Patch "hns: avoid stack overflow with CONFIG_KASAN" has been added to the 4.4-stable tree

This is a note to let you know that I've just added the patch titled

hns: avoid stack overflow with CONFIG_KASAN

to the 4.4-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
hns-avoid-stack-overflow-with-config_kasan.patch
and it can be found in the queue-4.4 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.

From: Arnd Bergmann <[email protected]>
Date: Fri, 3 Feb 2017 17:35:46 +0100
Subject: hns: avoid stack overflow with CONFIG_KASAN

From: Arnd Bergmann <[email protected]>

commit b3f2d07f4649adcf6905953a10d217b5683e4077 upstream.

The use of ACCESS_ONCE() looks like a micro-optimization to force gcc to use an indexed load for the register address, but it has an absolutely detrimental effect on builds with gcc-5 and CONFIG_KASAN=y, leading to a very likely kernel stack overflow aside from very complex object code:

hisilicon/hns/hns_dsaf_gmac.c: In function 'hns_gmac_update_stats':
hisilicon/hns/hns_dsaf_gmac.c:419:1: error: the frame size of 2912 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] hisilicon/hns/hns_dsaf_ppe.c: In function 'hns_ppe_reset_common':
hisilicon/hns/hns_dsaf_ppe.c:390:1: error: the frame size of 1184 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] hisilicon/hns/hns_dsaf_ppe.c: In function 'hns_ppe_get_regs':
hisilicon/hns/hns_dsaf_ppe.c:621:1: error: the frame size of 3632 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] hisilicon/hns/hns_dsaf_rcb.c: In function 'hns_rcb_get_common_regs':
hisilicon/hns/hns_dsaf_rcb.c:970:1: error: the frame size of 2784 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] hisilicon/hns/hns_dsaf_gmac.c: In function 'hns_gmac_get_regs':
hisilicon/hns/hns_dsaf_gmac.c:641:1: error: the frame size of 5728 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] hisilicon/hns/hns_dsaf_rcb.c: In function 'hns_rcb_get_ring_regs':
hisilicon/hns/hns_dsaf_rcb.c:1021:1: error: the frame size of 2208 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] hisilicon/hns/hns_dsaf_main.c: In function 'hns_dsaf_comm_init':
hisilicon/hns/hns_dsaf_main.c:1209:1: error: the frame size of 1904 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] hisilicon/hns/hns_dsaf_xgmac.c: In function 'hns_xgmac_get_regs':
hisilicon/hns/hns_dsaf_xgmac.c:748:1: error: the frame size of 4704 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] hisilicon/hns/hns_dsaf_main.c: In function 'hns_dsaf_update_stats':
hisilicon/hns/hns_dsaf_main.c:2420:1: error: the frame size of 1088 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] hisilicon/hns/hns_dsaf_main.c: In function 'hns_dsaf_get_regs':
hisilicon/hns/hns_dsaf_main.c:2753:1: error: the frame size of 10768 bytes is larger than 1024 bytes [-Werror=frame-larger-than=]

This does not seem to happen any more with gcc-7, but removing the ACCESS_ONCE seems safe anyway and it avoids a serious issue for some people. I have verified that with gcc-5.3.1, the object code we get is better in the new version both with and without CONFIG_KASAN, as we no longer allocate a 1344 byte stack frame for hns_dsaf_get_regs() but otherwise have practically identical object code.

With gcc-7.0.0, removing ACCESS_ONCE has no effect, the object code is already good either way.

This patch is probably not urgent to get into 4.11 as only KASAN=y builds with certain compilers are affected, but I still think it makes sense to backport into older kernels.

---
 drivers/net/ethernet/hisilicon/hns/hns_dsaf_reg.h |    8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

@@ -900,9 +900,7 @@
 
 static inline void dsaf_write_reg(void __iomem *base, u32 reg, u32 value)
 {
-	u8 __iomem *reg_addr = ACCESS_ONCE(base);
-
-	writel(value, reg_addr + reg);
+	writel(value, base + reg);
 }
 
 #define dsaf_write_dev(a, reg, value) \

@@ -910,9 +908,7 @@ static inline void dsaf_write_reg(void _
 
 static inline u32 dsaf_read_reg(u8 __iomem *base, u32 reg)
 {
-	u8 __iomem *reg_addr = ACCESS_ONCE(base);
-
-	return readl(reg_addr + reg);
+	return readl(base + reg);
 }
 
 #define dsaf_read_dev(a, reg) \

Patches currently in stable-queue which might be from [email protected] are

queue-4.4/hns-avoid-stack-overflow-with-config_kasan.patch