Skip to contentSkip to conversation

FAILED: patch "[PATCH] MIPS: VDSO: Match data page cache colouring when D$ aliases" failed to apply to 4.4-stable tree

PATCH [email protected] 2 participants 4 messages Sep 16, 2018 – Sep 24, 2018

Subsystems (2)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 4.4-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

Quoted 1 line

From 0f02cfbc3d9e413d450d8d0fd660077c23f67eff Mon Sep 17 00:00:00 2001

From: Paul Burton <[email protected]>
Date: Thu, 30 Aug 2018 11:01:21 -0700
Subject: [PATCH] MIPS: VDSO: Match data page cache colouring when D$ aliases

When a system suffers from dcache aliasing a user program may observe stale VDSO data from an aliased cache line. Notably this can break the expectation that clock_gettime(CLOCK_MONOTONIC, ...) is, as its name suggests, monotonic.

In order to ensure that users observe updates to the VDSO data page as intended, align the user mappings of the VDSO data page such that their cache colouring matches that of the virtual address range which the kernel will use to update the data page - typically its unmapped address within kseg0.

This ensures that we don't introduce aliasing cache lines for the VDSO data page, and therefore that userland will observe updates without requiring cache invalidation.

Fixes ebb5e78cc634 ("MIPS: Initial implementation of a VDSO")

Patchwork: https://patchwork.linux-mips.org/patch/20344/

arch/mips/kernel/vdso.c +20 -0
index 019035d7225c..8f845f6e5f42 100644
@@ -13,6 +13,7 @@
 #include <linux/err.h>
 #include <linux/init.h>
 #include <linux/ioport.h>
+#include <linux/kernel.h>
 #include <linux/mm.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
@@ -20,6 +21,7 @@
 
 #include <asm/abi.h>
 #include <asm/mips-cps.h>
+#include <asm/page.h>
 #include <asm/vdso.h>
 
 /* Kernel-provided data used by the VDSO. */
@@ -128,12 +130,30 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 	vvar_size = gic_size + PAGE_SIZE;
 	size = vvar_size + image->size;
 
+	/*
+	 * Find a region that's large enough for us to perform the
+	 * colour-matching alignment below.
+	 */
+	if (cpu_has_dc_aliases)
+		size += shm_align_mask + 1;
+
 	base = get_unmapped_area(NULL, 0, size, 0, 0);
 	if (IS_ERR_VALUE(base)) {
 		ret = base;
 		goto out;
 	}
 
+	/*
+	 * If we suffer from dcache aliasing, ensure that the VDSO data page
+	 * mapping is coloured the same as the kernel's mapping of that memory.
+	 * This ensures that when the kernel updates the VDSO data userland
+	 * will observe it without requiring cache invalidations.
+	 */
+	if (cpu_has_dc_aliases) {
+		base = __ALIGN_MASK(base, shm_align_mask);
+		base += ((unsigned long)&vdso_data - gic_size) & shm_align_mask;
+	}
+
 	data_addr = base + gic_size;
 	vdso_addr = data_addr + PAGE_SIZE;
 
#1 lore raw

Hi,

Quoted 5 lines · On Sun, Sep 16, 2018 at 02:11:33PM +0200, [email protected] wrote:

The patch below does not apply to the 4.4-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

Please find attached patch 00578cd864d45 ("MIPS: VDSO: Drop gic_get_usm_range() usage")
and apply to {4.4,4.9}-stable tree for solving this conflict.

Thanks!

SZ Lin (林上智)

Quoted 91 lines

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 0f02cfbc3d9e413d450d8d0fd660077c23f67eff Mon Sep 17 00:00:00 2001
From: Paul Burton <[email protected]>
Date: Thu, 30 Aug 2018 11:01:21 -0700
Subject: [PATCH] MIPS: VDSO: Match data page cache colouring when D$ aliases

When a system suffers from dcache aliasing a user program may observe stale VDSO data from an aliased cache line. Notably this can break the expectation that clock_gettime(CLOCK_MONOTONIC, ...) is, as its name suggests, monotonic.

In order to ensure that users observe updates to the VDSO data page as intended, align the user mappings of the VDSO data page such that their cache colouring matches that of the virtual address range which the kernel will use to update the data page - typically its unmapped address within kseg0.

This ensures that we don't introduce aliasing cache lines for the VDSO data page, and therefore that userland will observe updates without requiring cache invalidation.

