Skip to contentSkip to conversation

usb: gadget: CONFIG_USB_S3C_HSOTG should depend on HAS_DMA

PATCH/RFC Geert Uytterhoeven 2 participants 3 messages May 6, 2014 USB

lore
Share & tools
lore
#0 lore raw

If NO_DMA=y:

drivers/built-in.o: In function `s3c_hsotg_map_dma':
s3c-hsotg.c:(.text+0x375b2c): undefined reference to usb_gadget_map_request' drivers/built-in.o: In function s3c_hsotg_unmap_dma':
s3c-hsotg.c:(.text+0x376a32): undefined reference to `usb_gadget_unmap_request' make[3]: *** [vmlinux] Error 1

Note that all callers of s3c_hsotg_map_dma()/s3c_hsotg_unmap_dma() are protected by a call into the using_dma() checking function:

static inline bool using_dma(struct s3c_hsotg *hsotg) {
return false; /* support is not complete */
}

but not all versions of gcc manage to optimize the check away, causing a link error.

---
Perhaps the DMA code should be removed completely, as its unused, or be protected by #ifdef CONFIG_HAS_DMA instead?

 drivers/usb/gadget/Kconfig |    1 +
 1 file changed, 1 insertion(+)
drivers/usb/gadget/Kconfig +1 -0
index 3557c7e5040d..0e9dead51a27 100644
@@ -302,6 +302,7 @@ config USB_PXA27X
 
 config USB_S3C_HSOTG
 	tristate "Designware/S3C HS/OtG USB Device controller"
+	depends on HAS_DMA
 	help
 	  The Designware USB2.0 high-speed gadget controller
 	  integrated into many SoCs.
Signature · 1 line
1.7.9.5
#1 lore raw

From: Geert Uytterhoeven

Quoted 11 lines

If NO_DMA=y:

drivers/built-in.o: In function `s3c_hsotg_map_dma':
s3c-hsotg.c:(.text+0x375b2c): undefined reference to usb_gadget_map_request' drivers/built-in.o: In function s3c_hsotg_unmap_dma':
s3c-hsotg.c:(.text+0x376a32): undefined reference to `usb_gadget_unmap_request' make[3]: *** [vmlinux] Error 1

Note that all callers of s3c_hsotg_map_dma()/s3c_hsotg_unmap_dma() are protected by a call into the using_dma() checking function:

Quoted 7 lines

static inline bool using_dma(struct s3c_hsotg *hsotg) {
return false; /* support is not complete */
}

but not all versions of gcc manage to optimize the check away, causing a link error.

A have seen gcc optimise away a call, but leave the undefined symbol in the symbol table.
That might be what is happening here.

I don't know a generic solution, I used 'objcopy -N__udivdi3' to remove the explicit unwanted reference from my driver.

David

#2 lore raw

Hi David,

Quoted 16 lines · David Laight
Quoted 16 lines

If NO_DMA=y:

drivers/built-in.o: In function `s3c_hsotg_map_dma':
s3c-hsotg.c:(.text+0x375b2c): undefined reference to usb_gadget_map_request' drivers/built-in.o: In function s3c_hsotg_unmap_dma':
s3c-hsotg.c:(.text+0x376a32): undefined reference to `usb_gadget_unmap_request' make[3]: *** [vmlinux] Error 1

Note that all callers of s3c_hsotg_map_dma()/s3c_hsotg_unmap_dma() are protected by a call into the using_dma() checking function:

static inline bool using_dma(struct s3c_hsotg *hsotg) {
return false; /* support is not complete */
}

Quoted 6 lines
Quoted 2 lines

but not all versions of gcc manage to optimize the check away, causing a link error.

A have seen gcc optimise away a call, but leave the undefined symbol in the symbol table.
That might be what is happening here.

Indeed, the unused functions s3c_hsotg_{,un}map_dma() are still there, but the calls to them have been optimized away.

Marking s3c_hsotg_{,un}map_dma() inline fixes that. I'll send a new patch doing that, as it's a bit silly to require "depends on HAS_DMA" for a driver which has DMA disabled.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds