The DRM_DP_AUX_CHARDEV and DRM_DP_CEC Kconfig symbols enable code that use DP helper functions, that are only present if CONFIG_DRM_DISPLAY_DP_HELPER is also enabled.
But these don't select the DRM_DISPLAY_DP_HELPER symbol, meaning that it is possible to enable any of them without CONFIG_DRM_DISPLAY_DP_HELPER.
That will lead to the following linking errors with the mentioned config:
LD vmlinux.o MODPOST vmlinux.symvers MODINFO modules.builtin.modinfo GEN modules.builtin LD .tmp_vmlinux.kallsyms1 KSYMS .tmp_vmlinux.kallsyms1.S AS .tmp_vmlinux.kallsyms1.S LD .tmp_vmlinux.kallsyms2 KSYMS .tmp_vmlinux.kallsyms2.S AS .tmp_vmlinux.kallsyms2.S LD vmlinux SYSMAP System.map SORTTAB vmlinux OBJCOPY arch/arm64/boot/Image MODPOST modules-only.symvers ERROR: modpost: "drm_dp_dpcd_write" [drivers/gpu/drm/display/drm_display_helper.ko] undefined! ERROR: modpost: "drm_dp_read_desc" [drivers/gpu/drm/display/drm_display_helper.ko] undefined! ERROR: modpost: "drm_dp_dpcd_read" [drivers/gpu/drm/display/drm_display_helper.ko] undefined! make[1]: *** [scripts/Makefile.modpost:134: modules-only.symvers] Error 1 make[1]: *** Deleting file 'modules-only.symvers' make: *** [Makefile:1749: modules] Error 2
Note: It seems this has been an issue for a long time but was made easier to reproduce after the commit 1e0f66420b13 ("drm/display: Introduce a DRM display-helper module"). Adding a Fixes: tag just to make sure that this fix will be picked for stable once the mentioned change also lands there.
Fixes: 1e0f66420b13 ("drm/display: Introduce a DRM display-helper module") Signed-off-by: Javier Martinez Canillas javierm@redhat.com ---
Changes in v2: - Explain better the issue in the change description. - Only select DRM_DISPLAY_DP_HELPER and not DRM_DISPLAY_HELPER.
drivers/gpu/drm/display/Kconfig | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/display/Kconfig b/drivers/gpu/drm/display/Kconfig index f84f1b0cd23f..9fe80c4e555c 100644 --- a/drivers/gpu/drm/display/Kconfig +++ b/drivers/gpu/drm/display/Kconfig @@ -32,6 +32,7 @@ config DRM_DISPLAY_HDMI_HELPER config DRM_DP_AUX_CHARDEV bool "DRM DP AUX Interface" depends on DRM + select DRM_DISPLAY_DP_HELPER help Choose this option to enable a /dev/drm_dp_auxN node that allows to read and write values to arbitrary DPCD registers on the DP aux @@ -40,6 +41,7 @@ config DRM_DP_AUX_CHARDEV config DRM_DP_CEC bool "Enable DisplayPort CEC-Tunneling-over-AUX HDMI support" depends on DRM + select DRM_DISPLAY_DP_HELPER select CEC_CORE help Choose this option if you want to enable HDMI CEC support for
Hi
Am 27.04.22 um 23:55 schrieb Javier Martinez Canillas:
The DRM_DP_AUX_CHARDEV and DRM_DP_CEC Kconfig symbols enable code that use DP helper functions, that are only present if CONFIG_DRM_DISPLAY_DP_HELPER is also enabled.
But these don't select the DRM_DISPLAY_DP_HELPER symbol, meaning that it is possible to enable any of them without CONFIG_DRM_DISPLAY_DP_HELPER.
