linux-kernel - [PATCH v2] x86/head64: micro optimization: use memset instead of for loop

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue,  9 Feb 2016 19:11:57 +0600
From:	Alexander Kuleshov <kuleshovmail@...il.com>
To:	Ingo Molnar <mingo@...hat.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	"H . Peter Anvin" <hpa@...or.com>
Cc:	Andrey Ryabinin <ryabinin.a.a@...il.com>,
	Andy Lutomirski <luto@...nel.org>,
	Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
	Alexander Popov <alpopov@...ecurity.com>,
	linux-kernel@...r.kernel.org,
	Alexander Kuleshov <kuleshovmail@...il.com>
Subject: [PATCH v2] x86/head64: micro optimization: use memset instead of for loop

This patch simplifies clearing of page table directories, by
the using of the fast memset() from the arch/x86/lib/memset_64.S.
Besides this, we 25 bytes is 25 bytes less than original.

text       data     bss              dec            hex filename

9735144 4970776 15474688        30180608        1cc8500 vmlinux.old
9735119 4970776 15474688        30180583        1cc84e7 vmlinux

Signed-off-by: Alexander Kuleshov <kuleshovmail@...il.com>
---
Changelog:

v2: We got sizeof pud_p/pmd_p in the v1, but actually we need
to take sizeof of pud_p/pmd_p pointer, although their sizes are
the same.

 arch/x86/kernel/head64.c | 14 +++-----------
 1 file changed, 3 insertions(+), 11 deletions(-)

diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
index 2c0f340..1f4422d 100644
--- a/arch/x86/kernel/head64.c
+++ b/arch/x86/kernel/head64.c
@@ -40,13 +40,8 @@ pmdval_t early_pmd_flags = __PAGE_KERNEL_LARGE & ~(_PAGE_GLOBAL | _PAGE_NX);
 /* Wipe all early page tables except for the kernel symbol map */
 static void __init reset_early_page_tables(void)
 {
-	unsigned long i;
-
-	for (i = 0; i < PTRS_PER_PGD-1; i++)
-		early_level4_pgt[i].pgd = 0;
-
+	memset(early_level4_pgt, 0, sizeof(pgd_t)*(PTRS_PER_PGD-1));
 	next_early_pgt = 0;
-
 	write_cr3(__pa_nodebug(early_level4_pgt));
 }
 
@@ -54,7 +49,6 @@ static void __init reset_early_page_tables(void)
 int __init early_make_pgtable(unsigned long address)
 {
 	unsigned long physaddr = address - __PAGE_OFFSET;
-	unsigned long i;
 	pgdval_t pgd, *pgd_p;
 	pudval_t pud, *pud_p;
 	pmdval_t pmd, *pmd_p;
@@ -81,8 +75,7 @@ again:
 		}
 
 		pud_p = (pudval_t *)early_dynamic_pgts[next_early_pgt++];
-		for (i = 0; i < PTRS_PER_PUD; i++)
-			pud_p[i] = 0;
+		memset(pud_p, 0, sizeof(*pud_p) * PTRS_PER_PUD);
 		*pgd_p = (pgdval_t)pud_p - __START_KERNEL_map + phys_base + _KERNPG_TABLE;
 	}
 	pud_p += pud_index(address);
@@ -97,8 +90,7 @@ again:
 		}
 
 		pmd_p = (pmdval_t *)early_dynamic_pgts[next_early_pgt++];
-		for (i = 0; i < PTRS_PER_PMD; i++)
-			pmd_p[i] = 0;
+		memset(pmd_p, 0, sizeof(*pmd_p) * PTRS_PER_PMD);
 		*pud_p = (pudval_t)pmd_p - __START_KERNEL_map + phys_base + _KERNPG_TABLE;
 	}
 	pmd = (physaddr & PMD_MASK) + early_pmd_flags;
-- 
2.7.0.25.gfc10eb5