On Tue, 20 Feb 2018, Arnd Bergmann wrote:
> gcc-6 and higher warn about the way some loops are written in
> the ncr5380 driver:
>
> drivers/scsi/g_NCR5380.c: In function 'generic_NCR5380_pread':
> drivers/scsi/g_NCR5380.c:541:3: error: this 'while' clause does not guard... [-Werror=misleading-indentation]
> while (NCR5380_read(C400_CONTROL_STATUS_REG) & CSR_HOST_BUF_NOT_RDY);
> ^~~~~
> drivers/scsi/g_NCR5380.c:544:3: note: ...this statement, but the latter is misleadingly indented as if it is guarded by the 'while'
>
> This was addressed in mainline kernels as part of a rework on commit
> 12150797d064 ("ncr5380: Use runtime register mapping").
Acked-by: Finn Thain <fthain@xxxxxxxxxxxxxxxxxxx>
> We don't want the entire patch backported to stable kernels, but we can
> backport one hunk to get rid of the warning.
Or "backport something equivalent to one hunk".
>
> Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
> ---
> drivers/scsi/g_NCR5380.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/scsi/g_NCR5380.c b/drivers/scsi/g_NCR5380.c
> index f8d2478b11cc..87e081f8a386 100644
> --- a/drivers/scsi/g_NCR5380.c
> +++ b/drivers/scsi/g_NCR5380.c
> @@ -538,7 +538,10 @@ static inline int NCR5380_pread(struct Scsi_Host *instance, unsigned char *dst,
> printk(KERN_ERR "53C400r: Got 53C80_IRQ start=%d, blocks=%d\n", start, blocks);
> return -1;
> }
> - while (NCR5380_read(C400_CONTROL_STATUS_REG) & CSR_HOST_BUF_NOT_RDY);
> + while (NCR5380_read(C400_CONTROL_STATUS_REG) & CSR_HOST_BUF_NOT_RDY)
> + {
> + // FIXME - no timeout
> + }
>
> #ifndef SCSI_G_NCR5380_MEM
> {
>
