Doing kbuild 2.5 full build for SMP on 2.4.13 got lots of warnings and
a couple of errors. The .config had everything built in except for:
# I don't have the required input files to build the sound firmware.
# If anybody has them and can test, go for it.
CONFIG_PSS_HAVE_BOOT=n
CONFIG_MAUI_HAVE_BOOT=n
CONFIG_TRIX_HAVE_BOOT=n
# _SOUND_MSNDCLAS=y and _SOUND_MSNDPIN=y insist on having firmware that
# I don't have.
CONFIG_SOUND_MSNDCLAS=m
CONFIG_SOUND_MSNDPIN=m
# Broken code, missing ioctls
CONFIG_SCSI_CPQFCTS=n
# Broken code, missing #include <module.h>
CONFIG_SCSI_U14_34F=n
# Broken code, SCSI_ABORT_PENDING undefined
CONFIG_USB_HPUSBSCSI=n
The last three options give compile errors when they are turned on. I
have not checked if they are fixed in 2.4.14-pre3.
The link of vmlinux had a lot of duplicate objects:
drivers/net/ppp_deflate.o(.data+0x0): multiple definition of `deflate_copyright'
fs/jffs2/jffs2.o(.data+0x20): first defined here
drivers/net/ppp_deflate.o: In function `deflateInit_':
drivers/net/ppp_deflate.o(.text+0x0): multiple definition of `deflateInit_'
fs/jffs2/jffs2.o(.text+0xc90): first defined here
drivers/net/ppp_deflate.o: In function `deflateInit2_':
... lots more zlib conflicts ...
drivers/isdn/tpam/tpam.o: In function `hdlc_decode':
drivers/isdn/tpam/tpam.o(.text+0x3f20): multiple definition of `hdlc_decode'
drivers/isdn/hisax/hisax_st5481.o(.text+0x30b0): first defined here
ld: Warning: size of symbol `hdlc_decode' changed from 832 to 603 in drivers/isdn/tpam/tpam.o
drivers/isdn/tpam/tpam.o: In function `hdlc_encode':
drivers/isdn/tpam/tpam.o(.text+0x3d00): multiple definition of `hdlc_encode'
drivers/isdn/hisax/hisax_st5481.o(.text+0x33f0): first defined here
ld: Warning: size of symbol `hdlc_encode' changed from 1345 to 529 in drivers/isdn/tpam/tpam.o
drivers/mtd/chips/jedec_probe.o: In function `jedec_probe_init':
drivers/mtd/chips/jedec_probe.o(.text.init+0x0): multiple definition of `jedec_probe_init'
drivers/mtd/chips/jedec.o(.text.init+0x0): first defined here
The warnings, using gcc version 2.96 20000731 (Red Hat Linux 7.1 2.96-81):
arch/i386/kernel/dmi_scan.c:194: warning: `disable_ide_dma' defined but not used
drivers/atm/ambassador.c:301:10: warning: pasting "." and "start" does not give a valid preprocessing token
drivers/atm/ambassador.c:305:10: warning: pasting "." and "regions" does not give a valid preprocessing token
drivers/atm/ambassador.c:310:10: warning: pasting "." and "data" does not give a valid preprocessing token
drivers/atm/atmtcp.c:176: warning: `out_vcc' might be used uninitialized in this function
drivers/atm/atmtcp.c: In function `atmtcp_v_send':
drivers/char/applicom.c:257:2: warning: #warning "LEAK"
drivers/char/applicom.c:521:2: warning: #warning "Je suis stupide. DW. - copy*user in cli"
drivers/char/ip2/i2ellis.c:107: warning: `iiEllisCleanup' defined but not used
drivers/char/random.c:1248: warning: comparison of distinct pointer types lacks a cast
drivers/char/random.c: In function `xfer_secondary_pool':
drivers/i2c/i2c-core.c:1333: warning: implicit declaration of function `i2c_bitlp_init'
drivers/i2c/i2c-core.c: In function `i2c_init_all':
