[Devel] Re: [RFC v8][PATCH 05/12] x86 support for checkpoint/restart

[Oren Laadan]

Masahiko Takahashi wrote:
> Hi Oren,
> 
> I'm now trying to port your patchset to x86_64, and find a tiny
> inconsistency issue.
> 
> 
> On 2008-10-30 at 09:51 -0400, Oren Laadan wrote:
>> +/* dump the thread_struct of a given task */
>> +int cr_write_thread(struct cr_ctx *ctx, struct task_struct *t)
>> +{
>> +	struct cr_hdr h;
>> +	struct cr_hdr_thread *hh = cr_hbuf_get(ctx, sizeof(*hh));
>> +	struct thread_struct *thread;
>> +	struct desc_struct *desc;
>> +	int ntls = 0;
>> +	int n, ret;
>> +
>> +	h.type = CR_HDR_THREAD;
>> +	h.len = sizeof(*hh);
>> +	h.parent = task_pid_vnr(t);
>> +
>> +	thread = &t->thread;
>> +
>> +	/* calculate no. of TLS entries that follow */
>> +	desc = thread->tls_array;
>> +	for (n = GDT_ENTRY_TLS_ENTRIES; n > 0; n--, desc++) {
>> +		if (desc->a || desc->b)
>> +			ntls++;
>> +	}
>> +
>> +	hh->gdt_entry_tls_entries = GDT_ENTRY_TLS_ENTRIES;
>> +	hh->sizeof_tls_array = sizeof(thread->tls_array);
>> +	hh->ntls = ntls;
>> +
>> +	ret = cr_write_obj(ctx, &h, hh);
>> +	cr_hbuf_put(ctx, sizeof(*hh));
>> +	if (ret < 0)
>> +		return ret;
> 
> Please add
>    if (ntls == 0)
>             return ret;
> because, in restart phase, reading TLS entries from the image file
> is skipped if hh->ntls == 0, which may incur inconsistency and fail
> to restart.
> 

Will fix, thanks.

Oren.


_______________________________________________
Containers mailing list
Containers at lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers