2023-01-13 04:52:29

by Peter Foley

[permalink] [raw]
Subject: [PATCH 0/4] um: Various build fixes for allyesconfig

Fix some build failures I ran across when trying to build an allyesconfig
kernel for ARCH=um.

To: Richard Weinberger <[email protected]>
To: Anton Ivanov <[email protected]>
To: Johannes Berg <[email protected]>
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Peter Foley <[email protected]>

---
Peter Foley (4):
um: Make the definition of cpu_data more compatible
um: Avoid pcap multiple definition errors
um: Prevent building modules incompatible with MODVERSIONS
um: Use CFLAGS_vmlinux

arch/um/Makefile | 3 +--
arch/um/drivers/Kconfig | 2 ++
arch/um/drivers/pcap_kern.c | 4 ++--
arch/um/include/asm/processor-generic.h | 2 +-
arch/um/kernel/um_arch.c | 2 +-
5 files changed, 7 insertions(+), 6 deletions(-)
---
base-commit: d9fc1511728c15df49ff18e49a494d00f78b7cd4
change-id: 20230112-um-3f06b1bbbbb0

Best regards,
--
Peter Foley <[email protected]>


2023-01-13 04:52:48

by Peter Foley

[permalink] [raw]
Subject: [PATCH 1/4] um: Make the definition of cpu_data more compatible

Match the x86 implementation to improve build errors.
Noticed when building allyesconfig.

e.g.
../arch/um/include/asm/processor-generic.h:94:19: error: called object is not a function or function pointer
94 | #define cpu_data (&boot_cpu_data)
| ~^~~~~~~~~~~~~~~
../drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_topology.c:2157:16: note: in expansion of macro ‘cpu_data’
2157 | return cpu_data(first_cpu_of_numa_node).apicid;

Signed-off-by: Peter Foley <[email protected]>
---
arch/um/include/asm/processor-generic.h | 2 +-
arch/um/kernel/um_arch.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/um/include/asm/processor-generic.h b/arch/um/include/asm/processor-generic.h
index bb5f06480da9..7414154b8e9a 100644
--- a/arch/um/include/asm/processor-generic.h
+++ b/arch/um/include/asm/processor-generic.h
@@ -91,7 +91,7 @@ struct cpuinfo_um {

extern struct cpuinfo_um boot_cpu_data;

-#define cpu_data (&boot_cpu_data)
+#define cpu_data(cpu) boot_cpu_data
#define current_cpu_data boot_cpu_data
#define cache_line_size() (boot_cpu_data.cache_alignment)

diff --git a/arch/um/kernel/um_arch.c b/arch/um/kernel/um_arch.c
index 786b44dc20c9..8dcda617b8bf 100644
--- a/arch/um/kernel/um_arch.c
+++ b/arch/um/kernel/um_arch.c
@@ -96,7 +96,7 @@ static int show_cpuinfo(struct seq_file *m, void *v)

static void *c_start(struct seq_file *m, loff_t *pos)
{
- return *pos < nr_cpu_ids ? cpu_data + *pos : NULL;
+ return *pos < nr_cpu_ids ? &boot_cpu_data + *pos : NULL;
}

static void *c_next(struct seq_file *m, void *v, loff_t *pos)

--
2.39.0

2023-01-13 04:53:12

by Peter Foley

[permalink] [raw]
Subject: [PATCH 2/4] um: Avoid pcap multiple definition errors

Change the function name in pcap_kern to avoid conflicting with
libpcap.a.

e.g.
ld: /usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../lib64/libpcap.a(pcap.o): in function `pcap_init':
(.text+0x7f0): multiple definition of `pcap_init'; arch/um/drivers/pcap_kern.o:pcap_kern.c:(.text.unlikely+0x0): first defined here
---
arch/um/drivers/pcap_kern.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/um/drivers/pcap_kern.c b/arch/um/drivers/pcap_kern.c
index cfe4cb17694c..25ee2c97ca21 100644
--- a/arch/um/drivers/pcap_kern.c
+++ b/arch/um/drivers/pcap_kern.c
@@ -15,7 +15,7 @@ struct pcap_init {
char *filter;
};

-void pcap_init(struct net_device *dev, void *data)
+void pcap_init_kern(struct net_device *dev, void *data)
{
struct uml_net_private *pri;
struct pcap_data *ppri;
@@ -44,7 +44,7 @@ static int pcap_write(int fd, struct sk_buff *skb, struct uml_net_private *lp)
}

static const struct net_kern_info pcap_kern_info = {
- .init = pcap_init,
+ .init = pcap_init_kern,
.protocol = eth_protocol,
.read = pcap_read,
.write = pcap_write,

--
2.39.0

2023-01-13 04:53:47

by Peter Foley

[permalink] [raw]
Subject: [PATCH 3/4] um: Prevent building modules incompatible with MODVERSIONS

The manual ld invocation in arch/um/drivers doesn't play nicely with
genksyms. Given the problematic modules are deprecated anyway, just
prevent building them when using MODVERSIONS.