drivers/ide/pdc202xx.c:529: warning: `drive_pci' might be used uninitialized in this function
drivers/ide/pdc202xx.c: In function `config_chipset_for_dma':
drivers/ide/pdcraid.c:284: warning: label `outerr' defined but not used
drivers/ide/pdcraid.c:550: warning: unused variable `q'
drivers/ide/pdcraid.c:588: warning: unused variable `i'
drivers/ide/pdcraid.c:99: warning: unused variable `larg'
drivers/ide/pdcraid.c: In function `pdcraid0_make_request':
drivers/ide/pdcraid.c: In function `pdcraid_init':
drivers/ide/pdcraid.c: In function `pdcraid_init_one':
drivers/ide/pdcraid.c: In function `pdcraid_ioctl':
drivers/ieee1394/pcilynx.c:775: warning: `aux_ops' defined but not used
drivers/isdn/avmb1/b1pci.c:33: warning: `b1pci_pci_tbl' defined but not used
drivers/isdn/avmb1/c4.c:41: warning: `c4_pci_tbl' defined but not used
drivers/isdn/avmb1/t1pci.c:36: warning: `t1pci_pci_tbl' defined but not used
drivers/isdn/hisax/config.c:2079: warning: `hisax_pci_tbl' defined but not used
drivers/media/radio/radio-cadet.c:638: warning: `id_table' defined but not used
drivers/media/video/cpia.c:1305: warning: `proc_cpia_destroy' defined but not used
drivers/media/video/cpia.c:3290: warning: its scope is only this definition or declaration, which is probably not what you want.
drivers/media/video/cpia.c:3290: warning: `struct video_init' declared inside parameter list
drivers/media/video/msp3400.c:1233: warning: `rev2' might be used uninitialized in this function
drivers/media/video/msp3400.c: In function `msp_attach':
drivers/media/video/w9966.c:634: warning: `w9966_rReg_i2c' defined but not used
drivers/mtd/devices/doc1000.c:87:2: warning: #warning This is definitely not SMP safe. Lock the paging mechanism.
drivers/net/acenic.c:124: warning: `acenic_pci_tbl' defined but not used
drivers/net/acenic.c:1365: warning: right shift count >= width of type
drivers/net/acenic.c: In function `ace_init':
drivers/net/aironet4500_card.c:63: warning: `aironet4500_card_pci_tbl' defined but not used
drivers/net/arlan.c:1124: warning: `arlan_find_devices' defined but not used
drivers/net/arlan.c:22: warning: `probe' defined but not used
drivers/net/dgrs.c:124: warning: `dgrs_pci_tbl' defined but not used
drivers/net/fc/iph5526.c:114: warning: `iph5526_pci_tbl' defined but not used
drivers/net/fc/iph5526.c:227: warning: `driver_template' defined but not used
drivers/net/hamradio/6pack.c:703: warning: `msg_invparm' defined but not used
drivers/net/hp100.c:289: warning: `hp100_pci_tbl' defined but not used
drivers/net/irda/ali-ircc.c:467: warning: `ali_ircc_probe_43' defined but not used
drivers/net/irda/w83977af_ir.c:276: warning: `w83977af_close' defined but not used
drivers/net/pcmcia/xirc2ps_cs.c:2086:18: warning: extra tokens at end of #undef directive
drivers/net/rrunner.c:121: warning: `rrunner_pci_tbl' defined but not used
drivers/net/rrunner.c:1241: warning: comparison of distinct pointer types lacks a cast
drivers/net/rrunner.c:1252: warning: comparison of distinct pointer types lacks a cast
drivers/net/rrunner.c: In function `rr_dump':
drivers/net/sb1000.c:140: warning: `id_table' defined but not used
drivers/net/sk98lin/skvpd.c:245: warning: `VpdWriteDWord' defined but not used