Fixes ebb5e78cc634 ("MIPS: Initial implementation of a VDSO")

Patchwork: https://patchwork.linux-mips.org/patch/20344/

arch/mips/kernel/vdso.c +20 -0
index 019035d7225c..8f845f6e5f42 100644
@@ -13,6 +13,7 @@
 #include <linux/err.h>
 #include <linux/init.h>
 #include <linux/ioport.h>
+#include <linux/kernel.h>
 #include <linux/mm.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
@@ -20,6 +21,7 @@
 
 #include <asm/abi.h>
 #include <asm/mips-cps.h>
+#include <asm/page.h>
 #include <asm/vdso.h>
 
 /* Kernel-provided data used by the VDSO. */
@@ -128,12 +130,30 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 	vvar_size = gic_size + PAGE_SIZE;
 	size = vvar_size + image->size;
 
+	/*
+	 * Find a region that's large enough for us to perform the
+	 * colour-matching alignment below.
+	 */
+	if (cpu_has_dc_aliases)
+		size += shm_align_mask + 1;
+
 	base = get_unmapped_area(NULL, 0, size, 0, 0);
 	if (IS_ERR_VALUE(base)) {
 		ret = base;
 		goto out;
 	}
 
+	/*
+	 * If we suffer from dcache aliasing, ensure that the VDSO data page
+	 * mapping is coloured the same as the kernel's mapping of that memory.
+	 * This ensures that when the kernel updates the VDSO data userland
+	 * will observe it without requiring cache invalidations.
+	 */
+	if (cpu_has_dc_aliases) {
+		base = __ALIGN_MASK(base, shm_align_mask);
+		base += ((unsigned long)&vdso_data - gic_size) & shm_align_mask;
+	}
+
 	data_addr = base + gic_size;
 	vdso_addr = data_addr + PAGE_SIZE;
 
Signature · 5 lines
SZ Lin (林上智) <[email protected]>, http://people.debian.org/~szlin

Debian Developer, debian.org.tw Administrator

4096R/ 178F 8338 B314 01E3 04FC 44BA A959 B38A 9561 F3F9
Quoted 1 line

From 890151c23e16a48ad0ba05ff60633234c1934c8b Mon Sep 17 00:00:00 2001

From: Paul Burton <[email protected]>
Date: Sat, 12 Aug 2017 21:36:30 -0700
Subject: [PATCH] MIPS: VDSO: Drop gic_get_usm_range() usage

commit 00578cd864d45ae4b8fa3f684f8d6f783dd8d15d upstream

We don't really need gic_get_usm_range() to abstract discovery of the address of the GIC user-visible section now that we have access to its base address globally.

Switch to calculating it ourselves, which will allow us to stop requiring the irqchip driver to care about a counter exposed to userland for use via the VDSO.

Cc Jason Cooper <[email protected]>
Cc Marc Zyngier <[email protected]>
Cc Thomas Gleixner <[email protected]>

Patchwork: https://patchwork.linux-mips.org/patch/17040/

---
 arch/mips/kernel/vdso.c | 15 +++++----------
 1 file changed, 5 insertions(+), 10 deletions(-)
arch/mips/kernel/vdso.c +5 -10
index 5649a9e429e0..1266fd56c919 100644
@@ -13,13 +13,13 @@
 #include <linux/err.h>
 #include <linux/init.h>
 #include <linux/ioport.h>
-#include <linux/irqchip/mips-gic.h>
 #include <linux/mm.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
 #include <linux/timekeeper_internal.h>
 
 #include <asm/abi.h>
+#include <asm/mips-cps.h>
 #include <asm/vdso.h>
 
 /* Kernel-provided data used by the VDSO. */
@@ -99,9 +99,8 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 {
 	struct mips_vdso_image *image = current->thread.abi->vdso;
 	struct mm_struct *mm = current->mm;
-	unsigned long gic_size, vvar_size, size, base, data_addr, vdso_addr;
+	unsigned long gic_size, vvar_size, size, base, data_addr, vdso_addr, gic_pfn;
 	struct vm_area_struct *vma;
-	struct resource gic_res;
 	int ret;
 
 	down_write(&mm->mmap_sem);
@@ -114,7 +113,7 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 	 * only map a page even though the total area is 64K, as we only need
 	 * the counter registers at the start.
 	 */
-	gic_size = gic_present ? PAGE_SIZE : 0;
+	gic_size = mips_gic_present() ? PAGE_SIZE : 0;
 	vvar_size = gic_size + PAGE_SIZE;
 	size = vvar_size + image->size;
 
@@ -137,13 +136,9 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 
 	/* Map GIC user page. */
 	if (gic_size) {
-		ret = gic_get_usm_range(&gic_res);
-		if (ret)
-			goto out;
+		gic_pfn = virt_to_phys(mips_gic_base + MIPS_GIC_USER_OFS) >> PAGE_SHIFT;
 
-		ret = io_remap_pfn_range(vma, base,
-					 gic_res.start >> PAGE_SHIFT,
-					 gic_size,
+		ret = io_remap_pfn_range(vma, base, gic_pfn, gic_size,
 					 pgprot_noncached(PAGE_READONLY));
 		if (ret)
 			goto out;
Signature · 1 line
2.19.0
author maintainer
#2 lore raw
Quoted 92 lines · SZ Lin (林上智)

Hi,

Quoted 5 lines · On Sun, Sep 16, 2018 at 02:11:33PM +0200, [email protected] wrote:

The patch below does not apply to the 4.4-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

Please find attached patch 00578cd864d45 ("MIPS: VDSO: Drop gic_get_usm_range() usage")
and apply to {4.4,4.9}-stable tree for solving this conflict.

