linux-kernel - Re: [PATCH 3/3] riscv: mm: Do not restrict mmap address based on hint

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <mhng-ad0c8c3b-568a-484d-bc3d-49b56a11dcd6@palmer-ri-x1c9>
Date: Tue, 03 Sep 2024 07:27:03 -0700 (PDT)
From: Palmer Dabbelt <palmer@...osinc.com>
To: cyy@...self.name
CC: Charlie Jenkins <charlie@...osinc.com>, corbet@....net,
  Paul Walmsley <paul.walmsley@...ive.com>, aou@...s.berkeley.edu, shuah@...nel.org, rsworktech@...look.com,
  alexghiti@...osinc.com, linux-doc@...r.kernel.org, linux-riscv@...ts.infradead.org,
  linux-kernel@...r.kernel.org, linux-kselftest@...r.kernel.org
Subject:     Re: [PATCH 3/3] riscv: mm: Do not restrict mmap address based on hint

On Mon, 26 Aug 2024 19:24:38 PDT (-0700), cyy@...self.name wrote:
>
>
>> On Aug 27, 2024, at 00:36, Charlie Jenkins <charlie@...osinc.com> wrote:
>> 
>> The hint address should not forcefully restrict the addresses returned
>> by mmap as this causes mmap to report ENOMEM when there is memory still
>> available.
>> 
>
> Fixing in this way will break userspace on Sv57 machines as some
> issues mentioned in the patch [1].
>
> I suggest restricting to BIT(47) by default, like patch [2], to
> align with kernel behavior on x86 and aarch64, and this does exist
> on x86 and aarch64 for quite a long time. In that way, we will also
> solve the problem mentioned in the first patch [1], as QEMU enables
> Sv57 by default now and will not break userspace.
>
> [1] https://lore.kernel.org/linux-riscv/20230809232218.849726-1-charlie@rivosinc.com/
> [2] https://lore.kernel.org/linux-riscv/tencent_B2D0435BC011135736262764B511994F4805@qq.com/

I'm going to pick this up as it's a revert and a bug fix, so we can 
backport it.  If the right answer is to just forget about the sv39 
userspace and only worry about sv48 userspace then your patches are 
likely the way to go, but there's a handful of discussions around that 
which might take a bit.

>
> Thanks,
> Yangyu Chen
>
>> Signed-off-by: Charlie Jenkins <charlie@...osinc.com>
>> Fixes: b5b4287accd7 ("riscv: mm: Use hint address in mmap if available")
>> Fixes: add2cc6b6515 ("RISC-V: mm: Restrict address space for sv39,sv48,sv57")
>> Closes: https://lore.kernel.org/linux-kernel/ZbxTNjQPFKBatMq+@ghost/T/#mccb1890466bf5a488c9ce7441e57e42271895765
>> ---
>> arch/riscv/include/asm/processor.h | 26 ++------------------------
>> 1 file changed, 2 insertions(+), 24 deletions(-)
>> 
>> diff --git a/arch/riscv/include/asm/processor.h b/arch/riscv/include/asm/processor.h
>> index 8702b8721a27..efa1b3519b23 100644
>> --- a/arch/riscv/include/asm/processor.h
>> +++ b/arch/riscv/include/asm/processor.h
>> @@ -14,36 +14,14 @@
>> 
>> #include <asm/ptrace.h>
>> 
>> -/*
>> - * addr is a hint to the maximum userspace address that mmap should provide, so
>> - * this macro needs to return the largest address space available so that
>> - * mmap_end < addr, being mmap_end the top of that address space.
>> - * See Documentation/arch/riscv/vm-layout.rst for more details.
>> - */
>> #define arch_get_mmap_end(addr, len, flags) \
>> ({ \
>> - unsigned long mmap_end; \
>> - typeof(addr) _addr = (addr); \
>> - if ((_addr) == 0 || is_compat_task() || \
>> -    ((_addr + len) > BIT(VA_BITS - 1))) \
>> - mmap_end = STACK_TOP_MAX; \
>> - else \
>> - mmap_end = (_addr + len); \
>> - mmap_end; \
>> + STACK_TOP_MAX; \
>> })
>> 
>> #define arch_get_mmap_base(addr, base) \
>> ({ \
>> - unsigned long mmap_base; \
>> - typeof(addr) _addr = (addr); \
>> - typeof(base) _base = (base); \
>> - unsigned long rnd_gap = DEFAULT_MAP_WINDOW - (_base); \
>> - if ((_addr) == 0 || is_compat_task() || \
>> -    ((_addr + len) > BIT(VA_BITS - 1))) \
>> - mmap_base = (_base); \
>> - else \
>> - mmap_base = (_addr + len) - rnd_gap; \
>> - mmap_base; \
>> + base; \
>> })
>> 
>> #ifdef CONFIG_64BIT
>> 
>> -- 
>> 2.45.0
>>