drivers/net/skfp/skfddi.c:185: warning: `skfddi_pci_tbl' defined but not used
drivers/net/tokenring/olympic.c:1725:22: warning: no newline at end of file
drivers/net/tokenring/tmsisa.c:44: warning: `portlist' defined but not used
drivers/net/tulip/tulip_core.c:1326: warning: label `early_out' defined but not used
drivers/net/tulip/tulip_core.c: In function `tulip_mwi_config':
drivers/net/via-rhine.c:1590: warning: unused variable `np'
drivers/net/via-rhine.c: In function `via_rhine_remove_one':
drivers/net/wan/comx-hw-mixcom.c:570: warning: label `err_restore_flags' defined but not used
drivers/net/wan/comx-hw-mixcom.c: In function `MIXCOM_open':
drivers/net/wan/sdladrv.c:216: warning: `sdladrv_pci_tbl' defined but not used
drivers/net/winbond-840.c:146: warning: `version' defined but not used
drivers/parport/parport_cs.c:109: warning: `parport_cs_ops' defined but not used
drivers/scsi/advansys.c:5555: warning: `req_cnt' might be used uninitialized in this function
drivers/scsi/advansys.c: In function `advansys_detect':
drivers/scsi/dpt_i2o.c:115: warning: `DebugFlags' defined but not used
drivers/scsi/eata_dma.c:1070: warning: `hd' might be used uninitialized in this function
drivers/scsi/eata_dma.c: In function `register_HBA':
drivers/scsi/g_NCR5380.c:963: warning: `id_table' defined but not used
drivers/scsi/NCR5380.c:795: warning: `NCR5380_print_options' defined but not used
drivers/scsi/osst.c:1385: warning: `expected' might be used uninitialized in this function
drivers/scsi/osst.c: In function `osst_reposition_and_retry':
drivers/scsi/qla1280.c:1615: warning: `qla1280_do_dpc' defined but not used
drivers/scsi/scsi_debug.c:656: warning: unused variable `size'
drivers/scsi/scsi_debug.c: In function `scsi_debug_biosparam':
drivers/scsi/scsi.h:402:27: scsi_obsolete.h: No such file or directory
drivers/scsi/sym53c416.c:613: warning: `id_table' defined but not used
drivers/scsi/tmscsim.c:280: warning: `tmscsim_pci_tbl' defined but not used
drivers/scsi/u14-34f.c:1759: parse error before string constant
drivers/scsi/u14-34f.c:1759: warning: data definition has no type or storage class
drivers/scsi/u14-34f.c:1759: warning: function declaration isn't a prototype
drivers/scsi/u14-34f.c:1759: warning: type defaults to `int' in declaration of `MODULE_LICENSE'
drivers/sound/ad1848.c:2870: warning: `id_table' defined but not used
drivers/sound/cmpci.c:1457: warning: unused variable `s'
drivers/sound/cmpci.c: In function `cm_release_mixdev':
drivers/sound/cs4281/cs4281m.c:4478: warning: initialization from incompatible pointer type
drivers/sound/cs4281/cs4281m.c:4479: warning: initialization from incompatible pointer type
drivers/sound/rme96xx.c:1218: warning: unused variable `hwp'
drivers/sound/rme96xx.c: In function `rme96xx_release':
drivers/sound/sb_card.c:531: warning: `id_table' defined but not used
drivers/sound/trident.c:3937: warning: unused variable `mask'
drivers/sound/trident.c: In function `trident_probe':
drivers/telephony/ixj.h:41: warning: `ixj_h_rcsid' defined but not used
drivers/usb/hpusbscsi.c:351: (Each undeclared identifier is reported only once
drivers/usb/hpusbscsi.c:351: for each function it appears in.)
