[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Fri, 24 Apr 2020 17:52:41 +0700
From: Phong Tran <tranmanphong@...il.com>
To: Will Deacon <will@...nel.org>, Mark Rutland <mark.rutland@....com>,
ruscur@...sell.cc
Cc: tranmanphong@...il.com, steve.capper@....com, steven.price@....com,
keescook@...omium.org, greg@...ah.com, akpm@...ux-foundation.org,
alexios.zavras@...el.com, broonie@...nel.org,
kernel-hardening@...ts.openwall.com, linux-arm-kernel@...ts.infradead.org,
linux-kernel@...r.kernel.org, tglx@...utronix.de
Subject: Re: Re: [PATCH v2] arm64: add check_wx_pages debugfs for CHECK_WX
On 4/22/20 10:26 PM, Will Deacon wrote:
> On Wed, Apr 22, 2020 at 03:35:27PM +0100, Mark Rutland wrote:
>> On Wed, Apr 22, 2020 at 12:35:58AM +0700, Phong Tran wrote:
>>> follow the suggestion from
>>> https://github.com/KSPP/linux/issues/35
>>>
>>> Signed-off-by: Phong Tran <tranmanphong@...il.com>
>>> ---
>>> Change since v1:
>>> - Update the Kconfig help text
>>> - Don't check the return value of debugfs_create_file()
>>> - Tested on QEMU aarch64
>>
>> As on v1, I think that this should be generic across all architectures
>> with CONFIG_DEBUG_WX. Adding this only on arm64 under
>> CONFIG_PTDUMP_DEBUGFS (which does not generally imply a WX check)
>> doesn't seem right.
>>
>> Maybe we need a new ARCH_HAS_CHECK_WX config to make that simpler, but
>> that seems simple to me.
>
> Agreed. When I asked about respinning, I assumed this would be done in
> generic code as you requested on the first version. Phong -- do you think
> you can take a look at that, please?
>
sure, plan to make change in mm/ptdump.c with KConfig
"ARCH_HAS_CHECK_WX" as suggestion.
Then need align with this patch in PowerPC arch also
https://lore.kernel.org/kernel-hardening/20200402084053.188537-3-ruscur@russell.cc/
diff --git a/arch/arm64/Kconfig.debug b/arch/arm64/Kconfig.debug
index a1efa246c9ed..50f18e7ff2ae 100644
--- a/arch/arm64/Kconfig.debug
+++ b/arch/arm64/Kconfig.debug
@@ -25,6 +25,7 @@ config ARM64_RANDOMIZE_TEXT_OFFSET
config DEBUG_WX
bool "Warn on W+X mappings at boot"
+ select ARCH_HAS_CHECK_WX
select PTDUMP_CORE
---help---
Generate a warning if any W+X mappings are found at boot.
diff --git a/mm/Kconfig.debug b/mm/Kconfig.debug
index 0271b22e063f..40c9ac5a4db2 100644
--- a/mm/Kconfig.debug
+++ b/mm/Kconfig.debug
@@ -138,3 +138,6 @@ config PTDUMP_DEBUGFS
kernel.
If in doubt, say N.
+
+config ARCH_HAS_CHECK_WX
+ bool
diff --git a/mm/ptdump.c b/mm/ptdump.c
index 26208d0d03b7..8f54db007aaa 100644
--- a/mm/ptdump.c
+++ b/mm/ptdump.c
@@ -137,3 +137,29 @@ void ptdump_walk_pgd(struct ptdump_state *st,
struct mm_struct *mm, pgd_t *pgd)
/* Flush out the last page */
st->note_page(st, 0, -1, 0);
}
+
+#ifdef CONFIG_ARCH_HAS_CHECK_WX
+#include <linux/debugfs.h>
+#include <asm/ptdump.h>
+
+static int check_wx_debugfs_set(void *data, u64 val)
+{
+ if (val != 1ULL)
+ return -EINVAL;
+
+ ptdump_check_wx();
+
+ return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(check_wx_fops, NULL, check_wx_debugfs_set,
"%llu\n");
+
+static int ptdump_debugfs_check_wx_init(void)
+{
+ debugfs_create_file("check_wx_pages", 0200, NULL,
+ NULL, &check_wx_fops) ? 0 : -ENOMEM;
+ return 0;
+}
+
+device_initcall(ptdump_debugfs_check_wx_init);
+#endif
Regards,
Phong.
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
