Re: [PATCH v2 3/3] powerpc/kexec_file: Prevent kexec range truncation

[Jinjie Ruan]

在 2026/7/29 14:44, Sourabh Jain 写道:
> 
> 
> On 29/07/26 11:56, Jinjie Ruan wrote:
>>
>> 在 2026/7/29 12:43, Sourabh Jain 写道:
>>>
>>> On 29/07/26 06:59, Jinjie Ruan wrote:
>>>> Sashiko AI review pointed out the following issue.
>>>>
>>>> The __merge_memory_ranges() function incorrectly handles overlapping
>>>> memory ranges when merging them. Although sort_memory_ranges() sorts
>>>> all
>>>> ranges by their start address in ascending order beforehand, the merge
>>>> logic remains defective in two ways:
>>>>
>>>> 1. It compares the current range's start against the previous element
>>>> (i-1)
>>>>      instead of the running target index (idx)
>>>>
>>>> 2. It unconditionally overwrites 'ranges[idx].end' with
>>>> 'ranges[i].end'.
>>>>
>>>> This logic flaw leads to critical memory truncation when a larger
>>>> memory
>>>> range completely subsumes subsequent smaller ranges.
>>>>
>>>> For example, consider a sorted input array with three ranges:
>>>>     Range A (idx=0): [0x1000 - 0x9000]
>>>>     Range B (i=1):   [0x2000 - 0x5000] (completely inside Range A)
>>>>     Range C (i=2):   [0x6000 - 0x8000] (completely inside Range A)
>>>>
>>>> 1. When i=1 (Range B):
>>>>      ranges[1].start (0x2000) <= ranges[0].end + 1 (0x9001) is TRUE.
>>>>      The code executes: ranges[0].end = ranges[1].end, which
>>>> erroneously
>>>>      shrinks Range A's end from 0x9000 down to 0x5000.
>>>>
>>>> 2. When i=2 (Range C):
>>>>      ranges[2].start (0x6000) <= ranges[1].end + 1 (0x5001) is FALSE.
>>>>      The code falls into the else block, creating a broken new range.
>>>>
>>>> As a result, valid memory fragments [0x5001 - 0x5fff] and [0x8001 -
>>>> 0x9000]
>>>> are completely lost from the kexec exclude lists, potentially allowing
>>>> the crash kernel to overwrite active memory, causing data corruption
>>>> or crashes.
>>>>
>>>> Fix this by ensuring the start of the current range is compared
>>>> against the
>>>> end of the active merged range (idx), and use max() to safely
>>>> prevent the
>>>> outer boundary from being truncated.
>>>>
>>>> Cc: Sourabh Jain <sourabhjain@xxxxxxxxxxxxx>
>>>> Cc: Hari Bathini <hbathini@xxxxxxxxxxxxx>
>>>> Cc: Michael Ellerman <mpe@xxxxxxxxxxxxxx>
>>>> Cc: stable@xxxxxxxxxxxxxxx
>>>> Fixes: 180adfc532a8 ("powerpc/kexec_file: Add helper functions for
>>>> getting memory ranges")
>>>> Signed-off-by: Jinjie Ruan <ruanjinjie@xxxxxxxxxx>
>>>> ---
>>>>    arch/powerpc/kexec/ranges.c | 12 +++++-------
>>>>    1 file changed, 5 insertions(+), 7 deletions(-)
>>>>
>>>> diff --git a/arch/powerpc/kexec/ranges.c b/arch/powerpc/kexec/ranges.c
>>>> index e5fea23b191b..539061d14a77 100644
>>>> --- a/arch/powerpc/kexec/ranges.c
>>>> +++ b/arch/powerpc/kexec/ranges.c
>>>> @@ -21,6 +21,7 @@
>>>>    #include <linux/of.h>
>>>>    #include <linux/slab.h>
>>>>    #include <linux/memblock.h>
>>>> +#include <linux/minmax.h>
>>>>    #include <linux/crash_core.h>
>>>>    #include <asm/sections.h>
>>>>    #include <asm/kexec_ranges.h>
>>>> @@ -105,19 +106,16 @@ static void __merge_memory_ranges(struct
>>>> crash_mem *mem_rngs)
>>>>        struct range *ranges;
>>>>        int i, idx;
>>>>    -    if (!mem_rngs)
>>>> +    if (!mem_rngs || mem_rngs->nr_ranges <= 1)
>>>>            return;
>>> Although the below loop handles this but it is good to return early when
>>> there is
>>> only range.
>>>
>>>>          idx = 0;
>>>> -    ranges = &(mem_rngs->ranges[0]);
>>>> +    ranges = mem_rngs->ranges;
>>>>        for (i = 1; i < mem_rngs->nr_ranges; i++) {
>>>> -        if (ranges[i].start <= (ranges[i-1].end + 1))
>>>> -            ranges[idx].end = ranges[i].end;
>>>> +        if (ranges[i].start <= (ranges[idx].end + 1))
>>>> +            ranges[idx].end = max(ranges[idx].end, ranges[i].end);
>>> Yeah this changes is needed.
>>>
>>>>            else {
>>>>                idx++;
>>>> -            if (i == idx)
>>>> -                continue;
>>> Do we really need to remove the above condition?
>>>
>>> Isn't this condition is helpful till we find an overlap?
>> Hi Sourabh,
>>
>> I believe there is no functional change here. The else branch itself
>> already indicates that there will be no overlap this time, so we can
>> safely use ranges[i] as the next memory region.
> 
> I agree that keeping or removing it doesn't change the functionality.
> 
> My point is that it helps avoid copying a range onto itself. For example:
> 
> Range A (idx = 0): [0x1000 - 0x2000]
> Range B (idx = 1): [0x3000 - 0x5000]
> Range C (idx = 2): [0x6000 - 0x8000]
> 
> For all three ranges,  idx = i = x and kernel will do:
> ranges[x] = ranges[x];
> 
> This self-assignment can be avoided if we keep the condition instead of
> removing it. Isn't it?

Yes, with the continue we can avoid the self-assignment.

If this review is fine, can it be updated and submitted directly, or
does it need to be updated in the new version?

> 
> - Sourabh Jain
> 
> 
>> Whether we remove it or keep it is fine.
>>
>>> - Sourabh Jain
>>>
>>>> -
>>>>                ranges[idx] = ranges[i];
>>>>            }
>>>>        }
>