2023-05-27 12:45:24

by Zhen Lei

Subject: [PATCH 2/6] kexec: delete a useless check in crash_shrink_memory()

The check '(crashk_res.parent != NULL)' is added by
commit e05bd3367bd3 ("kexec: fix Oops in crash_shrink_memory()"), but it's
stale now. Because if 'crashk_res' is not reserved, it will be zero in
size and will be intercepted by the above 'if (new_size >= old_size)'.

Ago:
if (new_size >= end - start + 1)

Now:
old_size = (end == 0) ? 0 : end - start + 1;
if (new_size >= old_size)

Signed-off-by: Zhen Lei <[email protected]>
---
kernel/kexec_core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
index 22acee18195a591..d1ab139dd49035e 100644
--- a/kernel/kexec_core.c
+++ b/kernel/kexec_core.c
@@ -1137,7 +1137,7 @@ int crash_shrink_memory(unsigned long new_size)
end = start + new_size;
crash_free_reserved_phys_range(end, crashk_res.end);

- if ((start == end) && (crashk_res.parent != NULL))
+ if (start == end)
release_resource(&crashk_res);

ram_res->start = end;
--
2.25.1

2023-05-31 00:34:19

by Baoquan He

Subject: Re: [PATCH 2/6] kexec: delete a useless check in crash_shrink_memory()

On 05/27/23 at 08:34pm, Zhen Lei wrote:
> The check '(crashk_res.parent != NULL)' is added by
> commit e05bd3367bd3 ("kexec: fix Oops in crash_shrink_memory()"), but it's
> stale now. Because if 'crashk_res' is not reserved, it will be zero in
> size and will be intercepted by the above 'if (new_size >= old_size)'.
>
> Ago:
> if (new_size >= end - start + 1)
>
> Now:
> old_size = (end == 0) ? 0 : end - start + 1;
> if (new_size >= old_size)

Hmm, I would strongly suggest we keep that check. Even though the
current code like above can do the acutal checking, but its actual usage
is not obvious for checking of crashk_res existence. In the future,
someone may change above calculation and don't notice the hidden
functionality at all behind the calculation. The cost of the check is
almost zero, right?

>
> Signed-off-by: Zhen Lei <[email protected]>
> ---
> kernel/kexec_core.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
> index 22acee18195a591..d1ab139dd49035e 100644
> --- a/kernel/kexec_core.c
> +++ b/kernel/kexec_core.c
> @@ -1137,7 +1137,7 @@ int crash_shrink_memory(unsigned long new_size)
> end = start + new_size;
> crash_free_reserved_phys_range(end, crashk_res.end);
>
> - if ((start == end) && (crashk_res.parent != NULL))
> + if (start == end)
> release_resource(&crashk_res);
>
> ram_res->start = end;
> --
> 2.25.1
>

2023-05-31 02:36:06

by Zhen Lei

Subject: Re: [PATCH 2/6] kexec: delete a useless check in crash_shrink_memory()

On 2023/5/31 8:17, Baoquan He wrote:
> On 05/27/23 at 08:34pm, Zhen Lei wrote:
>> The check '(crashk_res.parent != NULL)' is added by
>> commit e05bd3367bd3 ("kexec: fix Oops in crash_shrink_memory()"), but it's
>> stale now. Because if 'crashk_res' is not reserved, it will be zero in
>> size and will be intercepted by the above 'if (new_size >= old_size)'.
>>
>> Ago:
>> if (new_size >= end - start + 1)
>>
>> Now:
>> old_size = (end == 0) ? 0 : end - start + 1;
>> if (new_size >= old_size)
>
> Hmm, I would strongly suggest we keep that check. Even though the
> current code like above can do the acutal checking, but its actual usage
> is not obvious for checking of crashk_res existence. In the future,
> someone may change above calculation and don't notice the hidden
> functionality at all behind the calculation. The cost of the check is
> almost zero, right?

The cost of the check is negligible. The only downside is that it's hard to
understand why it's added, and I only found out why by looking at the history
log. In my opinion, the above 'Now:' is the right fix.

>
>>
>> Signed-off-by: Zhen Lei <[email protected]>
>> ---
>> kernel/kexec_core.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
>> index 22acee18195a591..d1ab139dd49035e 100644
>> --- a/kernel/kexec_core.c
>> +++ b/kernel/kexec_core.c
>> @@ -1137,7 +1137,7 @@ int crash_shrink_memory(unsigned long new_size)
>> end = start + new_size;
>> crash_free_reserved_phys_range(end, crashk_res.end);
>>
>> - if ((start == end) && (crashk_res.parent != NULL))
>> + if (start == end)
>> release_resource(&crashk_res);
>>
>> ram_res->start = end;
>> --
>> 2.25.1
>>
>
> .
>

--
Regards,
Zhen Lei

2023-05-31 07:51:22

by Baoquan He

Subject: Re: [PATCH 2/6] kexec: delete a useless check in crash_shrink_memory()

