linux-kernel - Re: [BUG] ALSA: echoaudio: warning when the driver fails to probe

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <s5ha6csc7lq.wl-tiwai@suse.de>
Date:   Mon, 11 Apr 2022 11:06:09 +0200
From:   Takashi Iwai <tiwai@...e.de>
To:     Zheyu Ma <zheyuma97@...il.com>
Cc:     perex@...ex.cz, tiwai@...e.com, alsa-devel@...a-project.org,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [BUG] ALSA: echoaudio: warning when the driver fails to probe

On Mon, 11 Apr 2022 10:49:53 +0200,
Zheyu Ma wrote:
> 
> On Mon, Apr 11, 2022 at 3:41 PM Takashi Iwai <tiwai@...e.de> wrote:
> >
> > On Sun, 10 Apr 2022 11:13:55 +0200,
> > Zheyu Ma wrote:
> > >
> > > Hello,
> > >
> > > I found a bug in echoaudio.c.
> > > When the driver fails at the function snd_echo_create(), it should
> > > release resources requested before, otherwise we will get the
> > > following warning:
> > >
> > > [    3.262866] remove_proc_entry: removing non-empty directory
> > > 'irq/21', leaking at least 'snd_indigodj'
> > > [    3.263577] WARNING: CPU: 3 PID: 261 at fs/proc/generic.c:717
> > > remove_proc_entry+0x389/0x3f0
> > > [    3.267098] RIP: 0010:remove_proc_entry+0x389/0x3f0
> > > [    3.269976] Call Trace:
> > > [    3.269979]  <TASK>
> > > [    3.269988]  unregister_irq_proc+0x14c/0x170
> > > [    3.269997]  irq_free_descs+0x94/0xe0
> > > [    3.270004]  mp_unmap_irq+0xb6/0x100
> > > [    3.270011]  acpi_unregister_gsi_ioapic+0x27/0x40
> > > [    3.270017]  acpi_pci_irq_disable+0x1d3/0x320
> > > [    3.270025]  pci_disable_device+0x1ad/0x380
> > > [    3.270034]  pcim_release+0x566/0x6d0
> > > [    3.270046]  devres_release_all+0x1f1/0x2c0
> > > [    3.270057]  really_probe+0xe0/0x920
> >
> > Could you try the patch below?
> 
> The following patch works for me, the previous warning disappears, thank you.
> But I got another error, I have no idea about it.

OK, that's bad, it's basically the destructor order problem.
Could you try the patch below instead of the previous one?


thanks,

Takashi

---
--- a/sound/pci/echoaudio/echoaudio.c
+++ b/sound/pci/echoaudio/echoaudio.c
@@ -1889,13 +1889,20 @@ static void snd_echo_free(struct snd_card *card)
 	if (chip->comm_page)
 		rest_in_peace(chip);
 
-	if (chip->irq >= 0)
-		free_irq(chip->irq, chip);
-
 	/* release chip data */
 	free_firmware_cache(chip);
 }
 
+static void do_free_irq(void *p)
+{
+	struct echoaudio *chip = p;
+
+	if (chip->irq >= 0) {
+		free_irq(chip->irq, chip);
+		chip->irq = -1;
+	}
+}
+
 /* <--snd_echo_probe() */
 static int snd_echo_create(struct snd_card *card,
 			   struct pci_dev *pci)
@@ -1945,8 +1952,9 @@ static int snd_echo_create(struct snd_card *card,
 	card->sync_irq = chip->irq;
 	dev_dbg(card->dev, "pci=%p irq=%d subdev=%04x Init hardware...\n",
 		chip->pci, chip->irq, chip->pci->subsystem_device);
-
-	card->private_free = snd_echo_free;
+	err = devm_add_action(card->dev, do_free_irq, chip);
+	if (err < 0)
+		return err;
 
 	/* Create the DSP comm page - this is the area of memory used for most
 	of the communication with the DSP, which accesses it via bus mastering */
@@ -1957,6 +1965,7 @@ static int snd_echo_create(struct snd_card *card,
 		return -ENOMEM;
 	chip->comm_page_phys = chip->commpage_dma_buf->addr;
 	chip->comm_page = (struct comm_page *)chip->commpage_dma_buf->area;
+	card->private_free = snd_echo_free;
 
 	err = init_hw(chip, chip->pci->device, chip->pci->subsystem_device);
 	if (err >= 0)