Re: Remove lots of IS_ERR_VALUE abuses and compilation warning.

From: arvind Yadav
Date: Thu Jul 07 2016 - 13:06:47 EST


Yes, You are right,
-Now Return type of 'qe_muram_alloc' is 'unsigned long', That Was trying to assigned in ucc_fast_tx_virtual_fifo_base_offset and ucc_fast_rx_virtual_fifo_base_offset. These variable are 'unsigned int'. So before assignment need a proper type casting.

-Passing value in IS_ERR_VALUE() is wrong. So this is also need a proper type casting before passing an argument.

I have done the changes and re-submitted anther patch, Please review It.

Thanks,
Arvind

On Thursday 07 July 2016 09:21 PM, Guenter Roeck wrote:
On Thu, Jul 07, 2016 at 08:57:29PM +0530, Arvind Yadav wrote:
Passing value in IS_ERR_VALUE() is wrong, as they
pass an 'int' into a function that takes an 'unsigned long'
argument. This happens to work because the type is sign-extended
on 64-bit architectures before it gets converted into an
unsigned type.

Passing an 'unsigned short' or 'unsigned int'
argument into IS_ERR_VALUE() is guaranteed to be broken, as are
8-bit integers and types that are wider than 'unsigned long'.

Any user will get compilation warning for that do not pass an
'unsigned long' argument.

Commit '287980e49f; - This change is alreday fixes lots of other
worst abusers.

Couple of generic comments:

- Your patch subject lines don't include the affected drivers/modules.
As such, most of them will be ignored because maintainers won't realize
that you are talking with them. Some may ask you to resubmit with proper
subject lines.
Commit 287980e49f is different; it addresses the problem in several
drivers in a single commit.
- If you patch a single file, I think it would be better to adjust the
description accordingly. In this patch, the offending variable type is
u32. The patch description is therefore misleading; the code here simply does
not work.
- When you resubmit a patch, you don't include a version, not a change log.
This means additional work for maintainers, who have to figure out which
patch to apply.

Specific comment:

The allocator in question returns -ENOMEM in an unsigned long. This is assigned
to u32. A proper fix would be to assign the return value to an unsigned
long and to use IS_ER_VALUE() to check if it reports an error, and to only
assign it to ucc_fast_rx_virtual_fifo_base_offset if there was no error.

Also, unless I am missing something - since ucc_fast_rx_virtual_fifo_base_offset
is defined as u32, it is somewhat unlikely that it is ever < 0.

Guenter

Signed-off-by: Arvind Yadav <arvind.yadav.cs@xxxxxxxxx>
---
drivers/soc/fsl/qe/ucc_fast.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/soc/fsl/qe/ucc_fast.c b/drivers/soc/fsl/qe/ucc_fast.c
index a768931..7cc783c 100644
--- a/drivers/soc/fsl/qe/ucc_fast.c
+++ b/drivers/soc/fsl/qe/ucc_fast.c
@@ -268,7 +268,7 @@ int ucc_fast_init(struct ucc_fast_info * uf_info, struct ucc_fast_private ** ucc
/* Allocate memory for Tx Virtual Fifo */
uccf->ucc_fast_tx_virtual_fifo_base_offset =
qe_muram_alloc(uf_info->utfs, UCC_FAST_VIRT_FIFO_REGS_ALIGNMENT);
- if (IS_ERR_VALUE(uccf->ucc_fast_tx_virtual_fifo_base_offset)) {
+ if (uccf->ucc_fast_tx_virtual_fifo_base_offset < 0) {
printk(KERN_ERR "%s: cannot allocate MURAM for TX FIFO\n",
__func__);
uccf->ucc_fast_tx_virtual_fifo_base_offset = 0;
@@ -281,7 +281,7 @@ int ucc_fast_init(struct ucc_fast_info * uf_info, struct ucc_fast_private ** ucc
qe_muram_alloc(uf_info->urfs +
UCC_FAST_RECEIVE_VIRTUAL_FIFO_SIZE_FUDGE_FACTOR,
UCC_FAST_VIRT_FIFO_REGS_ALIGNMENT);
- if (IS_ERR_VALUE(uccf->ucc_fast_rx_virtual_fifo_base_offset)) {
+ if (uccf->ucc_fast_rx_virtual_fifo_base_offset < 0) {
printk(KERN_ERR "%s: cannot allocate MURAM for RX FIFO\n",
__func__);
uccf->ucc_fast_rx_virtual_fifo_base_offset = 0;