Skip to contentSkip to conversation

fix kbuild warning in sisfb.o

PATCH In review Darren Jenkins\ 5 participants 6 messages Mar 31, 2006 – Apr 1, 2006

Subsystems (3)
Kbuild Kernel janitors Framebuffer layer
lore
Share & tools
lore

Darren Jenkins\

author
#0 lore raw

PATCH 1/1 fix kbuild warning in sisfb.o

G'day list

This one was a lot harder to find, but much easier to fix.

WARNING: drivers/video/sis/sisfb.o - Section mismatch: reference to .init.data: from .data between 'sisfb_driver' (at offset 0xf8) and 'sisfb_ops'

This is caused by the 'pci_driver struct' in sis_main.c having a pointer to a 'pci_device_id struct' in 'sis_main.h' that is marked as __devinitdata.

The patch below just removes the __devinitdata annotation from the 'pci_device_id struct', which seems like the best solution here.

Note: This driver also contains a lot of

#if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)

which is in the TODO list
- Code that depends on LINUX_VERSION_CODE & KERNEL_VERSION < 2.6 can be deleted in most cases. (Mostly/all? done)

So I might follow with a patch that removes it.

Signed-off-by Darren Jenkins <[email protected]>
drivers/video/sis/sis_main.h renamed from drivers/video/sis/sis_main.h.orig +1 -1
@@ -125,7 +125,7 @@ static struct sisfb_chip_info {
 	{ XGI_40,     SIS_315_VGA, 1, HW_CURSOR_AREA_SIZE_315 * 4, SIS_CRT2_WENABLE_315, "XGI V3XT/V5/V8" },
 };
 
-static struct pci_device_id __devinitdata sisfb_pci_table[] = {
+static struct pci_device_id sisfb_pci_table[] = {
 #ifdef CONFIG_FB_SIS_300
 	{ PCI_VENDOR_ID_SI, PCI_DEVICE_ID_SI_300,     PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0},
 	{ PCI_VENDOR_ID_SI, PCI_DEVICE_ID_SI_540_VGA, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 1},

_______________________________________________
Kernel-janitors mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/kernel-janitors

#1 lore raw
Quoted 9 lines · Darren Jenkins

G'day list

This one was a lot harder to find, but much easier to fix.

WARNING: drivers/video/sis/sisfb.o - Section mismatch: reference to .init.data: from .data between 'sisfb_driver' (at offset 0xf8) and 'sisfb_ops'

Quoted 6 lines

This is caused by the 'pci_driver struct' in sis_main.c having a pointer to a 'pci_device_id struct' in 'sis_main.h' that is marked as __devinitdata.

The patch below just removes the __devinitdata annotation from the 'pci_device_id struct', which seems like the best solution here.

ACK, this is a bug that should be fixed.

Quoted 9 lines

Note: This driver also contains a lot of

#if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)

which is in the TODO list
- Code that depends on LINUX_VERSION_CODE & KERNEL_VERSION < 2.6 can be deleted in most cases. (Mostly/all? done)

So I might follow with a patch that removes it.

AFAIR, Thomas (Cc'ed) does still share this code between kernel 2.4 and 2.6, and the #ifdef's should therefore stay.

Quoted 7 lines
Signed-off-by Darren Jenkins <[email protected]>

drivers/video/sis/sis_main.h static struct sisfb_chip_info { · line 131

Quoted 8 lines
drivers/video/sis/sis_main.h renamed from drivers/video/sis/sis_main.h.orig +1 -1
@@ -125,7 +125,7 @@ static struct sisfb_chip_info {
 	{ XGI_40,     SIS_315_VGA, 1, HW_CURSOR_AREA_SIZE_315 * 4, SIS_CRT2_WENABLE_315, "XGI V3XT/V5/V8" },
 };
 
-static struct pci_device_id __devinitdata sisfb_pci_table[] = {
+static struct pci_device_id sisfb_pci_table[] = {
 #ifdef CONFIG_FB_SIS_300
 	{ PCI_VENDOR_ID_SI, PCI_DEVICE_ID_SI_300,     PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0},
 	{ PCI_VENDOR_ID_SI, PCI_DEVICE_ID_SI_540_VGA, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 1},

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

_______________________________________________
Kernel-janitors mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/kernel-janitors

#2 lore raw
Quoted 8 lines · Adrian Bunk
Quoted 6 lines

This is caused by the 'pci_driver struct' in sis_main.c having a pointer to a 'pci_device_id struct' in 'sis_main.h' that is marked as __devinitdata.

