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] [day] [month] [year] [list]
Message-ID: <ED6B71C237EB49E8+8e70fa30-f5fe-4d26-91c9-8498938a55b1@uniontech.com>
Date: Fri, 5 Dec 2025 17:20:46 +0800
From: Qiang Ma <maqianga@...ontech.com>
To: Sourabh Jain <sourabhjain@...ux.ibm.com>, akpm@...ux-foundation.org,
 bhe@...hat.com
Cc: kexec@...ts.infradead.org, linux-kernel@...r.kernel.org,
 Simon Horman <horms@...nel.org>
Subject: Re: [PATCH v3 0/3] kexec: print out debugging message if required for
 kexec_load


在 2025/12/4 20:18, Sourabh Jain 写道:
> Hello All,
>
> I read this patch series, including the older versions, and this is 
> what I
> think about the changes.
>
> There are two main reasons I see for why this patch series is proposed
> (these were not mentioned in the cover letter or commit messages):
>
> 1. Commit eb7622d908a09 ("kexec_file, riscv: print out debugging message
>    if required") and commit 6f8c1da071a4 ("kexec_file, arm64: print out
>    debugging message if required") moved some kexec-related debug 
> prints to
>    kexec_dprintk instead of pr_debug or pr_notice.
>
>    These commits made it easier to control the updated debug prints using
>    the -d option of kexec tool for kexec_file_load system call. But since
>    kexec_dprintk is not available for the kexec_load system call, 
> those prints
>    are no longer available for kexec_load at all. So patch series try 
> to bring
>    those logs for kexec_load system call.
>
> 2. Patch 03/03 prints additional logs about the kexec image type, start,
>     head, and flags via kexec_dprintk for the kexec_load system call.
>
> This is done by enabling the kexec_dprintk for kexec_load, so that logs
> from point 1 and the new logs from point 2 can be printed for kexec_load
> system call.
>
> However, the patch series does not explain why these additional debug 
> logs
> are needed for kexec_load. Since most of the work involved in loading
> kexec images through kexec_load happens in the kexec userspace tool, and
> the tool already provides enough logs to debug issues, the need for these
> extra kernel logs is not clear.
As pointed out in the point 1, these two commits not only move some
kexec-related debug prints to kexec_dprintk instead of pr_debug or 
pr_notice,
but also respectively remove and delete the debug print in kexec_image_info.
These debug print initially existed for both kexec_load and 
kexec_file_load in kernel space.
Just commit cbc2fe9d9cb2 ("kexec_file: The "add kexec_file flag to 
control debug printing")
series of patches removed or deleted some parts, but this series of 
patches only added them
to the generic code under the kexec_file_load path, not to the 
kexec_load path.
Therefore, the reason for adding this part of the debug print is to add 
this deleted debug print
to the kexec_load generic code. Regarding duplicate prints related to 
architecture, as mentioned
in previous posts[1], I want remove the print(type/head of kimage) in 
arm64, and then add it
to the general code.

[1]: 
https://lore.kernel.org/lkml/4D709A5E16BE5DEC+04521049-61de-411f-85c1-cfc049ff04c5@uniontech.com/

