* Re: [PATCH] tools/testing/cxl: Don't wrap cxl_core's own exported symbols
2026-07-21 8:40 [PATCH] tools/testing/cxl: Don't wrap cxl_core's own exported symbols Richard Cheng
@ 2026-07-21 16:22 ` Dave Jiang
2026-08-17 23:18 ` Alison Schofield
2026-08-19 14:53 ` shaikh kamaluddin
2 siblings, 0 replies; 5+ messages in thread
From: Dave Jiang @ 2026-07-21 16:22 UTC (permalink / raw)
To: Richard Cheng, dave, jic23, alison.schofield, vishal.l.verma,
djbw, danwilliams
Cc: iweiny, ming.li, linux-cxl, linux-kernel, newtonl, kristinc,
kaihengf, kobak
On 7/21/26 1:40 AM, Richard Cheng wrote:
> Building cxl_test with LLVM=1 (ld.lld) fails in modpost:
>
> """
> cxl_core: .export_symbol section references '__wrap_...', but it
> does not seem to be an export symbol
> """
>
> cxl_test --wraps 6 symbols that cxl_core both defines and exports.
> GNU ld only wraps undefined references, but ld.lld also redirects
> references to defined symbols, so cxl_core's own EXPORT_SYMBOL entries
> get rewritten to __wrap_* and modpost rejects them.
>
> cxl_core never calls these symbols, so wrap them only in the caller
> modules via LDFLAGS_<module>.o instead of the global ldflags-y.
>
> Signed-off-by: Richard Cheng <icheng@nvidia.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
> ---
> tools/testing/cxl/Kbuild | 16 ++++++++++------
> 1 file changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/tools/testing/cxl/Kbuild b/tools/testing/cxl/Kbuild
> index 2be1df80fcc9..e49ce644b4b1 100644
> --- a/tools/testing/cxl/Kbuild
> +++ b/tools/testing/cxl/Kbuild
> @@ -4,16 +4,20 @@ ldflags-y += --wrap=is_acpi_device_node
> ldflags-y += --wrap=acpi_evaluate_integer
> ldflags-y += --wrap=acpi_pci_find_root
> ldflags-y += --wrap=nvdimm_bus_register
> -ldflags-y += --wrap=cxl_await_media_ready
> -ldflags-y += --wrap=devm_cxl_add_rch_dport
> -ldflags-y += --wrap=cxl_endpoint_parse_cdat
> -ldflags-y += --wrap=devm_cxl_endpoint_decoders_setup
> ldflags-y += --wrap=hmat_get_extended_linear_cache_size
> -ldflags-y += --wrap=devm_cxl_add_dport_by_dev
> -ldflags-y += --wrap=devm_cxl_switch_port_decoders_setup
> ldflags-y += --wrap=walk_hmem_resources
> ldflags-y += --wrap=region_intersects
> ldflags-y += --wrap=region_intersects_soft_reserve
> +wrap_cxl_core := --wrap=cxl_await_media_ready
> +wrap_cxl_core += --wrap=devm_cxl_add_rch_dport
> +wrap_cxl_core += --wrap=devm_cxl_add_dport_by_dev
> +wrap_cxl_core += --wrap=devm_cxl_switch_port_decoders_setup
> +wrap_cxl_core += --wrap=devm_cxl_endpoint_decoders_setup
> +wrap_cxl_core += --wrap=cxl_endpoint_parse_cdat
> +LDFLAGS_cxl_acpi.o += $(wrap_cxl_core)
> +LDFLAGS_cxl_pmem.o += $(wrap_cxl_core)
> +LDFLAGS_cxl_port.o += $(wrap_cxl_core)
> +LDFLAGS_cxl_mem.o += $(wrap_cxl_core)
>
> DRIVERS := ../../../drivers
> DAX_HMEM_SRC := $(DRIVERS)/dax/hmem
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] tools/testing/cxl: Don't wrap cxl_core's own exported symbols
2026-07-21 8:40 [PATCH] tools/testing/cxl: Don't wrap cxl_core's own exported symbols Richard Cheng
2026-07-21 16:22 ` Dave Jiang
@ 2026-08-17 23:18 ` Alison Schofield
2026-08-19 14:53 ` shaikh kamaluddin
2 siblings, 0 replies; 5+ messages in thread
From: Alison Schofield @ 2026-08-17 23:18 UTC (permalink / raw)
To: Richard Cheng
Cc: dave, jic23, dave.jiang, vishal.l.verma, djbw, danwilliams,
iweiny, ming.li, linux-cxl, linux-kernel, newtonl, kristinc,
kaihengf, kobak
On Tue, Jul 21, 2026 at 04:40:09PM +0800, Richard Cheng wrote:
> Building cxl_test with LLVM=1 (ld.lld) fails in modpost:
>
> """
> cxl_core: .export_symbol section references '__wrap_...', but it
> does not seem to be an export symbol
> """
>
> cxl_test --wraps 6 symbols that cxl_core both defines and exports.