That will lead to the following linking errors with the mentioned config:
LD vmlinux.o MODPOST vmlinux.symvers MODINFO modules.builtin.modinfo GEN modules.builtin LD .tmp_vmlinux.kallsyms1 KSYMS .tmp_vmlinux.kallsyms1.S AS .tmp_vmlinux.kallsyms1.S LD .tmp_vmlinux.kallsyms2 KSYMS .tmp_vmlinux.kallsyms2.S AS .tmp_vmlinux.kallsyms2.S LD vmlinux SYSMAP System.map SORTTAB vmlinux OBJCOPY arch/arm64/boot/Image MODPOST modules-only.symvers ERROR: modpost: "drm_dp_dpcd_write" [drivers/gpu/drm/display/drm_display_helper.ko] undefined! ERROR: modpost: "drm_dp_read_desc" [drivers/gpu/drm/display/drm_display_helper.ko] undefined! ERROR: modpost: "drm_dp_dpcd_read" [drivers/gpu/drm/display/drm_display_helper.ko] undefined! make[1]: *** [scripts/Makefile.modpost:134: modules-only.symvers] Error 1 make[1]: *** Deleting file 'modules-only.symvers' make: *** [Makefile:1749: modules] Error 2
Note: It seems this has been an issue for a long time but was made easier to reproduce after the commit 1e0f66420b13 ("drm/display: Introduce a DRM display-helper module"). Adding a Fixes: tag just to make sure that this fix will be picked for stable once the mentioned change also lands there.
Fixes: 1e0f66420b13 ("drm/display: Introduce a DRM display-helper module") Signed-off-by: Javier Martinez Canillas javierm@redhat.com
Changes in v2:
Explain better the issue in the change description.
Only select DRM_DISPLAY_DP_HELPER and not DRM_DISPLAY_HELPER.
drivers/gpu/drm/display/Kconfig | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/display/Kconfig b/drivers/gpu/drm/display/Kconfig index f84f1b0cd23f..9fe80c4e555c 100644 --- a/drivers/gpu/drm/display/Kconfig +++ b/drivers/gpu/drm/display/Kconfig @@ -32,6 +32,7 @@ config DRM_DISPLAY_HDMI_HELPER config DRM_DP_AUX_CHARDEV bool "DRM DP AUX Interface" depends on DRM
- select DRM_DISPLAY_DP_HELPER
You cannot select DISPLAY_DP_HELPER without DISPLAY_HELPER.
Can't you simply make it depend on DISPLAY_DP_HELPER. The menu entry will show up as soon as there's a driver that selcets DISPLAY_DP_HELPER.
Best regards Thomas
help Choose this option to enable a /dev/drm_dp_auxN node that allows to read and write values to arbitrary DPCD registers on the DP aux @@ -40,6 +41,7 @@ config DRM_DP_AUX_CHARDEV config DRM_DP_CEC bool "Enable DisplayPort CEC-Tunneling-over-AUX HDMI support" depends on DRM
- select DRM_DISPLAY_DP_HELPER select CEC_CORE help Choose this option if you want to enable HDMI CEC support for
Hello Thomas,
Thanks for your feedback.
On 4/28/22 09:02, Thomas Zimmermann wrote:
[snip]
Changes in v2:
Explain better the issue in the change description.
Only select DRM_DISPLAY_DP_HELPER and not DRM_DISPLAY_HELPER.
drivers/gpu/drm/display/Kconfig | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/display/Kconfig b/drivers/gpu/drm/display/Kconfig index f84f1b0cd23f..9fe80c4e555c 100644 --- a/drivers/gpu/drm/display/Kconfig +++ b/drivers/gpu/drm/display/Kconfig @@ -32,6 +32,7 @@ config DRM_DISPLAY_HDMI_HELPER config DRM_DP_AUX_CHARDEV bool "DRM DP AUX Interface" depends on DRM
- select DRM_DISPLAY_DP_HELPER
You cannot select DISPLAY_DP_HELPER without DISPLAY_HELPER.
That was my original thought as well and what did in v1, but then I noticed that doing that it would force DRM_DISPLAY_HELPER to be set as built-in and not allow to be built as a module.
Can't you simply make it depend on DISPLAY_DP_HELPER. The menu entry will show up as soon as there's a driver that selcets DISPLAY_DP_HELPER.
I could but then that means that once won't be able to select these two config options unless some enable symbol selects DRM_DISPLAY_DP_HELPER.
