Linux-parisc archive on lore.kernel.org
 help / color / Atom feed
From: Mike Rapoport <rppt@kernel.org>
To: Helge Deller <deller@gmx.de>
Cc: "James E.J. Bottomley" <James.Bottomley@hansenpartnership.com>,
	Meelis Roos <mroos@linux.ee>, Jeroen Roovers <jer@gentoo.org>,
	Mikulas Patocka <mpatocka@redhat.com>,
	linux-parisc@vger.kernel.org, linux-kernel@vger.kernel.org,
	Mike Rapoport <rppt@linux.ibm.com>
Subject: [PATCH] parisc: map_pages(): cleanup page table initialization
Date: Sun, 12 Jan 2020 13:36:15 +0200
Message-ID: <20200112113615.24198-1-rppt@kernel.org> (raw)

From: Mike Rapoport <rppt@linux.ibm.com>

The current code uses '#if PTRS_PER_PMD == 1' to distinguish 2 vs 3 levels,
setup, it casts pgd to pgd to cope with page table folding and converts
addresses of page table entries from physical to virtual and back for no
good reason.

Simplify the accesses to the page table entries using proper unfolding of
the upper layers and replacing '#if PTRS_PER_PMD' with explicit
'#if CONFIG_PGTABLE_LEVELS == 3'

Signed-off-by: Mike Rapoport <rppt@linux.ibm.com>
---
Hi,

This patch is on top of the current parisc/for-next.
I've build-tested for generic-{32,64}_defconfig and was able to boot
qemu-system-parisc up to rootfs mount.

 arch/parisc/mm/init.c | 50 +++++++++++--------------------------------
 1 file changed, 12 insertions(+), 38 deletions(-)

diff --git a/arch/parisc/mm/init.c b/arch/parisc/mm/init.c
index 354cf060b67f..5224fb38d766 100644
--- a/arch/parisc/mm/init.c
+++ b/arch/parisc/mm/init.c
@@ -351,7 +351,6 @@ static void __init map_pages(unsigned long start_vaddr,
 			     unsigned long start_paddr, unsigned long size,
 			     pgprot_t pgprot, int force)
 {
-	pgd_t *pg_dir;
 	pmd_t *pmd;
 	pte_t *pg_table;
 	unsigned long end_paddr;
@@ -372,62 +371,37 @@ static void __init map_pages(unsigned long start_vaddr,
 
 	end_paddr = start_paddr + size;
 
-	pg_dir = pgd_offset_k(start_vaddr);
-
-#if PTRS_PER_PMD == 1
-	start_pmd = 0;
-#else
+	/* for 2-level configuration PTRS_PER_PMD is 0 so start_pmd will be 0 */
 	start_pmd = ((start_vaddr >> PMD_SHIFT) & (PTRS_PER_PMD - 1));
-#endif
 	start_pte = ((start_vaddr >> PAGE_SHIFT) & (PTRS_PER_PTE - 1));
 
 	address = start_paddr;
 	vaddr = start_vaddr;
 	while (address < end_paddr) {
-#if PTRS_PER_PMD == 1
-		pmd = (pmd_t *)__pa(pg_dir);
-#else
-		pmd = (pmd_t *)pgd_address(*pg_dir);
-
-		/*
-		 * pmd is physical at this point
-		 */
+		pgd_t *pgd = pgd_offset_k(vaddr);
+		p4d_t *p4d = p4d_offset(pgd, vaddr);
+		pud_t *pud = pud_offset(p4d, vaddr);
 
-		if (!pmd) {
+#if CONFIG_PGTABLE_LEVELS == 3
+		if (pud_none(*pud)) {
 			pmd = memblock_alloc(PAGE_SIZE << PMD_ORDER,
 					     PAGE_SIZE << PMD_ORDER);
 			if (!pmd)
 				panic("pmd allocation failed.\n");
-			pmd = (pmd_t *) __pa(pmd);
+			pud_populate(NULL, pud, pmd);
 		}
-
-		pud_populate(NULL, (pud_t *)pg_dir, __va(pmd));
 #endif
-		pg_dir++;
-
-		/* now change pmd to kernel virtual addresses */
 
-		pmd = (pmd_t *)__va(pmd) + start_pmd;
+		pmd = pmd_offset(pud, vaddr);
 		for (tmp1 = start_pmd; tmp1 < PTRS_PER_PMD; tmp1++, pmd++) {
-
-			/*
-			 * pg_table is physical at this point
-			 */
-
-			pg_table = (pte_t *)pmd_address(*pmd);
-			if (!pg_table) {
-				pg_table = memblock_alloc(PAGE_SIZE,
-							  PAGE_SIZE);
+			if (pmd_none(*pmd)) {
+				pg_table = memblock_alloc(PAGE_SIZE, PAGE_SIZE);
 				if (!pg_table)
 					panic("page table allocation failed\n");
-				pg_table = (pte_t *) __pa(pg_table);
+				pmd_populate_kernel(NULL, pmd, pg_table);
 			}
 
-			pmd_populate_kernel(NULL, pmd, __va(pg_table));
-
-			/* now change pg_table to kernel virtual addresses */
-
-			pg_table = (pte_t *) __va(pg_table) + start_pte;
+			pg_table = pte_offset_kernel(pmd, vaddr);
 			for (tmp2 = start_pte; tmp2 < PTRS_PER_PTE; tmp2++, pg_table++) {
 				pte_t pte;
 				pgprot_t prot;

base-commit: 9cf86a639ffdd9c38d510d35efcc15ed6dfb2efa
-- 
2.24.0


                 reply index

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200112113615.24198-1-rppt@kernel.org \
    --to=rppt@kernel.org \
    --cc=James.Bottomley@hansenpartnership.com \
    --cc=deller@gmx.de \
    --cc=jer@gentoo.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-parisc@vger.kernel.org \
    --cc=mpatocka@redhat.com \
    --cc=mroos@linux.ee \
    --cc=rppt@linux.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-parisc archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-parisc/0 linux-parisc/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-parisc linux-parisc/ https://lore.kernel.org/linux-parisc \
		linux-parisc@vger.kernel.org
	public-inbox-index linux-parisc

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-parisc


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git