Re: tracing: Create a sparse bitmask for pid filtering

[Colin Ian King <colin.king@canonical.com>]

On 07/10/2021 14:51, Steven Rostedt wrote:
> On Thu, 7 Oct 2021 12:26:32 +0100
> Colin Ian King <colin.king@canonical.com> wrote:
> 
>> Hi,
>>
>> Static analysis on linux-next with Coverity has identified two issues
>> with reads of initialized pointers in the following commit:
>>
>> commit 8d6e90983ade25ec7925211ac31d9ccaf64b7edf
>> Author: Steven Rostedt (VMware) <rostedt@goodmis.org>
>> Date:   Thu Sep 23 22:20:57 2021 -0400
>>
>>       tracing: Create a sparse bitmask for pid filtering
>>
>> The analysis is as follows:
>>
>> 332 static void pid_list_refill_irq(struct irq_work *iwork)
>> 333 {
>>
>>      1. Condition 0 /* !!(!__builtin_types_compatible_p() &&
>> !__builtin_types_compatible_p()) */, taking false branch.
> 
> What does the above mean?
> 
>>
>> 334        struct trace_pid_list *pid_list = container_of(iwork, struct
>> trace_pid_list,
>> 335                                                       refill_irqwork);
>>
>>      2. var_decl: Declaring variable upper without initializer.
> 
> Hmm, I think this is legit. I should have both upper and lower initialized
> as NULL.
> 
>>
>> 336        union upper_chunk *upper;
>> 337        union lower_chunk *lower;
>> 338        union upper_chunk **upper_next = &upper;
>> 339        union lower_chunk **lower_next = &lower;
>> 340        int upper_count;
>> 341        int lower_count;
>> 342        int ucnt = 0;
>> 343        int lcnt = 0;
>> 344
>> 345 again:
>> 346        raw_spin_lock(&pid_list->lock);
>> 347        upper_count = CHUNK_ALLOC - pid_list->free_upper_chunks;
>> 348        lower_count = CHUNK_ALLOC - pid_list->free_lower_chunks;
>> 349        raw_spin_unlock(&pid_list->lock);
>> 350
>>
>>      3. Condition upper_count <= 0, taking false branch.
> 
> What does the above mean?
> 
>>
>> 351        if (upper_count <= 0 && lower_count <= 0)
>> 352                return;
>> 353
>>
>>      4. Condition upper_count-- > 0, taking true branch.
>>
>> 354        while (upper_count-- > 0) {
>> 355                union upper_chunk *chunk;
>> 356
>> 357                chunk = kzalloc(sizeof(*chunk), GFP_KERNEL);
>>
>>      5. Condition !chunk, taking true branch.
>> 358                if (!chunk)
>>      6. Breaking from loop.
>>
>> 359                        break;
>> 360                *upper_next = chunk;
>> 361                upper_next = &chunk->next;
>> 362                ucnt++;
>> 363        }
>> 364
>>
>>      7. Condition lower_count-- > 0, taking true branch.
>>
>> 365        while (lower_count-- > 0) {
>> 366                union lower_chunk *chunk;
>> 367
>> 368                chunk = kzalloc(sizeof(*chunk), GFP_KERNEL);
>>
>>      8. Condition !chunk, taking true branch.
>>
>> 369                if (!chunk)
>>
>>      9. Breaking from loop.
>>
>> 370                        break;
>> 371                *lower_next = chunk;
>> 372                lower_next = &chunk->next;
>> 373                lcnt++;
>> 374        }
>> 375
>> 376        raw_spin_lock(&pid_list->lock);
>>
>>       Uninitialized pointer read (UNINIT)
>>       10. uninit_use: Using uninitialized value upper.
> 
> Agreed.
> 
>>
>> 377        if (upper) {
>> 378                *upper_next = pid_list->upper_list;
>> 379                pid_list->upper_list = upper;
>> 380                pid_list->free_upper_chunks += ucnt;
>> 381        }
>>
>>       Uninitialized pointer read (UNINIT)
>>       11. uninit_use: Using uninitialized value lower.
> 
> Agreed.
> 
>>
>> 382        if (lower) {
>> 383                *lower_next = pid_list->lower_list;
>> 384                pid_list->lower_list = lower;
>> 385                pid_list->free_lower_chunks += lcnt;
>> 386        }
>> 387        raw_spin_unlock(&pid_list->lock);
>> 388
>>
>> Colin
> 
> So is this just a fancy way of saying that upper and lower were
> uninitialized?

Basically, yes. But it shows how the static analyzer determined this :-)

> 
> -- Steve
>