In my opinion, DRM_DP_AUX_CHARDEV and DRM_DP_CEC are different than all other options that select DRM_DISPLAY_DP_HELPER, since those are drivers and want to have both DRM_DISPLAY_DP_HELPER and DRM_DISPLAY_HELPER set.
But DRM_DP_AUX_CHARDEV and DRM_DP_CEC are just included in drm_display_helper.o if enabled, and depend on symbols that are present if CONFIG_DRM_DISPLAY_DP_HELPER is enabled. So just need the latter, if DRM_DISPLAY_HELPER is not enabled then it will just be a no-op.
Having written that though I noticed that a "depends on DRM_DISPLAY_HELPER" makes sense. If you agree I can add it and post a v3.
Now, pondering more about this issue, probably the most correct thing to do is for the drivers that make use of the symbols exported by DRM_DP_{AUX_CHARDEV,CEC} to select these. What do you think ? -- Best regards,
Javier Martinez Canillas Linux Engineering Red Hat
Hi
Am 28.04.22 um 09:26 schrieb Javier Martinez Canillas:
Hello Thomas,
Thanks for your feedback.
On 4/28/22 09:02, Thomas Zimmermann wrote:
[snip]
Changes in v2:
Explain better the issue in the change description.
Only select DRM_DISPLAY_DP_HELPER and not DRM_DISPLAY_HELPER.
drivers/gpu/drm/display/Kconfig | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/display/Kconfig b/drivers/gpu/drm/display/Kconfig index f84f1b0cd23f..9fe80c4e555c 100644 --- a/drivers/gpu/drm/display/Kconfig +++ b/drivers/gpu/drm/display/Kconfig @@ -32,6 +32,7 @@ config DRM_DISPLAY_HDMI_HELPER config DRM_DP_AUX_CHARDEV bool "DRM DP AUX Interface" depends on DRM
- select DRM_DISPLAY_DP_HELPER
You cannot select DISPLAY_DP_HELPER without DISPLAY_HELPER.
That was my original thought as well and what did in v1, but then I noticed that doing that it would force DRM_DISPLAY_HELPER to be set as built-in and not allow to be built as a module.
It was a rhetorical only. I didn't mean to actually set DISPLAY_HELPER.
Can't you simply make it depend on DISPLAY_DP_HELPER. The menu entry will show up as soon as there's a driver that selcets DISPLAY_DP_HELPER.
I could but then that means that once won't be able to select these two config options unless some enable symbol selects DRM_DISPLAY_DP_HELPER.
In my opinion, DRM_DP_AUX_CHARDEV and DRM_DP_CEC are different than all other options that select DRM_DISPLAY_DP_HELPER, since those are drivers and want to have both DRM_DISPLAY_DP_HELPER and DRM_DISPLAY_HELPER set.
But DRM_DP_AUX_CHARDEV and DRM_DP_CEC are just included in drm_display_helper.o if enabled, and depend on symbols that are present if CONFIG_DRM_DISPLAY_DP_HELPER is enabled. So just need the latter, if DRM_DISPLAY_HELPER is not enabled then it will just be a no-op.
Having written that though I noticed that a "depends on DRM_DISPLAY_HELPER" makes sense. If you agree I can add it and post a v3.
Yes please. These options enable features of the DP code. If there's no driver with DP, it doesn't make sense to allow them.
I know that there could be an odd situation where userspace might not have DP, but still wants the chardev file of aux bus. But that situation existed already when the code was located within KMS helpers.
Now, pondering more about this issue, probably the most correct thing to do is for the drivers that make use of the symbols exported by DRM_DP_{AUX_CHARDEV,CEC} to select these. What do you think ?
That's not considered good style. Select should not be used for anything that is user-configurable. [1]
Best regards Thomas
[1] https://elixir.bootlin.com/linux/v5.17.4/source/Documentation/kbuild/kconfig...
-- Best regards,
Javier Martinez Canillas Linux Engineering Red Hat
On 4/28/22 09:45, Thomas Zimmermann wrote:
[snip]
You cannot select DISPLAY_DP_HELPER without DISPLAY_HELPER.