drivers/usb/hpusbscsi.c:351: `SCSI_ABORT_PENDING' undeclared (first use in this function)
drivers/usb/hpusbscsi.c:352: warning: control reaches end of non-void function
drivers/usb/hpusbscsi.c: In function `hpusbscsi_scsi_abort':
drivers/usb/serial/belkin_sa.c:106: warning: `id_table_combined' defined but not used
drivers/usb/serial/digi_acceleport.c:480: warning: `id_table_combined' defined but not used
drivers/usb/serial/ftdi_sio.c:137: warning: `id_table_combined' defined but not used
drivers/usb/serial/io_tables.h:33: warning: `id_table_combined' defined but not used
drivers/usb/serial/keyspan.h:349: warning: `keyspan_ids_combined' defined but not used
drivers/usb/serial/keyspan_pda.c:146: warning: `id_table_combined' defined but not used
drivers/usb/serial/mct_u232.c:129: warning: `id_table_combined' defined but not used
drivers/usb/serial/visor.c:171: warning: `id_table' defined but not used
drivers/usb/serial/whiteheat.c:116: warning: `id_table_combined' defined but not used
drivers/video/aty128fb.c:1616: warning: suggest parentheses around assignment used as truth value
drivers/video/aty128fb.c:216: warning: `font' defined but not used
drivers/video/aty128fb.c:217: warning: `mode' defined but not used
drivers/video/aty128fb.c:218: warning: `nomtrr' defined but not used
drivers/video/aty128fb.c: In function `aty128fb_setup':
drivers/video/aty/atyfb_base.c:2524: warning: suggest parentheses around assignment used as truth value
drivers/video/aty/atyfb_base.c: In function `atyfb_setup':
drivers/video/aty/mach64_accel.c:341:1: warning: pasting "fbcon_aty8" and "=" does not give a valid preprocessing token
drivers/video/aty/mach64_accel.c:344:1: warning: pasting "fbcon_aty16" and "=" does not give a valid preprocessing token
drivers/video/aty/mach64_accel.c:347:1: warning: pasting "fbcon_aty24" and "=" does not give a valid preprocessing token
drivers/video/aty/mach64_accel.c:350:1: warning: pasting "fbcon_aty32" and "=" does not give a valid preprocessing token
drivers/video/hgafb.c:798: warning: suggest parentheses around assignment used as truth value
drivers/video/hgafb.c: In function `hgafb_setup':
drivers/video/matrox/matroxfb_base.c:2358: warning: suggest parentheses around assignment used as truth value
drivers/video/matrox/matroxfb_base.c: In function `matroxfb_setup':
drivers/video/matrox/matroxfb_g450.c:7: warning: `matroxfb_g450_get_reg' defined but not used
drivers/video/radeonfb.c:1873: warning: `fbcon_radeon_clear' defined but not used
drivers/video/radeonfb.c:540: warning: suggest parentheses around assignment used as truth value
drivers/video/radeonfb.c: In function `radeonfb_setup':
drivers/video/riva/fbdev.c:2048: warning: suggest parentheses around assignment used as truth value
drivers/video/riva/fbdev.c: In function `rivafb_setup':
drivers/video/sis/sis_main.c:1729: warning: suggest parentheses around assignment used as truth value
drivers/video/sis/sis_main.c: In function `sisfb_setup':
drivers/video/sstfb.c:1700: warning: suggest parentheses around assignment used as truth value
drivers/video/sstfb.c: In function `sstfb_setup':
drivers/video/tdfxfb.c:2089: warning: suggest parentheses around assignment used as truth value
drivers/video/tdfxfb.c: In function `tdfxfb_setup':
drivers/video/vesafb.c:460: warning: suggest parentheses around assignment used as truth value
drivers/video/vesafb.c: In function `vesafb_setup':
drivers/video/vfb.c:385: warning: suggest parentheses around assignment used as truth value
drivers/video/vfb.c: In function `vfb_setup':
drivers/video/vga16fb.c:695: warning: suggest parentheses around assignment used as truth value
drivers/video/vga16fb.c: In function `vga16fb_setup':
fs/affs/super.c:273:2: warning: #warning
fs/ncpfs/ncplib_kernel.c:56: warning: `ncp_add_mem_fromfs' defined but not used
On Sun, Oct 28, 2001 at 08:58:36PM +1100, Keith Owens wrote:
>
> drivers/usb/serial/belkin_sa.c:106: warning: `id_table_combined' defined but not used
> drivers/usb/serial/digi_acceleport.c:480: warning: `id_table_combined' defined but not used
> drivers/usb/serial/ftdi_sio.c:137: warning: `id_table_combined' defined but not used
> drivers/usb/serial/io_tables.h:33: warning: `id_table_combined' defined but not used
> drivers/usb/serial/keyspan.h:349: warning: `keyspan_ids_combined' defined but not used
> drivers/usb/serial/keyspan_pda.c:146: warning: `id_table_combined' defined but not used
> drivers/usb/serial/mct_u232.c:129: warning: `id_table_combined' defined but not used
> drivers/usb/serial/visor.c:171: warning: `id_table' defined but not used
> drivers/usb/serial/whiteheat.c:116: warning: `id_table_combined' defined but not used
These, and lots of the other pci_id table warnings are due to the tables
being used for MODULE_DEVICE_TABLE() information. When the code is not
compiled as modules, those tables are not needed.
