This is a note to let you know that I've just added the patch titled
x86/mm: Make mmap(MAP_32BIT) work correctly
to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git%3Ba=su...
The filename of the patch is: x86-mm-make-mmap-map_32bit-work-correctly.patch and it can be found in the queue-4.9 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree, please let stable@vger.kernel.org know about it.
From foo@baz Sun Mar 18 16:55:33 CET 2018
From: Dmitry Safonov dsafonov@virtuozzo.com Date: Mon, 6 Mar 2017 17:17:20 +0300 Subject: x86/mm: Make mmap(MAP_32BIT) work correctly
From: Dmitry Safonov dsafonov@virtuozzo.com
[ Upstream commit 3e6ef9c80946f781fc25e8490c9875b1d2b61158 ]
mmap(MAP_32BIT) is broken due to the dependency on the TIF_ADDR32 thread flag.
For 64bit applications MAP_32BIT will force legacy bottom-up allocations and the 1GB address space restriction even if the application issued a compat syscall, which should not be subject of these restrictions.
For 32bit applications, which issue 64bit syscalls the newly introduced mmap base separation into 64-bit and compat bases changed the behaviour because now a 64-bit mapping is returned, but due to the TIF_ADDR32 dependency MAP_32BIT is ignored. Before the separation a 32-bit mapping was returned, so the MAP_32BIT handling was irrelevant.
Replace the check for TIF_ADDR32 with a check for the compat syscall. That solves both the 64-bit issuing a compat syscall and the 32-bit issuing a 64-bit syscall problems.
[ tglx: Massaged changelog ]
Signed-off-by: Dmitry Safonov dsafonov@virtuozzo.com Cc: 0x7f454c46@gmail.com Cc: linux-mm@kvack.org Cc: Andy Lutomirski luto@kernel.org Cc: Cyrill Gorcunov gorcunov@openvz.org Cc: Borislav Petkov bp@suse.de Cc: "Kirill A. Shutemov" kirill.shutemov@linux.intel.com Link: http://lkml.kernel.org/r/20170306141721.9188-5-dsafonov@virtuozzo.com Signed-off-by: Thomas Gleixner tglx@linutronix.de Signed-off-by: Sasha Levin alexander.levin@microsoft.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- arch/x86/kernel/sys_x86_64.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
--- a/arch/x86/kernel/sys_x86_64.c +++ b/arch/x86/kernel/sys_x86_64.c @@ -100,7 +100,7 @@ out: static void find_start_end(unsigned long flags, unsigned long *begin, unsigned long *end) { - if (!test_thread_flag(TIF_ADDR32) && (flags & MAP_32BIT)) { + if (!in_compat_syscall() && (flags & MAP_32BIT)) { /* This is usually used needed to map code in small model, so it needs to be in the first 31bit. Limit it to that. This means we need to move the @@ -175,7 +175,7 @@ arch_get_unmapped_area_topdown(struct fi return addr;
/* for MAP_32BIT mappings we force the legacy mmap base */ - if (!test_thread_flag(TIF_ADDR32) && (flags & MAP_32BIT)) + if (!in_compat_syscall() && (flags & MAP_32BIT)) goto bottomup;
/* requesting a specific address */
Patches currently in stable-queue which might be from dsafonov@virtuozzo.com are
queue-4.9/x86-mm-make-mmap-map_32bit-work-correctly.patch