From: Keita Suzuki <keitasuzuki.p...@sslab.ics.keio.ac.jp>

[ Upstream commit bc28369c6189009b66d9619dd9f09bd8c684bb98 ]

When mfd_add_devices() fail, pcr->slots should also be freed. However,
the current implementation does not free the member, leading to a memory
leak.

Fix this by adding a new goto label that frees pcr->slots.

Signed-off-by: Keita Suzuki <keitasuzuki.p...@sslab.ics.keio.ac.jp>
Link: 
https://lore.kernel.org/r/20200909071853.4053-1-keitasuzuki.p...@sslab.ics.keio.ac.jp
Signed-off-by: Greg Kroah-Hartman <gre...@linuxfoundation.org>
Signed-off-by: Sasha Levin <sas...@kernel.org>
---
 drivers/misc/cardreader/rtsx_pcr.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/misc/cardreader/rtsx_pcr.c 
b/drivers/misc/cardreader/rtsx_pcr.c
index 37ccc67f4914b..f2b2805942f50 100644
--- a/drivers/misc/cardreader/rtsx_pcr.c
+++ b/drivers/misc/cardreader/rtsx_pcr.c
@@ -1562,12 +1562,14 @@ static int rtsx_pci_probe(struct pci_dev *pcidev,
        ret = mfd_add_devices(&pcidev->dev, pcr->id, rtsx_pcr_cells,
                        ARRAY_SIZE(rtsx_pcr_cells), NULL, 0, NULL);
        if (ret < 0)
-               goto disable_irq;
+               goto free_slots;
 
        schedule_delayed_work(&pcr->idle_work, msecs_to_jiffies(200));
 
        return 0;
 
+free_slots:
+       kfree(pcr->slots);
 disable_irq:
        free_irq(pcr->irq, (void *)pcr);
 disable_msi:
-- 
2.25.1

Reply via email to