Thanks!

SZ Lin (林上智)

Quoted 76 lines

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 0f02cfbc3d9e413d450d8d0fd660077c23f67eff Mon Sep 17 00:00:00 2001
From: Paul Burton <[email protected]>
Date: Thu, 30 Aug 2018 11:01:21 -0700
Subject: [PATCH] MIPS: VDSO: Match data page cache colouring when D$ aliases

When a system suffers from dcache aliasing a user program may observe stale VDSO data from an aliased cache line. Notably this can break the expectation that clock_gettime(CLOCK_MONOTONIC, ...) is, as its name suggests, monotonic.

In order to ensure that users observe updates to the VDSO data page as intended, align the user mappings of the VDSO data page such that their cache colouring matches that of the virtual address range which the kernel will use to update the data page - typically its unmapped address within kseg0.

This ensures that we don't introduce aliasing cache lines for the VDSO data page, and therefore that userland will observe updates without requiring cache invalidation.

Fixes ebb5e78cc634 ("MIPS: Initial implementation of a VDSO")

Patchwork: https://patchwork.linux-mips.org/patch/20344/

arch/mips/kernel/vdso.c +9 -0
index 019035d7225c..8f845f6e5f42 100644
@@ -13,6 +13,7 @@
 #include <linux/err.h>
 #include <linux/init.h>
 #include <linux/ioport.h>
+#include <linux/kernel.h>
 #include <linux/mm.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
@@ -20,6 +21,7 @@
 
 #include <asm/abi.h>
 #include <asm/mips-cps.h>
+#include <asm/page.h>
 #include <asm/vdso.h>
 
 /* Kernel-provided data used by the VDSO. */
@@ -128,12 +130,30 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 	vvar_size = gic_size + PAGE_SIZE;
 	size = vvar_size + image->size;
 
+	/*
+	 * Find a region that's large enough for us to perform the
+	 * colour-matching alignment below.
+	 */
+	if (cpu_has_dc_aliases)
+		size += shm_align_mask + 1;
+
 	base = get_unmapped_area(NULL, 0, size, 0, 0);
 	if (IS_ERR_VALUE(base)) {
 		ret = base;
 		goto out;
 	}
 
Quoted 46 lines
Quoted 15 lines
arch/mips/kernel/vdso.c +11 -0
+	/*
+	 * If we suffer from dcache aliasing, ensure that the VDSO data page
+	 * mapping is coloured the same as the kernel's mapping of that memory.
+	 * This ensures that when the kernel updates the VDSO data userland
+	 * will observe it without requiring cache invalidations.
+	 */
+	if (cpu_has_dc_aliases) {
+		base = __ALIGN_MASK(base, shm_align_mask);
+		base += ((unsigned long)&vdso_data - gic_size) & shm_align_mask;
+	}
+
 	data_addr = base + gic_size;
 	vdso_addr = data_addr + PAGE_SIZE;
 
Signature · 5 lines
SZ Lin (林上智) <[email protected]>, http://people.debian.org/~szlin

Debian Developer, debian.org.tw Administrator

4096R/ 178F 8338 B314 01E3 04FC 44BA A959 B38A 9561 F3F9
Quoted 1 line

From 890151c23e16a48ad0ba05ff60633234c1934c8b Mon Sep 17 00:00:00 2001

From: Paul Burton <[email protected]>
Date: Sat, 12 Aug 2017 21:36:30 -0700
Subject: [PATCH] MIPS: VDSO: Drop gic_get_usm_range() usage

commit 00578cd864d45ae4b8fa3f684f8d6f783dd8d15d upstream

We don't really need gic_get_usm_range() to abstract discovery of the address of the GIC user-visible section now that we have access to its base address globally.

Switch to calculating it ourselves, which will allow us to stop requiring the irqchip driver to care about a counter exposed to userland for use via the VDSO.

Cc Jason Cooper <[email protected]>
Cc Marc Zyngier <[email protected]>
Cc Thomas Gleixner <[email protected]>

Patchwork: https://patchwork.linux-mips.org/patch/17040/

I do not understand. What "bug" is this fixing that this needs to be backported for?

thanks,

greg k-h

author maintainer
#3 lore raw
In reply to SZ Lin (林上智)
Quoted 11 lines · SZ Lin (林上智)

Hi,

Quoted 5 lines · On Sun, Sep 16, 2018 at 02:11:33PM +0200, [email protected] wrote:

The patch below does not apply to the 4.4-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

Please find attached patch 00578cd864d45 ("MIPS: VDSO: Drop gic_get_usm_range() usage")
and apply to {4.4,4.9}-stable tree for solving this conflict.

Now applied, thanks.

greg k-h