linux-kernel - Re: [PATCH v6 0/5] riscv: Introduce KASLR

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: <93D8412E2714688D+f87fdf42-4c77-ec81-c46f-30eda8ec44ee@tinylab.org>
Date:   Thu, 17 Aug 2023 21:27:01 +0800
From:   Song Shuai <songshuaishuai@...ylab.org>
To:     Alexandre Ghiti <alexghiti@...osinc.com>
Cc:     Paul Walmsley <paul.walmsley@...ive.com>,
        Palmer Dabbelt <palmer@...belt.com>,
        Ard Biesheuvel <ardb@...nel.org>,
        Kees Cook <keescook@...omium.org>,
        linux-riscv@...ts.infradead.org, linux-kernel@...r.kernel.org,
        linux-efi@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        Albert Ou <aou@...s.berkeley.edu>, zong.li@...ive.com,
        conor.dooley@...rochip.com
Subject: Re: [PATCH v6 0/5] riscv: Introduce KASLR



在 2023/8/17 21:10, Alexandre Ghiti 写道:
> Hi Song,
> 
> On Tue, Aug 15, 2023 at 1:24 PM Song Shuai <songshuaishuai@...ylab.org> wrote:
>>
>>
>> Hi, Alex:
>>
>> 在 2023/7/22 20:38, Alexandre Ghiti 写道:
>>> The following KASLR implementation allows to randomize the kernel mapping:
>>>
>>> - virtually: we expect the bootloader to provide a seed in the device-tree
>>> - physically: only implemented in the EFI stub, it relies on the firmware to
>>>     provide a seed using EFI_RNG_PROTOCOL. arm64 has a similar implementation
>>>     hence the patch 3 factorizes KASLR related functions for riscv to take
>>>     advantage.
>>>
>>> The new virtual kernel location is limited by the early page table that only
>>> has one PUD and with the PMD alignment constraint, the kernel can only take
>>> < 512 positions.
>>>
>>
>> I have gone through the code and tested this series with RiscVVirt edk2.
>> All seems good to me, you can add :
>>
>> Tested-by: Song Shuai <songshuaishuai@...ylab.org>
>>
>> And a few questions about patch 2 ("riscv: Dump out kernel offset
>> information on panic"):
>>
>> 1. The dump_kernel_offset() function would output "Kernel Offset: 0x0
>> from 0xffffffff80000000" when booting with "nokaslr" option.
>>
>> How about disabling the registration of "kernel_offset_notifier" with
>> "nokaslr" option?
> 
> I'd rather keep it as it shows the "nokaslr" flag was taken into account.
> 
>>
>> 2. Inspired by patch 2, I added the Crash KASLR support based on this
>> series [1].
>> So is it necessary to keep this patch if we have Crash KASLR support?
> 
> I don't understand your question here

Crash can automatically calculate virt_offset by comparing the vmlinux 
and vmcore. If this patch is just intended to assist Crash in setting 
the "--kaslr offset," it might be deleted; if not just keep it. 	

> 
>>
>>
>> [1]:
>> https://lore.kernel.org/linux-riscv/20230815104800.705753-1-songshuaishuai@tinylab.org/T/#u
>>
>>> base-commit-tag: v6.5-rc1
>>>
>>> Changes in v6:
>>>     * Fix reintroduced build failures by compiling kaslr.c only for arm64
>>>       and riscv, as suggested by Ard
>>>
>>> Changes in v5:
>>>     * Renamed efi-stub-kaslr.c into kaslr.c and fix commit log of patch 3,
>>>       as suggested by Ard
>>>     * Removed stubs since the kaslr functions were moved to their own file
>>>       (and then does not trigger any build failure for architectures that do
>>>       not call those functions since they are in their own compilation unit)
>>>
>>> Changes in v4:
>>>     * Fix efi_get_kimg macro that returned nothing
>>>     * Moved new kaslr functions into their own files to avoid zboot link
>>>       failures, as suggested by Ard
>>>
>>> Changes in v3:
>>>     * Rebase on top of 6.4-rc2
>>>     * Make RANDOMIZE_BASE depend on 64bit
>>>     * Fix efi_icache_sync and efi_get_kimg_min_align which were undefined
>>>       in x86 (and certainly other archs)
>>>     * Add patch 4 to fix warning on rv32
>>>
>>> Changes in v2:
>>>     * Rebase on top of 6.3-rc1
>>>     * Add a riscv cache sync after memcpying the kernel
>>>     * Add kaslr_offset implementation for KCOV
>>>     * Add forward declaration to quiet LLVM
>>>
>>> Alexandre Ghiti (5):
>>>     riscv: Introduce virtual kernel mapping KASLR
>>>     riscv: Dump out kernel offset information on panic
>>>     arm64: libstub: Move KASLR handling functions to kaslr.c
>>>     libstub: Fix compilation warning for rv32
>>>     riscv: libstub: Implement KASLR by using generic functions
>>>
>>>    arch/arm64/include/asm/efi.h              |   2 +
>>>    arch/riscv/Kconfig                        |  19 +++
>>>    arch/riscv/include/asm/efi.h              |   2 +
>>>    arch/riscv/include/asm/page.h             |   3 +
>>>    arch/riscv/kernel/image-vars.h            |   1 +
>>>    arch/riscv/kernel/pi/Makefile             |   2 +-
>>>    arch/riscv/kernel/pi/cmdline_early.c      |  13 ++
>>>    arch/riscv/kernel/pi/fdt_early.c          |  30 ++++
>>>    arch/riscv/kernel/setup.c                 |  25 ++++
>>>    arch/riscv/mm/init.c                      |  36 ++++-
>>>    drivers/firmware/efi/libstub/Makefile     |   4 +-
>>>    drivers/firmware/efi/libstub/arm64-stub.c | 117 ++--------------
>>>    drivers/firmware/efi/libstub/efistub.h    |   8 ++
>>>    drivers/firmware/efi/libstub/kaslr.c      | 159 ++++++++++++++++++++++
>>>    drivers/firmware/efi/libstub/riscv-stub.c |  33 ++---
>>>    15 files changed, 328 insertions(+), 126 deletions(-)
>>>    create mode 100644 arch/riscv/kernel/pi/fdt_early.c
>>>    create mode 100644 drivers/firmware/efi/libstub/kaslr.c
>>>
>>
>> --
>> Thanks
>> Song Shuai
> 
> Thanks for testing this and your suggestions!
> 
> Alex
> 

-- 
Thanks
Song Shuai