Hm, I guess I should look into some kind of macro to keep this from
happening...
thanks,
greg k-h
On Sun, 28 Oct 2001 10:03:17 -0800,
Greg KH <[email protected]> wrote:
>On Sun, Oct 28, 2001 at 08:58:36PM +1100, Keith Owens wrote:
>> drivers/usb/serial/belkin_sa.c:106: warning: `id_table_combined' defined but not used
>These, and lots of the other pci_id table warnings are due to the tables
>being used for MODULE_DEVICE_TABLE() information. When the code is not
>compiled as modules, those tables are not needed.
Against 2.4.13 or 2.4.13-ac*. It adds a dummy reference (which is then
discarded) for module tables when code is built in. This removes the
spurious warning messages.
Index: 13.1/include/linux/module.h
--- 13.1/include/linux/module.h Tue, 09 Oct 2001 11:09:32 +1000 kaos (linux-2.4/c/b/46_module.h 1.1.1.1.2.6.1.1 644)
+++ 13.13(w)/include/linux/module.h Mon, 29 Oct 2001 12:04:23 +1100 kaos (linux-2.4/c/b/46_module.h 1.1.1.1.2.8 644)
@@ -11,6 +11,7 @@
#include <linux/spinlock.h>
#include <linux/list.h>
+#ifndef CONFIG_KBUILD_2_5
#ifdef __GENKSYMS__
# define _set_ver(sym) sym
# undef MODVERSIONS
@@ -21,6 +22,7 @@
# include <linux/modversions.h>
# endif
#endif /* __GENKSYMS__ */
+#endif /* CONFIG_KBUILD_2_5 */
#include <asm/atomic.h>
@@ -257,8 +259,6 @@ static const unsigned long __module_##gt
__attribute__ ((unused)) = sizeof(struct gtype##_id); \
static const struct gtype##_id * __module_##gtype##_table \
__attribute__ ((unused)) = name
-#define MODULE_DEVICE_TABLE(type,name) \
- MODULE_GENERIC_TABLE(type##_device,name)
/*
* The following license idents are currently accepted as indicating free
@@ -312,8 +312,15 @@ static const char __module_using_checksu
#define MODULE_SUPPORTED_DEVICE(name)
#define MODULE_PARM(var,type)
#define MODULE_PARM_DESC(var,desc)
-#define MODULE_GENERIC_TABLE(gtype,name)
-#define MODULE_DEVICE_TABLE(type,name)
+
+/* Create a dummy reference to the table to suppress gcc unused warnings. Put
+ * the reference in the .data.exit section which is discarded when code is built
+ * in, so the reference does not bloat the running kernel. Note: cannot be
+ * const, other exit data may be writable.
+ */
+#define MODULE_GENERIC_TABLE(gtype,name) \
+static struct gtype##_id * __module_##gtype##_table \
+ __attribute__ ((unused, __section__(".data.exit"))) = name
#ifndef __GENKSYMS__
@@ -327,6 +334,9 @@ extern struct module *module_list;
#endif /* !__GENKSYMS__ */
#endif /* MODULE */
+
+#define MODULE_DEVICE_TABLE(type,name) \
+ MODULE_GENERIC_TABLE(type##_device,name)
/* Export a symbol either from the kernel or a module.
Keith Owens wrote:
> @@ -11,6 +11,7 @@
> #include <linux/spinlock.h>
> #include <linux/list.h>
>
> +#ifndef CONFIG_KBUILD_2_5
> #ifdef __GENKSYMS__
> # define _set_ver(sym) sym
> # undef MODVERSIONS
> @@ -21,6 +22,7 @@
> # include <linux/modversions.h>
> # endif
> #endif /* __GENKSYMS__ */
> +#endif /* CONFIG_KBUILD_2_5 */
>
> #include <asm/atomic.h>
>
> @@ -257,8 +259,6 @@ static const unsigned long __module_##gt
I don't think we need this part of the patch.