If commit cbc2fe9d9cb2 ("kexec_file: The "add kexec_file flag to control 
debug printing" series of patches
is indeed intended to remove those kexec_image_info in kexec_load, not 
to lose them.
I think patch 3 can be removed, but patch 1, 2 are still required.
As Sourabh Jain mentioned in the point 1, the debugging print that uses 
kexec_dprintk
in the kexec_load path currently existing under arm64 is equivalent to 
being unavailable.
>
> So adding the need for the additional debug logs in the cover letter or
> commit message is needed to decide whether it is worth adding the code
> change proposed in this patch series Also, I found it hard to map the 
> commit message to the changes introduced in the commits. And finally 
> If I review the patches only based on the code changes: 01/03 The 
> commit message refers to kexec_dbg_print 
There is an error in the patch description. It should be 
kexec_file_dbg_print. I will fix it.
> as kexec_core_dbg_print. 02/03 I think kexec_dbg_print should be set 
> to false at the end of kimage_free, not at the start. 03/03 There is 
> duplication of debug logs for architectures that already print the 
> same information. For example, arm64. - 
In the kimage_file_alloc_init()/kimage_alloc_init(), if there is an 
abnormal exit before executing *rimage = image,
the image is empty, but kexec_dbg_print has been initialized.
In this case, It needs to be placed at the beginning of kimage_free to 
reset it.
> Sourabh Jain
> On 26/11/25 14:14, Qiang Ma wrote:
>> Overview:
>> =========
>> The commit a85ee18c7900 ("kexec_file: print out debugging message
>> if required") has added general code printing in kexec_file_load(),
>> but not in kexec_load().
>>      Since kexec_load and kexec_file_load are not triggered 
>> simultaneously,
>> we can unify the debug flag of kexec and kexec_file as kexec_dbg_print.
>>      Next, we need to do some things in this patchset:
>>      1. rename kexec_file_dbg_print to kexec_dbg_print
>> 2. Add KEXEC_DEBUG
>> 3. Initialize kexec_dbg_print for kexec
>> 4. Fix uninitialized struct kimage *image pointer
>> 5. Set the reset of kexec_dbg_print to kimage_free
>>
>> Testing:
>> =========
>> I did testing on x86_64, arm64 and loongarch. On x86_64, the printed 
>> messages
>> look like below:
>>
>> unset CONFIG_KEXEC_FILE:
>> [   81.502374] kexec: kexec_load: type:0, start:0x23fff7700 
>> head:0x10a4b9002 flags:0x3e0010
>>
>> set CONFIG_KEXEC_FILE
>> [   36.774228] kexec_file: kernel: 0000000066c386c8 kernel_size: 
>> 0xd78400
>> [   36.821814] kexec-bzImage64: Loaded purgatory at 0x23fffb000
>> [   36.821826] kexec-bzImage64: Loaded boot_param, command line and 
>> misc at 0x23fff9000 bufsz=0x12d0 memsz=0x2000
>> [   36.821829] kexec-bzImage64: Loaded 64bit kernel at 0x23d400000 
>> bufsz=0xd73400 memsz=0x2ab7000
>> [   36.821918] kexec-bzImage64: Loaded initrd at 0x23bd0b000 
>> bufsz=0x16f40a8 memsz=0x16f40a8
>> [   36.821920] kexec-bzImage64: Final command line is: 
>> root=/dev/mapper/test-root crashkernel=auto rd.lvm.lv=test/root
>> [   36.821925] kexec-bzImage64: E820 memmap:
>> [   36.821926] kexec-bzImage64: 0000000000000000-000000000009ffff (1)
>> [   36.821928] kexec-bzImage64: 0000000000100000-0000000000811fff (1)
>> [   36.821930] kexec-bzImage64: 0000000000812000-0000000000812fff (2)
>> [   36.821931] kexec-bzImage64: 0000000000813000-00000000bee38fff (1)
>> [   36.821933] kexec-bzImage64: 00000000bee39000-00000000beec2fff (2)
>> [   36.821934] kexec-bzImage64: 00000000beec3000-00000000bf8ecfff (1)
>> [   36.821935] kexec-bzImage64: 00000000bf8ed000-00000000bfb6cfff (2)
>> [   36.821936] kexec-bzImage64: 00000000bfb6d000-00000000bfb7efff (3)
>> [   36.821937] kexec-bzImage64: 00000000bfb7f000-00000000bfbfefff (4)
>> [   36.821938] kexec-bzImage64: 00000000bfbff000-00000000bff7bfff (1)
>> [   36.821939] kexec-bzImage64: 00000000bff7c000-00000000bfffffff (2)
>> [   36.821940] kexec-bzImage64: 00000000feffc000-00000000feffffff (2)
>> [   36.821941] kexec-bzImage64: 00000000ffc00000-00000000ffffffff (2)
>> [   36.821942] kexec-bzImage64: 0000000100000000-000000023fffffff (1)
>> [   36.872348] kexec_file: nr_segments = 4
>> [   36.872356] kexec_file: segment[0]: buf=0x000000005314ece7 
>> bufsz=0x4000 mem=0x23fffb000 memsz=0x5000
>> [   36.872370] kexec_file: segment[1]: buf=0x000000006e59b143 
>> bufsz=0x12d0 mem=0x23fff9000 memsz=0x2000
>> [   36.872374] kexec_file: segment[2]: buf=0x00000000eb7b1fc3 
>> bufsz=0xd73400 mem=0x23d400000 memsz=0x2ab7000
>> [   36.882172] kexec_file: segment[3]: buf=0x000000006af76441 
>> bufsz=0x16f40a8 mem=0x23bd0b000 memsz=0x16f5000
>> [   36.889113] kexec_file: kexec_file_load: type:0, start:0x23fffb150 
>> head:0x101a2e002 flags:0x8
>>
>> Changes in v3:
>> ==========
>> - Rename kexec_core_dbg_print to kexec_dbg_print
>> - Remove unnecessary segments prints
>> - Remove patch "kexec_file: Fix the issue of mismatch between loop 
>> variable types"
>>
>> Qiang Ma (3):
>>    kexec: Fix uninitialized struct kimage *image pointer
>>    kexec: add kexec flag to control debug printing
>>    kexec: print out debugging message if required for kexec_load
>>
>>   include/linux/kexec.h      | 9 +++++----
>>   include/uapi/linux/kexec.h | 1 +
>>   kernel/kexec.c             | 8 +++++++-
>>   kernel/kexec_core.c        | 4 +++-
>>   kernel/kexec_file.c        | 4 +---
>>   5 files changed, 17 insertions(+), 9 deletions(-)
>>
>
>

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