The patch below just removes the __devinitdata annotation from the 'pci_device_id struct', which seems like the best solution here.

ACK, this is a bug that should be fixed.

If it is, then Documentation/pci.txt needs to be updated:

The struct pci_driver shouldn't be marked with any of these tags. The ID table array should be marked __devinitdata.

_______________________________________________
Kernel-janitors mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/kernel-janitors

#3 lore raw
In reply to Darren Jenkins\
Quoted 9 lines · Darren Jenkins

G'day list

This one was a lot harder to find, but much easier to fix.

WARNING: drivers/video/sis/sisfb.o - Section mismatch: reference to .init.data: from .data between 'sisfb_driver' (at offset 0xf8) and 'sisfb_ops'

Quoted 6 lines

This is caused by the 'pci_driver struct' in sis_main.c having a pointer to a 'pci_device_id struct' in 'sis_main.h' that is marked as __devinitdata.

The patch below just removes the __devinitdata annotation from the 'pci_device_id struct', which seems like the best solution here.

No, this should be just fine. devinitdata should be just fine for PCI device ids.

thanks,

greg k-h

_______________________________________________
Kernel-janitors mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/kernel-janitors

author
#4 lore raw
In reply to Darren Jenkins\

Sam

I don't know weather you are aware of this one yet, but we have just found an issue where 'struct pci_device_id' should be marked __devinitdata (see Documentation/pci.txt) as the pci code apparently only uses it in the driver initialisation, but the 'struct pci_driver' will still hold a pointer to it, giving a kbuild warning. See below. This definitely looks like a false positive that should be removed.

Darren Jenkins

Quoted 17 lines · Greg KH
Quoted 9 lines · Darren Jenkins

G'day list

This one was a lot harder to find, but much easier to fix.

WARNING: drivers/video/sis/sisfb.o - Section mismatch: reference to .init.data: from .data between 'sisfb_driver' (at offset 0xf8) and 'sisfb_ops'

Quoted 6 lines

This is caused by the 'pci_driver struct' in sis_main.c having a pointer to a 'pci_device_id struct' in 'sis_main.h' that is marked as __devinitdata.

The patch below just removes the __devinitdata annotation from the 'pci_device_id struct', which seems like the best solution here.

Quoted 6 lines

No, this should be just fine. devinitdata should be just fine for PCI device ids.

thanks,

greg k-h

_______________________________________________
Kernel-janitors mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/kernel-janitors

#5 lore raw
In reply to Darren Jenkins\
Quoted 8 lines · Darren Jenkins

Sam

I don't know weather you are aware of this one yet, but we have just found an issue where 'struct pci_device_id' should be marked __devinitdata (see Documentation/pci.txt) as the pci code apparently only uses it in the driver initialisation, but the 'struct pci_driver' will still hold a pointer to it, giving a kbuild warning. See below. This definitely looks like a false positive that should be removed.

Hi Darren.
Following patch should fix it.
It recognize that the variable is named *_driver, and in that case it will not warn about references to .init.data (+.init.text and .exit.text).

This is only compiletine tested since I'm a bit busy atm. Let me know if this solves it as expected and especially if it seems to silence warnings that is not false positives.

Sam

scripts/mod/modpost.c +3 -2
index 0b92ddf..81448e5 100644
@@ -492,7 +492,7 @@ static int strrcmp(const char *s, const 
  *   These functions may often be marked __init and we do not want to
  *   warn here.
  *   the pattern is identified by:
- *   tosec   = .init.text | .exit.text
+ *   tosec   = .init.text | .exit.text | .init.data
  *   fromsec = .data
  *   atsym = *_driver, *_ops, *_probe, *probe_one
  **/

@@ -522,7 +522,8 @@ static int secref_whitelist(const char *
 
 	/* Check for pattern 2 */
 	if ((strcmp(tosec, ".init.text") != 0) &&
-	    (strcmp(tosec, ".exit.text") != 0))
+	    (strcmp(tosec, ".exit.text") != 0) &&
+	    (strcmp(tosec, ".init.data") != 0))
 		f2 = 0;
 	if (strcmp(fromsec, ".data") != 0)
 		f2 = 0;

_______________________________________________
Kernel-janitors mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/kernel-janitors