MASK, addr,
> > > + phys - map_offset, prot,
> > > + pgtable_alloc,
> > > + flags & (~NO_SEC_REMAPPINGS));
> > > +
> > > + if (next < (addr & PMD_MASK) + PMD_SIZE)
> > > + alloc_init_cont_pte(pmdp, next,
> > > + (addr & PUD_MASK) + PUD_SIZE,
> > > + next - addr + phys,
> > > + prot, pgtable_alloc,
> > > + flags & (~NO_SEC_REMAPPINGS));
> > > + }
> > > + alloc_init_cont_pte(pmdp, addr, next, phys, prot,
> > > + pgtable_alloc, flags);
> > > + phys += next - addr;
> > > + } while (pmdp++, addr = next, addr != end);
> > > +}
> >
> > There is still to much duplicated code here and in init_pud_remap().
> >
> > Did you consider something like this:
> >
> > void __init map_crashkernel(void)
> > {
> > int flags = NO_EXEC_MAPPINGS | NO_BLOCK_MAPPINGS | NO_CONT_MAPPINGS;
> > u64 size;
> >
> > /*
> > * check if crash kernel supported, reserved etc
> > */
> >
> >
> > size = crashk_res.end + 1 - crashk_res.start;
> >
> > __remove_pgd_mapping(swapper_pg_dir, __phys_to_virt(start), size);
> > __create_pgd_mapping(swapper_pg_dir, crashk_res.start,
> > __phys_to_virt(crashk_res.start), size,
> > PAGE_KERNEL, early_pgtable_alloc, flags);
> > }
> >
> I'm trying do this.
> But I think it's the Inverse Process of mem mapping and also generates
> duplicated code(Boundary judgment, pagetable modify).
>
> When removing the pgd mapping, it may split pud/pmd section which also needs
> [[[rebuild and clear]]] some pagetable.
Well, __remove_pgd_mapping() is probably an overkill, but
unmap_hotplug_pmd_range() and unmap_hotplug_pud_range() should do, depending
on the size of the crash kernel.
--
Sincerely yours,
Mike.
Powered by blists - more mailing lists