That was my original thought as well and what did in v1, but then I noticed that doing that it would force DRM_DISPLAY_HELPER to be set as built-in and not allow to be built as a module.
It was a rhetorical only. I didn't mean to actually set DISPLAY_HELPER.
Ah, sorry for misunderstanding.
Can't you simply make it depend on DISPLAY_DP_HELPER. The menu entry will show up as soon as there's a driver that selcets DISPLAY_DP_HELPER.
I could but then that means that once won't be able to select these two config options unless some enable symbol selects DRM_DISPLAY_DP_HELPER.
In my opinion, DRM_DP_AUX_CHARDEV and DRM_DP_CEC are different than all other options that select DRM_DISPLAY_DP_HELPER, since those are drivers and want to have both DRM_DISPLAY_DP_HELPER and DRM_DISPLAY_HELPER set.
But DRM_DP_AUX_CHARDEV and DRM_DP_CEC are just included in drm_display_helper.o if enabled, and depend on symbols that are present if CONFIG_DRM_DISPLAY_DP_HELPER is enabled. So just need the latter, if DRM_DISPLAY_HELPER is not enabled then it will just be a no-op.
Having written that though I noticed that a "depends on DRM_DISPLAY_HELPER" makes sense. If you agree I can add it and post a v3.
Yes please. These options enable features of the DP code. If there's no driver with DP, it doesn't make sense to allow them.
I know that there could be an odd situation where userspace might not have DP, but still wants the chardev file of aux bus. But that situation existed already when the code was located within KMS helpers.
Agreed.
Now, pondering more about this issue, probably the most correct thing to do is for the drivers that make use of the symbols exported by DRM_DP_{AUX_CHARDEV,CEC} to select these. What do you think ?
That's not considered good style. Select should not be used for anything that is user-configurable. [1]
Right. So giving even more thought to this, now I think that we should just include drm_dp_aux_dev.o, drm_dp_cec.o (and probably drm_dp_aux_bus.o?) unconditionally to drm_display_helper-$(CONFIG_DRM_DISPLAY_DP_HELPER).
After all, these are not big objects and drm_display_helper can now be built as module.
I don't see that much value to have separate user-configurable config options...
Hi
Am 28.04.22 um 09:52 schrieb Javier Martinez Canillas:
On 4/28/22 09:45, Thomas Zimmermann wrote:
[snip]
You cannot select DISPLAY_DP_HELPER without DISPLAY_HELPER.
That was my original thought as well and what did in v1, but then I noticed that doing that it would force DRM_DISPLAY_HELPER to be set as built-in and not allow to be built as a module.
It was a rhetorical only. I didn't mean to actually set DISPLAY_HELPER.
Ah, sorry for misunderstanding.
Can't you simply make it depend on DISPLAY_DP_HELPER. The menu entry will show up as soon as there's a driver that selcets DISPLAY_DP_HELPER.
I could but then that means that once won't be able to select these two config options unless some enable symbol selects DRM_DISPLAY_DP_HELPER.
In my opinion, DRM_DP_AUX_CHARDEV and DRM_DP_CEC are different than all other options that select DRM_DISPLAY_DP_HELPER, since those are drivers and want to have both DRM_DISPLAY_DP_HELPER and DRM_DISPLAY_HELPER set.
But DRM_DP_AUX_CHARDEV and DRM_DP_CEC are just included in drm_display_helper.o if enabled, and depend on symbols that are present if CONFIG_DRM_DISPLAY_DP_HELPER is enabled. So just need the latter, if DRM_DISPLAY_HELPER is not enabled then it will just be a no-op.
Having written that though I noticed that a "depends on DRM_DISPLAY_HELPER" makes sense. If you agree I can add it and post a v3.
Yes please. These options enable features of the DP code. If there's no driver with DP, it doesn't make sense to allow them.
I know that there could be an odd situation where userspace might not have DP, but still wants the chardev file of aux bus. But that situation existed already when the code was located within KMS helpers.
Agreed.
Now, pondering more about this issue, probably the most correct thing to do is for the drivers that make use of the symbols exported by DRM_DP_{AUX_CHARDEV,CEC} to select these. What do you think ?
