[PATCH 2.6.2-rc1-mm2] fs/xfs/xfs_log_recover.c

[PATCH 2.6.2-rc1-mm2] fs/xfs/xfs_log_recover.c

[Bryan Whitehead]

This fixes a warning on compile of the xfs fs module.

--- fs/xfs/xfs_log_recover.c.orig       2004-01-23 23:17:35.402907768 -0800
+++ fs/xfs/xfs_log_recover.c    2004-01-23 23:19:09.368622808 -0800
@@ -1542,21 +1542,14 @@
                case XFS_LI_BUF:
                        flags = buf_f->blf_flags;
                        break;
                case XFS_LI_6_1_BUF:
                case XFS_LI_5_3_BUF:
                        obuf_f = (xfs_buf_log_format_v1_t*)buf_f;
                        flags = obuf_f->blf_flags;
-                       break;
-               }
-
-               switch (ITEM_TYPE(itemq)) {
-               case XFS_LI_BUF:
-               case XFS_LI_6_1_BUF:
-               case XFS_LI_5_3_BUF:
                        if (!(flags & XFS_BLI_CANCEL)) {
                                xlog_recover_insert_item_frontq(&trans->r_itemq,
                                                                itemq);
                                break;
                        }
                case XFS_LI_INODE:
                case XFS_LI_6_1_INODE:


--
Bryan Whitehead
driver@megahappy.net

Re: [PATCH 2.6.2-rc1-mm2] fs/xfs/xfs_log_recover.c

[Christoph Hellwig]

On Fri, Jan 23, 2004 at 11:31:11PM -0800, Bryan Whitehead wrote:
> 
> This fixes a warning on compile of the xfs fs module.

This patch looks very strange.  What error do you get without it?

Re: [PATCH 2.6.2-rc1-mm2] fs/xfs/xfs_log_recover.c

[Bryan Whitehead]

I should have been more clear in the frist email. Sorry.

The variable "flags" is used in an if statement without having a value 
assigned.

There is 2 switch statements that do this: (ITEM_TYPE(itemq) They execute 
one right after the other with no returns. So it is redundant.

I think the patch fixes a cut/paste accident...

Here is a more complete diff so you can see what is going on:
--- fs/xfs/xfs_log_recover.c.orig       2004-01-23 23:17:35.402907768 
-0800
+++ fs/xfs/xfs_log_recover.c    2004-01-23 23:19:09.368622808 -0800
@@ -1539,27 +1539,20 @@
                itemq_next = itemq->ri_next;
                buf_f = (xfs_buf_log_format_t *)itemq->ri_buf[0].i_addr;
                switch (ITEM_TYPE(itemq)) {
                case XFS_LI_BUF:
                        flags = buf_f->blf_flags;
                        break;
                case XFS_LI_6_1_BUF:
                case XFS_LI_5_3_BUF:
                        obuf_f = (xfs_buf_log_format_v1_t*)buf_f;
                        flags = obuf_f->blf_flags;
-                       break;
-               }
-
-               switch (ITEM_TYPE(itemq)) {
-               case XFS_LI_BUF:
-               case XFS_LI_6_1_BUF:
-               case XFS_LI_5_3_BUF:
                        if (!(flags & XFS_BLI_CANCEL)) {
                                
xlog_recover_insert_item_frontq(&trans->r_itemq,
                                                                itemq);
                                break;
                        }
                case XFS_LI_INODE:
                case XFS_LI_6_1_INODE:
                case XFS_LI_5_3_INODE:
                case XFS_LI_DQUOT:
                case XFS_LI_QUOTAOFF:



On Sat, 24 Jan 2004, Christoph Hellwig wrote:

> On Fri, Jan 23, 2004 at 11:31:11PM -0800, Bryan Whitehead wrote:
> > 
> > This fixes a warning on compile of the xfs fs module.
> 
> This patch looks very strange.  What error do you get without it?
> 

-- 
Bryan Whitehead
Email:driver@megahappy.net
WorkE:driver@jpl.nasa.gov

Re: [PATCH 2.6.2-rc1-mm2] fs/xfs/xfs_log_recover.c

[Bryan Whitehead]

Oh, one more thing. After removed the superflurous switch statement flags 
will have a value which fixes the warning.

If you look at the flow of code it is impossible for flags to not have a 
value when it gets to the if statement as the previous switch statement 
would give it a value. But since that switch statement only catches that 
one case anyway, it's better to merge the 2 switch statemnets.

On Sat, 24 Jan 2004, Christoph Hellwig wrote:

> On Fri, Jan 23, 2004 at 11:31:11PM -0800, Bryan Whitehead wrote:
> > 
> > This fixes a warning on compile of the xfs fs module.
> 
> This patch looks very strange.  What error do you get without it?
> 

-- 
Bryan Whitehead
Email:driver@megahappy.net
WorkE:driver@jpl.nasa.gov

Re: [PATCH 2.6.2-rc1-mm2] fs/xfs/xfs_log_recover.c

[Bryan Whitehead]

If I'm off base on this then sorry for the trouble. Just thought I'd 
take a hit at getting rid of some compiler warnings..

Bryan Whitehead wrote:
> Oh, one more thing. After removed the superflurous switch statement flags 
> will have a value which fixes the warning.
> 
> If you look at the flow of code it is impossible for flags to not have a 
> value when it gets to the if statement as the previous switch statement 
> would give it a value. But since that switch statement only catches that 
> one case anyway, it's better to merge the 2 switch statemnets.
> 
> On Sat, 24 Jan 2004, Christoph Hellwig wrote:
> 
> 
>>On Fri, Jan 23, 2004 at 11:31:11PM -0800, Bryan Whitehead wrote:
>>
>>>This fixes a warning on compile of the xfs fs module.
>>
>>This patch looks very strange.  What error do you get without it?
>>
> 
> 


-- 
Bryan Whitehead
Email:driver@megahappy.net
WorkE:driver@jpl.nasa.gov