e.g.
MODPOST Module.symvers
arch/um/drivers/.pcap.o.cmd: No such file or directory
---
arch/um/drivers/Kconfig | 2 ++
1 file changed, 2 insertions(+)

diff --git a/arch/um/drivers/Kconfig b/arch/um/drivers/Kconfig
index a4f0a19fbe14..36911b1fddcf 100644
--- a/arch/um/drivers/Kconfig
+++ b/arch/um/drivers/Kconfig
@@ -261,6 +261,7 @@ config UML_NET_VECTOR
config UML_NET_VDE
bool "VDE transport (obsolete)"
depends on UML_NET
+ depends on !MODVERSIONS
select MAY_HAVE_RUNTIME_DEPS
help
This User-Mode Linux network transport allows one or more running
@@ -309,6 +310,7 @@ config UML_NET_MCAST
config UML_NET_PCAP
bool "pcap transport (obsolete)"
depends on UML_NET
+ depends on !MODVERSIONS
select MAY_HAVE_RUNTIME_DEPS
help
The pcap transport makes a pcap packet stream on the host look

--
2.39.0

2023-01-13 04:54:29

by Peter Foley

[permalink] [raw]
Subject: [PATCH 4/4] um: Use CFLAGS_vmlinux

link-vmlinux.sh doesn't use LDFLAGS_vmlinux when linking the kernel for
UML. Move the LDFLAGS_EXESTACK options into CFLAGS_vmlinux so they're
actually respected.

e.g.
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: warning: .tmp_vmlinux.kallsyms3.o: missing .note.GNU-stack section implies executable stack
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: NOTE: This behaviour is deprecated and will be removed in a future version of the linker
/usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: warning: vmlinux has a LOAD segment with RWX permissions
---
arch/um/Makefile | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/arch/um/Makefile b/arch/um/Makefile
index f1d4d67157be..93dfb23bd263 100644
--- a/arch/um/Makefile
+++ b/arch/um/Makefile
@@ -139,11 +139,10 @@ ifeq ($(CONFIG_LD_IS_BFD),y)
LDFLAGS_EXECSTACK += $(call ld-option,--no-warn-rwx-segments)
endif

-LD_FLAGS_CMDLINE = $(foreach opt,$(KBUILD_LDFLAGS),-Wl,$(opt))
+LD_FLAGS_CMDLINE = $(foreach opt,$(KBUILD_LDFLAGS) $(LDFLAGS_EXECSTACK),-Wl,$(opt))

# Used by link-vmlinux.sh which has special support for um link
export CFLAGS_vmlinux := $(LINK-y) $(LINK_WRAPS) $(LD_FLAGS_CMDLINE)
-export LDFLAGS_vmlinux := $(LDFLAGS_EXECSTACK)

# When cleaning we don't include .config, so we don't include
# TT or skas makefiles and don't clean skas_ptregs.h.

--
2.39.0

2023-01-13 07:16:10

by David Gow

[permalink] [raw]
Subject: Re: [PATCH 4/4] um: Use CFLAGS_vmlinux

On Fri, 13 Jan 2023 at 12:49, Peter Foley <[email protected]> wrote:
>
> link-vmlinux.sh doesn't use LDFLAGS_vmlinux when linking the kernel for
> UML. Move the LDFLAGS_EXESTACK options into CFLAGS_vmlinux so they're
> actually respected.
>
> e.g.
> /usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: warning: .tmp_vmlinux.kallsyms3.o: missing .note.GNU-stack section implies executable stack
> /usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: NOTE: This behaviour is deprecated and will be removed in a future version of the linker
> /usr/lib/gcc/x86_64-pc-linux-gnu/12/../../../../x86_64-pc-linux-gnu/bin/ld: warning: vmlinux has a LOAD segment with RWX permissions
> ---

Thanks very much for fixing this -- the LDFLAGS/CFLAGS fun here always
trips me up!

Reviewed-by: David Gow <[email protected]>

Cheers,
-- David


> arch/um/Makefile | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/arch/um/Makefile b/arch/um/Makefile
> index f1d4d67157be..93dfb23bd263 100644
> --- a/arch/um/Makefile
> +++ b/arch/um/Makefile
> @@ -139,11 +139,10 @@ ifeq ($(CONFIG_LD_IS_BFD),y)
> LDFLAGS_EXECSTACK += $(call ld-option,--no-warn-rwx-segments)
> endif
>
> -LD_FLAGS_CMDLINE = $(foreach opt,$(KBUILD_LDFLAGS),-Wl,$(opt))
> +LD_FLAGS_CMDLINE = $(foreach opt,$(KBUILD_LDFLAGS) $(LDFLAGS_EXECSTACK),-Wl,$(opt))
>
> # Used by link-vmlinux.sh which has special support for um link
> export CFLAGS_vmlinux := $(LINK-y) $(LINK_WRAPS) $(LD_FLAGS_CMDLINE)
> -export LDFLAGS_vmlinux := $(LDFLAGS_EXECSTACK)
>
> # When cleaning we don't include .config, so we don't include
> # TT or skas makefiles and don't clean skas_ptregs.h.
>
> --
> 2.39.0
>
> _______________________________________________
> linux-um mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-um


Attachments:
smime.p7s (3.91 kB)
S/MIME Cryptographic Signature