That's not considered good style. Select should not be used for anything that is user-configurable. [1]
Right. So giving even more thought to this, now I think that we should just include drm_dp_aux_dev.o, drm_dp_cec.o (and probably drm_dp_aux_bus.o?) unconditionally to drm_display_helper-$(CONFIG_DRM_DISPLAY_DP_HELPER).
After all, these are not big objects and drm_display_helper can now be built as module.
I don't see that much value to have separate user-configurable config options...
I don't know the side effects of this. We're exporting another device file after all.
For know I'd make them depend on DRM_DISPLAY_DP_HELPER. If someone complains we can revert and fix the linker error by adding stub functions for the missing symbols.
Best regards Thomas
On 4/28/22 10:04, Thomas Zimmermann wrote:
[snip]
Right. So giving even more thought to this, now I think that we should just include drm_dp_aux_dev.o, drm_dp_cec.o (and probably drm_dp_aux_bus.o?) unconditionally to drm_display_helper-$(CONFIG_DRM_DISPLAY_DP_HELPER).
After all, these are not big objects and drm_display_helper can now be built as module.
I don't see that much value to have separate user-configurable config options...
I don't know the side effects of this. We're exporting another device file after all.
For know I'd make them depend on DRM_DISPLAY_DP_HELPER. If someone complains we can revert and fix the linker error by adding stub functions for the missing symbols.
Ok, I'll just add the depends on then to fix the linking errors and then either adding the stubs when disabled or just making it part of the DP_HELPER can be done as a follow-up.
(cc Jani and Lyude)
Am 28.04.22 um 09:52 schrieb Javier Martinez Canillas:
On 4/28/22 09:45, Thomas Zimmermann wrote:
[snip]
You cannot select DISPLAY_DP_HELPER without DISPLAY_HELPER.
That was my original thought as well and what did in v1, but then I noticed that doing that it would force DRM_DISPLAY_HELPER to be set as built-in and not allow to be built as a module.
It was a rhetorical only. I didn't mean to actually set DISPLAY_HELPER.
Ah, sorry for misunderstanding.
Can't you simply make it depend on DISPLAY_DP_HELPER. The menu entry will show up as soon as there's a driver that selcets DISPLAY_DP_HELPER.
I could but then that means that once won't be able to select these two config options unless some enable symbol selects DRM_DISPLAY_DP_HELPER.
In my opinion, DRM_DP_AUX_CHARDEV and DRM_DP_CEC are different than all other options that select DRM_DISPLAY_DP_HELPER, since those are drivers and want to have both DRM_DISPLAY_DP_HELPER and DRM_DISPLAY_HELPER set.
But DRM_DP_AUX_CHARDEV and DRM_DP_CEC are just included in drm_display_helper.o if enabled, and depend on symbols that are present if CONFIG_DRM_DISPLAY_DP_HELPER is enabled. So just need the latter, if DRM_DISPLAY_HELPER is not enabled then it will just be a no-op.
Having written that though I noticed that a "depends on DRM_DISPLAY_HELPER" makes sense. If you agree I can add it and post a v3.
Yes please. These options enable features of the DP code. If there's no driver with DP, it doesn't make sense to allow them.
I know that there could be an odd situation where userspace might not have DP, but still wants the chardev file of aux bus. But that situation existed already when the code was located within KMS helpers.
Agreed.
Now, pondering more about this issue, probably the most correct thing to do is for the drivers that make use of the symbols exported by DRM_DP_{AUX_CHARDEV,CEC} to select these. What do you think ?
That's not considered good style. Select should not be used for anything that is user-configurable. [1]
Right. So giving even more thought to this, now I think that we should just include drm_dp_aux_dev.o, drm_dp_cec.o (and probably drm_dp_aux_bus.o?) unconditionally to drm_display_helper-$(CONFIG_DRM_DISPLAY_DP_HELPER).
After all, these are not big objects and drm_display_helper can now be built as module.
I don't see that much value to have separate user-configurable config options...
dri-devel@lists.freedesktop.org
-
Javier Martinez Canillas -
Thomas Zimmermann