[PATCH] octeontx2-af: Fix a resource leak in an error handling path in 'cgx_probe()'

If an error occurs after the call to 'pci_alloc_irq_vectors()', we must
call 'pci_free_irq_vectors()' in order to avoid a	resource leak.

The same sequence is already in place in the corresponding 'cgx_remove()'
function.

Fixes: 1463f382f58d ("octeontx2-af: Add support for CGX link management")
Signed-off-by: Christophe JAILLET <redacted>
---
 drivers/net/ethernet/marvell/octeontx2/af/cgx.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
index 742f0c1f60df..6d55e3d0b7ea 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c

@@ -825,7 +825,7 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (!cgx->cgx_cmd_workq) {
 		dev_err(dev, "alloc workqueue failed for cgx cmd");
 		err = -ENOMEM;
-		goto err_release_regions;
+		goto err_free_irq_vectors;
 	}
 
 	list_add(&cgx->cgx_list, &cgx_list);

@@ -841,6 +841,8 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 err_release_lmac:
 	cgx_lmac_exit(cgx);
 	list_del(&cgx->cgx_list);
+err_free_irq_vectors:
+	pci_free_irq_vectors(pdev);
 err_release_regions:
 	pci_release_regions(pdev);
 err_disable_device:

-- 
2.19.1

From: Christophe JAILLET <redacted>
Date: Sat, 29 Dec 2018 17:42:22 +0100

If an error occurs after the call to 'pci_alloc_irq_vectors()', we must
call 'pci_free_irq_vectors()' in order to avoid a	resource leak.

The same sequence is already in place in the corresponding 'cgx_remove()'
function.

Fixes: 1463f382f58d ("octeontx2-af: Add support for CGX link management")
Signed-off-by: Christophe JAILLET <redacted>

Applied and queued up for -stable.

Please fix the date and time on your computer, putting improper dates into
the email headers of your patch postings makes your patch appear out
of order in my patchwork queue.

Thank you.

Le 04/01/2019 à 21:46, David Miller a écrit :

From: Christophe JAILLET <redacted>
Date: Sat, 29 Dec 2018 17:42:22 +0100

quoted

If an error occurs after the call to 'pci_alloc_irq_vectors()', we must
call 'pci_free_irq_vectors()' in order to avoid a	resource leak.

The same sequence is already in place in the corresponding 'cgx_remove()'
function.

Fixes: 1463f382f58d ("octeontx2-af: Add support for CGX link management")
Signed-off-by: Christophe JAILLET <redacted>

Applied and queued up for -stable.

Please fix the date and time on your computer, putting improper dates into
the email headers of your patch postings makes your patch appear out
of order in my patchwork queue.

Thank you.

Hi,

yes, I've spotted this time issue.
For some reason my VM get out of synch when restored from a saved state.
It was not the case before. It is likely due to an update of VirtualBox (i.e. 5.x --> 6.0) performed a few days ago.
Will try to give it a look.

For now, I have to manually do some ntpdate to get the time/date fixed... until the next save/restore of the VM.

Sorry for the inconvenience.

CJ