> GNU ld only wraps undefined references, but ld.lld also redirects
> references to defined symbols, so cxl_core's own EXPORT_SYMBOL entries
> get rewritten to __wrap_* and modpost rejects them.
>
> cxl_core never calls these symbols, so wrap them only in the caller
> modules via LDFLAGS_<module>.o instead of the global ldflags-y.
Reviewed-by: Alison Schofield <alison.schofield@intel.com>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] tools/testing/cxl: Don't wrap cxl_core's own exported symbols
2026-07-21 8:40 [PATCH] tools/testing/cxl: Don't wrap cxl_core's own exported symbols Richard Cheng
2026-07-21 16:22 ` Dave Jiang
2026-08-17 23:18 ` Alison Schofield
@ 2026-08-19 14:53 ` shaikh kamaluddin
2026-08-20 2:09 ` Richard Cheng
2 siblings, 1 reply; 5+ messages in thread
From: shaikh kamaluddin @ 2026-08-19 14:53 UTC (permalink / raw)
To: Richard Cheng
Cc: dave, jic23, dave.jiang, alison.schofield, vishal.l.verma, djbw,
danwilliams, iweiny, ming.li, linux-cxl, linux-kernel, newtonl,
kristinc, kaihengf, kobak
On Tue, Jul 21, 2026 at 04:40:09PM +0800, Richard Cheng wrote:
> Building cxl_test with LLVM=1 (ld.lld) fails in modpost:
>
> """
> cxl_core: .export_symbol section references '__wrap_...', but it
> does not seem to be an export symbol
> """
>
> cxl_test --wraps 6 symbols that cxl_core both defines and exports.
> GNU ld only wraps undefined references, but ld.lld also redirects
> references to defined symbols, so cxl_core's own EXPORT_SYMBOL entries
> get rewritten to __wrap_* and modpost rejects them.
>
> cxl_core never calls these symbols, so wrap them only in the caller
> modules via LDFLAGS_<module>.o instead of the global ldflags-y.
>
> Signed-off-by: Richard Cheng <icheng@nvidia.com>
> ---
> tools/testing/cxl/Kbuild | 16 ++++++++++------
> 1 file changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/tools/testing/cxl/Kbuild b/tools/testing/cxl/Kbuild
> index 2be1df80fcc9..e49ce644b4b1 100644
> --- a/tools/testing/cxl/Kbuild
> +++ b/tools/testing/cxl/Kbuild
> @@ -4,16 +4,20 @@ ldflags-y += --wrap=is_acpi_device_node
> ldflags-y += --wrap=acpi_evaluate_integer
> ldflags-y += --wrap=acpi_pci_find_root
> ldflags-y += --wrap=nvdimm_bus_register
> -ldflags-y += --wrap=cxl_await_media_ready
> -ldflags-y += --wrap=devm_cxl_add_rch_dport
> -ldflags-y += --wrap=cxl_endpoint_parse_cdat
> -ldflags-y += --wrap=devm_cxl_endpoint_decoders_setup
> ldflags-y += --wrap=hmat_get_extended_linear_cache_size
> -ldflags-y += --wrap=devm_cxl_add_dport_by_dev
> -ldflags-y += --wrap=devm_cxl_switch_port_decoders_setup
> ldflags-y += --wrap=walk_hmem_resources
> ldflags-y += --wrap=region_intersects
> ldflags-y += --wrap=region_intersects_soft_reserve
> +wrap_cxl_core := --wrap=cxl_await_media_ready
> +wrap_cxl_core += --wrap=devm_cxl_add_rch_dport
> +wrap_cxl_core += --wrap=devm_cxl_add_dport_by_dev
> +wrap_cxl_core += --wrap=devm_cxl_switch_port_decoders_setup
> +wrap_cxl_core += --wrap=devm_cxl_endpoint_decoders_setup
> +wrap_cxl_core += --wrap=cxl_endpoint_parse_cdat
> +LDFLAGS_cxl_acpi.o += $(wrap_cxl_core)
> +LDFLAGS_cxl_pmem.o += $(wrap_cxl_core)
> +LDFLAGS_cxl_port.o += $(wrap_cxl_core)
> +LDFLAGS_cxl_mem.o += $(wrap_cxl_core)
>
> DRIVERS := ../../../drivers
> DAX_HMEM_SRC := $(DRIVERS)/dax/hmem
> --
> 2.43.0
>
Hi Richard,
I was hitting the same tools/testing/cxl build failure on current cxl/next.
modpost was failing on wrapped CXL symbols, for example:
.export_symbol section references '__wrap_devm_cxl_add_rch_dport',
but it does not seem to be an export symbol
.export_symbol section references '__wrap_devm_cxl_add_dport_by_dev',
but it does not seem to be an export symbol
.export_symbol section references '__wrap_cxl_await_media_ready',
but it does not seem to be an export symbol
I found this patch while investigating the failure and applied it to my tree.
With the patch applied, the wrapped-symbol build issue is resolved and tools/testing/cxl
builds successfully for me.
