Skip to contentSkip to conversation

PATCH: fix some 32bit isms

PATCH Alan Cox 5 participants 8 messages Jul 28, 2004 – Aug 6, 2004

Subsystems (2)
Memory technology devices (MTD) LSI Logic MPT fusion drivers (FC/SAS/SPI)
lore
Share & tools
lore

Alan Cox

author
#0 lore raw

Fairly self explanatory. int is not size_t.

Alan

OSDL Developer Certificate of Origin 1.0 included herein by reference

drivers/message/fusion/mptbase.c +1 -1
diff -u --new-file --recursive --exclude-from /usr/src/exclude linux.vanilla-2.6.8-rc2/drivers/message/fusion/mptbase.c linux-2.6.8-rc2/drivers/message/fusion/mptbase.c
@@ -2417,7 +2417,7 @@
 	} else {
 		printk(MYIOC_s_ERR_FMT 
 		     "Invalid IOC facts reply, msgLength=%d offsetof=%d!\n",
-		     ioc->name, facts->MsgLength, (offsetof(IOCFactsReply_t,
+		     ioc->name, facts->MsgLength, (int)(offsetof(IOCFactsReply_t,
 		     RequestFrameSize)/sizeof(u32)));
 		return -66;
 	}
drivers/mtd/inftlmount.c +8 -6
diff -u --new-file --recursive --exclude-from /usr/src/exclude linux.vanilla-2.6.8-rc2/drivers/mtd/inftlmount.c linux-2.6.8-rc2/drivers/mtd/inftlmount.c
@@ -58,7 +58,7 @@
 	u8 buf[SECTORSIZE];
 	struct INFTLMediaHeader *mh = &inftl->MediaHdr;
 	struct INFTLPartition *ip;
-	int retlen;
+	size_t retlen;
 
 	DEBUG(MTD_DEBUG_LEVEL3, "INFTL: find_boot_record(inftl=0x%x)\n",
 		(int)inftl);

@@ -288,7 +288,7 @@
 		inftl->PUtable = kmalloc(inftl->nb_blocks * sizeof(u16), GFP_KERNEL);
 		if (!inftl->PUtable) {
 			printk(KERN_WARNING "INFTL: allocation of PUtable "
-				"failed (%d bytes)\n",
+				"failed (%ld bytes)\n",
 				inftl->nb_blocks * sizeof(u16));
 			return -ENOMEM;
 		}

@@ -297,7 +297,7 @@
 		if (!inftl->VUtable) {
 			kfree(inftl->PUtable);
 			printk(KERN_WARNING "INFTL: allocation of VUtable "
-				"failed (%d bytes)\n",
+				"failed (%ld bytes)\n",
 				inftl->nb_blocks * sizeof(u16));
 			return -ENOMEM;
 		}

@@ -348,7 +348,8 @@
 static int check_free_sectors(struct INFTLrecord *inftl, unsigned int address,
 	int len, int check_oob)
 {
-	int i, retlen;
+	int i;
+	size_t retlen;
 	u8 buf[SECTORSIZE + inftl->mbd.mtd->oobsize];
 
 	DEBUG(MTD_DEBUG_LEVEL3, "INFTL: check_free_sectors(inftl=0x%x,"

@@ -382,7 +383,7 @@
  */
 int INFTL_formatblock(struct INFTLrecord *inftl, int block)
 {
-	int retlen;
+	size_t retlen;
 	struct inftl_unittail uci;
 	struct erase_info *instr = &inftl->instr;
 	int physblock;

@@ -551,7 +552,8 @@
 	int chain_length, do_format_chain;
 	struct inftl_unithead1 h0;
 	struct inftl_unittail h1;
-	int i, retlen;
+	int i;
+	size_t retlen;
 	u8 *ANACtable, ANAC;
 
 	DEBUG(MTD_DEBUG_LEVEL3, "INFTL: INFTL_mount(inftl=0x%x)\n", (int)s);
#1 lore raw
Followup to:  <[email protected]>
By author:    Alan Cox <[email protected]>
In newsgroup: linux.dev.kernel
Quoted 13 lines

Fairly self explanatory. int is not size_t.

Alan

OSDL Developer Certificate of Origin 1.0 included herein by reference

diff -u --new-file --recursive --exclude-from /usr/src/exclude linux.vanilla-2.6.8-rc2/drivers/message/fusion/mptbase.c linux-2.6.8-rc2/drivers/message/fusion/mptbase.c

drivers/message/fusion/mptbase.c line 2423

Quoted 8 lines
drivers/message/fusion/mptbase.c +1 -1
@@ -2417,7 +2417,7 @@
 	} else {
 		printk(MYIOC_s_ERR_FMT 
 		     "Invalid IOC facts reply, msgLength=%d offsetof=%d!\n",
-		     ioc->name, facts->MsgLength, (offsetof(IOCFactsReply_t,
+		     ioc->name, facts->MsgLength, (int)(offsetof(IOCFactsReply_t,
 		     RequestFrameSize)/sizeof(u32)));
 		return -66;
 	}

This is probably better fixed by changing %d to %zu.