Jeff
--
Jeff Garzik | Only so many songs can be sung
Building 1024 | with two lips, two lungs, and one tongue.
MandrakeSoft | - nomeansno
On Mon, 29 Oct 2001 01:17:15 -0500,
Jeff Garzik <[email protected]> wrote:
>Keith Owens wrote:
>> @@ -11,6 +11,7 @@
>> #include <linux/spinlock.h>
>> #include <linux/list.h>
>>
>> +#ifndef CONFIG_KBUILD_2_5
>> #ifdef __GENKSYMS__
>> # define _set_ver(sym) sym
>> # undef MODVERSIONS
>> @@ -21,6 +22,7 @@
>> # include <linux/modversions.h>
>> # endif
>> #endif /* __GENKSYMS__ */
>> +#endif /* CONFIG_KBUILD_2_5 */
>>
>> #include <asm/atomic.h>
>>
>> @@ -257,8 +259,6 @@ static const unsigned long __module_##gt
>
>I don't think we need this part of the patch.
It is a small change to module.h that is required for kbuild 2.5 and
has no effect on 2.4. Saves me having to send another patch later.
Keith Owens wrote:
>
> On Mon, 29 Oct 2001 01:17:15 -0500,
> Jeff Garzik <[email protected]> wrote:
> >Keith Owens wrote:
> >> @@ -11,6 +11,7 @@
> >> #include <linux/spinlock.h>
> >> #include <linux/list.h>
> >>
> >> +#ifndef CONFIG_KBUILD_2_5
> >> #ifdef __GENKSYMS__
> >> # define _set_ver(sym) sym
> >> # undef MODVERSIONS
> >> @@ -21,6 +22,7 @@
> >> # include <linux/modversions.h>
> >> # endif
> >> #endif /* __GENKSYMS__ */
> >> +#endif /* CONFIG_KBUILD_2_5 */
> >>
> >> #include <asm/atomic.h>
> >>
> >> @@ -257,8 +259,6 @@ static const unsigned long __module_##gt
> >
> >I don't think we need this part of the patch.
>
> It is a small change to module.h that is required for kbuild 2.5 and
> has no effect on 2.4. Saves me having to send another patch later.
If everybody did this Linus would get 1001 unnecessary changes for 1002
patches...
Look, this isn't 2.5. Everyone else has to separate out their 2.5
patches. You're not special, you need to do it too... Let's actually
take another half-second when sending patches to make sure Linus only
gets the changes destined for 2.4. Keep the patch noise down, and help
out not only Linus but everyone else reviewing 2.4 code.
Jeff, keeper of tree purity
--
Jeff Garzik | Only so many songs can be sung
Building 1024 | with two lips, two lungs, and one tongue.
MandrakeSoft | - nomeansno
On Sun, 28 Oct 2001, Greg KH wrote:
> These, and lots of the other pci_id table warnings are due to the tables
> being used for MODULE_DEVICE_TABLE() information. When the code is not
> compiled as modules, those tables are not needed.
>
> Hm, I guess I should look into some kind of macro to keep this from
> happening...
A couple of months ago I thought about this and could think of two
possible solutions:
o add a __moddevtable which expands to __devinitdata
__attribute((unused)).
Drawback: Needs changing of all drivers which produce the warning.
o add a variable which references the table within MODULE_DEVICE_TABLE.
I implemented this, minor drawback is that it costs 4 bytes per table.
However, IIRC Keith didn't like it at this time.
The best option, of course, is to move drivers to the new-style pci or
whatever interface, such that the table actually gets used. But the
middle of a stable series is not necessarily the best time to do so.
--Kai
On Mon, Oct 29, 2001 at 12:17:02PM +0100, Kai Germaschewski wrote:
>
> The best option, of course, is to move drivers to the new-style pci or
> whatever interface, such that the table actually gets used. But the
> middle of a stable series is not necessarily the best time to do so.
That's not the problem for the usb-serial drivers. The
MODULE_DEVICE_TABLE referenced structure is only needed if the drivers
are compiled as modules. The code uses a different usb_device_id
structure to register itself with the usb-serial core (see the code for
the reason why.)
Keith's proposed change looks like it will solve this problem nicely.
thanks,
greg k-h