发件人: lijiang <lijiang@redhat.com>
发送时间: 2023年10月19日 9:14:18
收件人: HAGIO KAZUHITO(萩尾一仁)
抄送: 陈冠有; crash-utility@redhat.com
主题: Re: [External Mail]Re: [PATCH v4] Fix: move huge compressed obj from page to zspage

[外部邮件] 此邮件来源于小米公司外部，请谨慎处理。若对邮件安全性存疑，请将邮件转发给misec@xiaomi.com进行反馈

On Thu, Oct 19, 2023 at 8:54 AM HAGIO KAZUHITO(萩尾　一仁) <k-hagio-ab@nec.com> wrote:

> This depends on kernel version number, it might be more friendly for the
> distribution to output some debug info as below:
>
> if (CRASHDEBUG(1))
> error(INFO, "ZRAM_FLAG_SHIFT=%lu\n", ZRAM_FLAG_SHIFT);
>
> Once the related kernel patches are backported to an old kernel, the
> similar checking may fail. But, at least we can know a little more with the
> crash debug info.

Sounds good, then I would like to show the shift value instead of mask
value, so changed a little. Is the attached patch OK?

Looks better, thank you, Kazu. So: Ack

Lianbo

6.1:
crash> rd 7fb60bdff000
zram_flag_shift: 13
...

RHEL8:
crash> rd 7ff668048000
zram_flag_shift: 24

4.18 and later kernels has ZRAM_LOCK, so I think usually the else-if
block isn't used and it's kind of a fail-safe for a future change.

Anyway, thank you for the good comments.

Thanks,
Kazu