-hpa

#2 lore raw
In reply to Alan Cox
Quoted 1 line · On Wed, 28 Jul 2004 09:59:41 -0400 Alan Cox <[email protected]> wrote:

Fairly self explanatory. int is not size_t.

And size_t is not "%ld" either.

Review

Quoted 2 lines
(quoted diff) +1 -1
-				"failed (%d bytes)\n",
+				"failed (%ld bytes)\n",

Rather, it is either "%Zd" or "%zd". Please fix :-)

#3 lore raw
In reply to Alan Cox
Quoted 5 lines · Alan Cox

printk(MYIOC_s_ERR_FMT
"Invalid IOC facts reply, msgLength=%d offsetof=%d!\n",
- ioc->name, facts->MsgLength, (offsetof(IOCFactsReply_t,
+ ioc->name, facts->MsgLength, (int)(offsetof(IOCFactsReply_t,

printk expects %zd for a size_t

Quoted 6 lines

RequestFrameSize)/sizeof(u32)));
...
printk(KERN_WARNING "INFTL: allocation of PUtable "
- "failed (%d bytes)\n",
+ "failed (%ld bytes)\n",
inftl->nb_blocks * sizeof(u16));

Some architectures will emit a warning here, and will perhaps print the wrong thing. We need to print size_t's with %zd. I'll fix that up.

author
#4 lore raw
Quoted 8 lines · Andrew Morton
Quoted 5 lines · Alan Cox

printk(MYIOC_s_ERR_FMT
"Invalid IOC facts reply, msgLength=%d offsetof=%d!\n",
- ioc->name, facts->MsgLength, (offsetof(IOCFactsReply_t,
+ ioc->name, facts->MsgLength, (int)(offsetof(IOCFactsReply_t,

printk expects %zd for a size_t

So I've now learned. I've been a bit out of touch with the 2.5/2.6 printk evolution

Quoted 2 lines

Some architectures will emit a warning here, and will perhaps print the wrong thing. We need to print size_t's with %zd. I'll fix that up.

Thanks

#5 lore raw
In reply to Andrew Morton
Followup to:  <[email protected]>
By author:    Andrew Morton <[email protected]>
In newsgroup: linux.dev.kernel
Quoted 10 lines
Quoted 5 lines · Alan Cox

printk(MYIOC_s_ERR_FMT
"Invalid IOC facts reply, msgLength=%d offsetof=%d!\n",
- ioc->name, facts->MsgLength, (offsetof(IOCFactsReply_t,
+ ioc->name, facts->MsgLength, (int)(offsetof(IOCFactsReply_t,

printk expects %zd for a size_t

It should be %zu, since size_t is unsigned.

%zd is appropriate for ssize_t.

-hpa

#6 lore raw
Quoted 11 lines · H. Peter Anvin
Followup to:  <[email protected]>
By author:    Andrew Morton <[email protected]>
In newsgroup: linux.dev.kernel
Quoted 8 lines
Quoted 5 lines · Alan Cox

printk(MYIOC_s_ERR_FMT
"Invalid IOC facts reply, msgLength=%d offsetof=%d!\n",
- ioc->name, facts->MsgLength, (offsetof(IOCFactsReply_t,
+ ioc->name, facts->MsgLength, (int)(offsetof(IOCFactsReply_t,

Quoted 6 lines
Quoted 2 lines

printk expects %zd for a size_t

It should be %zu, since size_t is unsigned.

%zd is appropriate for ssize_t.

It's signed (and not a ssize_t, while we are at it). If you want to be pedantic, use %td since we are dealing with ptrdiff_t here.

BTW, we ought to add 't' modifier in vsnpritf()...

#7 lore raw
Quoted 6 lines · On Sat, Aug 07, 2004 at 12:03:05AM +0100…
Quoted 3 lines

It should be %zu, since size_t is unsigned.

%zd is appropriate for ssize_t.

It's signed (and not a ssize_t, while we are at it). If you want to be pedantic, use %td since we are dealing with ptrdiff_t here.

*Ugh*

We are not doing homegrown offsetof(), so it's really size_t. My apologies.

Yes, that should be> %zu (OTOH, %zd will do the same unless we have 2Gb struct somewhere ;-)

Quoted 1 line

BTW, we ought to add 't' modifier in vsnpritf()...