RE: Re: [PATCH] trace: fix null pointer dereference in tracing_err_log_open()

[Mateusz Stachyra]

>> >From d6ef949d29b884dd77fe5e628dc71318de08868c Mon Sep 17 00:00:00 2001  
>> From: Mateusz Stachyra <m.stachyra@xxxxxxxxxxx>
>> Date: Mon, 3 Jul 2023 17:48:40 +0200
>> Subject: [PATCH] trace: fix null pointer dereference in tracing_err_log_open()
>> 
>> Fix an issue in function 'tracing_err_log_open'.
>> The function doesn't call 'seq_open' if file is opened only with
>> write permissions, which results in 'file->private_data' being left at null.
>> If we then use 'lseek' on that opened file, 'seq_lseek' dereferences
>> 'file->private_data' in 'mutex_lock(&m->lock)', resulting in a Kernel panic.
>> Writing to this node requires root privilages, therefore this bug
>> has very little security impact.
>> 
>> Tracefs node: /sys/kernel/tracing/error_log
>> 
>
>Nice catch, but I recommend a different solution.
>
>
>> Example Kernel panic:
>> 
>> Unable to handle kernel NULL pointer dereference at virtual address 0000000000000038
>> Call trace:
>>  mutex_lock+0x30/0x110
>>  seq_lseek+0x34/0xb8
>>  __arm64_sys_lseek+0x6c/0xb8
>>  invoke_syscall+0x58/0x13c
>>  el0_svc_common+0xc4/0x10c
>>  do_el0_svc+0x24/0x98
>>  el0_svc+0x24/0x88
>>  el0t_64_sync_handler+0x84/0xe4
>>  el0t_64_sync+0x1b4/0x1b8
>> Code: d503201f aa0803e0 aa1f03e1 aa0103e9 (c8e97d02)
>> ---[ end trace 561d1b49c12cf8a5 ]---
>> Kernel panic - not syncing: Oops: Fatal exception
>> 
>> Signed-off-by: Mateusz Stachyra <m.stachyra@xxxxxxxxxxx>
>> ---
>>  kernel/trace/trace.c | 10 +++++++++-
>>  1 file changed, 9 insertions(+), 1 deletion(-)
>> 
>> diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
>> index 5d2c5678b..bfa8e2d01 100644
>> --- a/kernel/trace/trace.c
>> +++ b/kernel/trace/trace.c
>> @@ -8097,8 +8097,16 @@ static int tracing_err_log_open(struct inode *inode, struct file *file)
>>                  return ret;
>>  
>>          /* If this file was opened for write, then erase contents */
>> -        if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC))
>> +        if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC)) {
>>                  clear_tracing_err_log(tr);
>> +                ret = seq_open(file, &tracing_err_log_seq_ops);
>> +                if (!ret) {
>> +                        struct seq_file *m = file->private_data;
>> +                        m->private = tr;
>> +                } else {
>> +                        trace_array_put(tr);
>> +                }
>> +        }
>>  
>>          if (file->f_mode & FMODE_READ) {
>>                  ret = seq_open(file, &tracing_err_log_seq_ops);
>> 
>> base-commit: 1ef6663a587ba3e57dc5065a477db1c64481eedd
>
>I believe this can be better fixed by:
>
>diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
>index 64a4dde073ef..999b7c73e324 100644
>--- a/kernel/trace/trace.c
>+++ b/kernel/trace/trace.c
>@@ -8135,7 +8135,7 @@ static const struct file_operations tracing_err_log_fops = {
>         .open           = tracing_err_log_open,
>         .write                = tracing_err_log_write,
>         .read           = seq_read,
>-        .llseek         = seq_lseek,
>+        .llseek                = tracing_lseek,
>         .release        = tracing_err_log_release,
> };
>
>as that tracing_lseek() is for this exact scenario.
>
>Care to send a v2?
>
>Thanks,
>
>-- Steve

Thanks for reply and for the suggestion.
I've submitted v2 at:
https://lore.kernel.org/all/20230704102706eucms1p30d7ecdcc287f46ad67679fc8491b2e0f@eucms1p3/

-- Mateusz