Tested-by: Shaikh Kamaluddin <shaikhkamal2012@gmail.com>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] tools/testing/cxl: Don't wrap cxl_core's own exported symbols
2026-08-19 14:53 ` shaikh kamaluddin
@ 2026-08-20 2:09 ` Richard Cheng
0 siblings, 0 replies; 5+ messages in thread
From: Richard Cheng @ 2026-08-20 2:09 UTC (permalink / raw)
To: shaikh kamaluddin
Cc: dave, jic23, dave.jiang, alison.schofield, vishal.l.verma, djbw,
danwilliams, iweiny, ming.li, linux-cxl, linux-kernel, newtonl,
kristinc, kaihengf, kobak
On Wed, Aug 19, 2026 at 08:23:47PM +0800, shaikh kamaluddin wrote:
> On Tue, Jul 21, 2026 at 04:40:09PM +0800, Richard Cheng wrote:
> > Building cxl_test with LLVM=1 (ld.lld) fails in modpost:
> >
> > """
> > cxl_core: .export_symbol section references '__wrap_...', but it
> > does not seem to be an export symbol
> > """
> >
> > cxl_test --wraps 6 symbols that cxl_core both defines and exports.
> > GNU ld only wraps undefined references, but ld.lld also redirects
> > references to defined symbols, so cxl_core's own EXPORT_SYMBOL entries
> > get rewritten to __wrap_* and modpost rejects them.
> >
> > cxl_core never calls these symbols, so wrap them only in the caller
> > modules via LDFLAGS_<module>.o instead of the global ldflags-y.
> >
> > Signed-off-by: Richard Cheng <icheng@nvidia.com>
> > ---
> > tools/testing/cxl/Kbuild | 16 ++++++++++------
> > 1 file changed, 10 insertions(+), 6 deletions(-)
> >
> > diff --git a/tools/testing/cxl/Kbuild b/tools/testing/cxl/Kbuild
> > index 2be1df80fcc9..e49ce644b4b1 100644
> > --- a/tools/testing/cxl/Kbuild
> > +++ b/tools/testing/cxl/Kbuild
> > @@ -4,16 +4,20 @@ ldflags-y += --wrap=is_acpi_device_node
> > ldflags-y += --wrap=acpi_evaluate_integer
> > ldflags-y += --wrap=acpi_pci_find_root
> > ldflags-y += --wrap=nvdimm_bus_register
> > -ldflags-y += --wrap=cxl_await_media_ready
> > -ldflags-y += --wrap=devm_cxl_add_rch_dport
> > -ldflags-y += --wrap=cxl_endpoint_parse_cdat
> > -ldflags-y += --wrap=devm_cxl_endpoint_decoders_setup
> > ldflags-y += --wrap=hmat_get_extended_linear_cache_size
> > -ldflags-y += --wrap=devm_cxl_add_dport_by_dev
> > -ldflags-y += --wrap=devm_cxl_switch_port_decoders_setup
> > ldflags-y += --wrap=walk_hmem_resources
> > ldflags-y += --wrap=region_intersects
> > ldflags-y += --wrap=region_intersects_soft_reserve
> > +wrap_cxl_core := --wrap=cxl_await_media_ready
> > +wrap_cxl_core += --wrap=devm_cxl_add_rch_dport
> > +wrap_cxl_core += --wrap=devm_cxl_add_dport_by_dev
> > +wrap_cxl_core += --wrap=devm_cxl_switch_port_decoders_setup
> > +wrap_cxl_core += --wrap=devm_cxl_endpoint_decoders_setup
> > +wrap_cxl_core += --wrap=cxl_endpoint_parse_cdat
> > +LDFLAGS_cxl_acpi.o += $(wrap_cxl_core)
> > +LDFLAGS_cxl_pmem.o += $(wrap_cxl_core)
> > +LDFLAGS_cxl_port.o += $(wrap_cxl_core)
> > +LDFLAGS_cxl_mem.o += $(wrap_cxl_core)
> >
> > DRIVERS := ../../../drivers
> > DAX_HMEM_SRC := $(DRIVERS)/dax/hmem
> > --
> > 2.43.0
> >
>
> Hi Richard,
>
> I was hitting the same tools/testing/cxl build failure on current cxl/next.
> modpost was failing on wrapped CXL symbols, for example:
>
> .export_symbol section references '__wrap_devm_cxl_add_rch_dport',
> but it does not seem to be an export symbol
>
> .export_symbol section references '__wrap_devm_cxl_add_dport_by_dev',
> but it does not seem to be an export symbol
>
> .export_symbol section references '__wrap_cxl_await_media_ready',
> but it does not seem to be an export symbol
>
> I found this patch while investigating the failure and applied it to my tree.
> With the patch applied, the wrapped-symbol build issue is resolved and tools/testing/cxl
> builds successfully for me.
>
> Tested-by: Shaikh Kamaluddin <shaikhkamal2012@gmail.com>
Hi Shaikh,
Happy to hear that, appreciate for this.
Best regards,
Richard Cheng.
^ permalink raw reply [flat|nested] 5+ messages in thread