]> git.ipfire.org Git - thirdparty/kernel/stable.git/commitdiff
scsi: ufs: Release clock if DMA map fails
authorCan Guo <cang@codeaurora.org>
Thu, 5 Dec 2019 02:14:33 +0000 (02:14 +0000)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 11 Jun 2020 07:22:21 +0000 (09:22 +0200)
commit 17c7d35f141ef6158076adf3338f115f64fcf760 upstream.

In queuecommand path, if DMA map fails, it bails out with clock held.  In
this case, release the clock to keep its usage paired.

[mkp: applied by hand]

Link: https://lore.kernel.org/r/0101016ed3d66395-1b7e7fce-b74d-42ca-a88a-4db78b795d3b-000000@us-west-2.amazonses.com
Reviewed-by: Bean Huo <beanhuo@micron.com>
Signed-off-by: Can Guo <cang@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
[EB: resolved cherry-pick conflict caused by newer kernels not having
 the clear_bit_unlock() line]
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/scsi/ufs/ufshcd.c

index 50d15748084a47d287a826e443447e6e86c52431..af4b0a2021d6c7f817a1847e0161278862b3b879 100644 (file)
@@ -1512,6 +1512,7 @@ static int ufshcd_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd)
 
        err = ufshcd_map_sg(hba, lrbp);
        if (err) {
+               ufshcd_release(hba);
                lrbp->cmd = NULL;
                clear_bit_unlock(tag, &hba->lrb_in_use);
                goto out;