On 05/31/23 at 10:19am, Leizhen (ThunderTown) wrote:
>
>
> On 2023/5/31 8:17, Baoquan He wrote:
> > On 05/27/23 at 08:34pm, Zhen Lei wrote:
> >> The check '(crashk_res.parent != NULL)' is added by
> >> commit e05bd3367bd3 ("kexec: fix Oops in crash_shrink_memory()"), but it's
> >> stale now. Because if 'crashk_res' is not reserved, it will be zero in
> >> size and will be intercepted by the above 'if (new_size >= old_size)'.
> >>
> >> Ago:
> >> if (new_size >= end - start + 1)
> >>
> >> Now:
> >> old_size = (end == 0) ? 0 : end - start + 1;
> >> if (new_size >= old_size)
> >
> > Hmm, I would strongly suggest we keep that check. Even though the
> > current code like above can do the acutal checking, but its actual usage
> > is not obvious for checking of crashk_res existence. In the future,
> > someone may change above calculation and don't notice the hidden
> > functionality at all behind the calculation. The cost of the check is
> > almost zero, right?
>
> The cost of the check is negligible. The only downside is that it's hard to
> understand why it's added, and I only found out why by looking at the history
> log. In my opinion, the above 'Now:' is the right fix.

It checks if the resource exists before releasing, just a normal
checking?
>
> >
> >>
> >> Signed-off-by: Zhen Lei <[email protected]>
> >> ---
> >> kernel/kexec_core.c | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
> >> index 22acee18195a591..d1ab139dd49035e 100644
> >> --- a/kernel/kexec_core.c
> >> +++ b/kernel/kexec_core.c
> >> @@ -1137,7 +1137,7 @@ int crash_shrink_memory(unsigned long new_size)
> >> end = start + new_size;
> >> crash_free_reserved_phys_range(end, crashk_res.end);
> >>
> >> - if ((start == end) && (crashk_res.parent != NULL))
> >> + if (start == end)
> >> release_resource(&crashk_res);
> >>
> >> ram_res->start = end;
> >> --
> >> 2.25.1
> >>
> >
> > .
> >
>
> --
> Regards,
> Zhen Lei
>

2023-05-31 08:38:44

by Zhen Lei

Subject: Re: [PATCH 2/6] kexec: delete a useless check in crash_shrink_memory()

On 2023/5/31 15:41, Baoquan He wrote:
> On 05/31/23 at 10:19am, Leizhen (ThunderTown) wrote:
>>
>>
>> On 2023/5/31 8:17, Baoquan He wrote:
>>> On 05/27/23 at 08:34pm, Zhen Lei wrote:
>>>> The check '(crashk_res.parent != NULL)' is added by
>>>> commit e05bd3367bd3 ("kexec: fix Oops in crash_shrink_memory()"), but it's
>>>> stale now. Because if 'crashk_res' is not reserved, it will be zero in
>>>> size and will be intercepted by the above 'if (new_size >= old_size)'.
>>>>
>>>> Ago:
>>>> if (new_size >= end - start + 1)
>>>>
>>>> Now:
>>>> old_size = (end == 0) ? 0 : end - start + 1;
>>>> if (new_size >= old_size)
>>>
>>> Hmm, I would strongly suggest we keep that check. Even though the
>>> current code like above can do the acutal checking, but its actual usage
>>> is not obvious for checking of crashk_res existence. In the future,
>>> someone may change above calculation and don't notice the hidden
>>> functionality at all behind the calculation. The cost of the check is
>>> almost zero, right?
>>
>> The cost of the check is negligible. The only downside is that it's hard to
>> understand why it's added, and I only found out why by looking at the history
>> log. In my opinion, the above 'Now:' is the right fix.
>
> It checks if the resource exists before releasing, just a normal
> checking?

If resource_size(&crashk_res) is zero, it means that crashk_res has not been
added(insert_resource) or has been deleted(release_resource). I've tested it. It's okay.

>>
>>>
>>>>
>>>> Signed-off-by: Zhen Lei <[email protected]>
>>>> ---
>>>> kernel/kexec_core.c | 2 +-
>>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>>
>>>> diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
>>>> index 22acee18195a591..d1ab139dd49035e 100644
>>>> --- a/kernel/kexec_core.c
>>>> +++ b/kernel/kexec_core.c
>>>> @@ -1137,7 +1137,7 @@ int crash_shrink_memory(unsigned long new_size)
>>>> end = start + new_size;
>>>> crash_free_reserved_phys_range(end, crashk_res.end);
>>>>
>>>> - if ((start == end) && (crashk_res.parent != NULL))
>>>> + if (start == end)
>>>> release_resource(&crashk_res);
>>>>
>>>> ram_res->start = end;
>>>> --
>>>> 2.25.1
>>>>
>>>
>>> .
>>>
>>
>> --
>> Regards,
>> Zhen Lei
>>
>
>
> .
>

--
Regards,
Zhen Lei