Re: [PATCH v2 2/3] randstruct: Force full rebuild when seed changes

2025-05-02 Thread Masahiro Yamada
On Sat, May 3, 2025 at 7:54 AM Kees Cook wrote: > > While the randstruct GCC plugin was being rebuilt if the randstruct > seed changed, Clangs build did not notice the change. Include the hash > header directly so that it becomes a universal build dependency and full > rebuilds will happen if it c

Re: [PATCH v2 1/3] gcc-plugins: Force full rebuild when plugins change

2025-05-02 Thread Masahiro Yamada
On Sat, May 3, 2025 at 7:54 AM Kees Cook wrote: > > There was no dependency between the plugins changing and the rest of the > kernel being built. Enforce this by including a synthetic header file > when using plugins, that is regenerated any time the plugins are built. > > This cannot be included

[PATCH 1/3] um: Remove obsolete legacy network transports

2025-05-02 Thread Tiwei Bie
These legacy network transports were marked as obsolete in commit 40814b98a570 ("um: Mark non-vector net transports as obsolete"). More than five years have passed since then. Remove these network transports to reduce the maintenance burden. Suggested-by: Anton Ivanov Signed-off-by: Tiwei Bie --

[PATCH 0/3] um: Remove legacy network transport support

2025-05-02 Thread Tiwei Bie
Tiwei Bie (3): um: Remove obsolete legacy network transports um: vector: Eliminate the dependency on uml_net um: Remove legacy network transport infrastructure .../virt/uml/user_mode_linux_howto_v2.rst | 47 +- arch/um/configs/i386_defconfig| 7 - arch/um/configs/x86_

[PATCH 2/3] um: vector: Eliminate the dependency on uml_net

2025-05-02 Thread Tiwei Bie
The only dependency on uml_net (i.e., the legacy network transport infrastructure) is the call to uml_net_setup_etheraddr(). Implement it inside vector to eliminate the uml_net dependency completely. It will allow us to remove uml_net in the next step. Signed-off-by: Tiwei Bie --- arch/um/driver

[PATCH 3/3] um: Remove legacy network transport infrastructure

2025-05-02 Thread Tiwei Bie
All legacy network transports have been removed. Vector transports provide the same capabilities with significantly higher network throughput. There is no reason to keep the legacy network transport infrastructure anymore. Remove it to reduce the maintenance burden. Signed-off-by: Tiwei Bie ---

Re: [PATCH] um: Merge os-Linux/drivers/ into drivers/

2025-05-02 Thread Tiwei Bie
On 2025/5/3 00:09, Johannes Berg wrote: > On Fri, 2025-05-02 at 10:08 +0100, Anton Ivanov wrote: >> >> On 02/05/2025 09:35, Thomas Meyer wrote: >>> Hi, >>> >>> Aren't those drivers under exactly this folder because those drivers will >>> work only with when the host OS is Linux? I mean this is all

[PATCH v2 1/3] gcc-plugins: Force full rebuild when plugins change

2025-05-02 Thread Kees Cook
There was no dependency between the plugins changing and the rest of the kernel being built. Enforce this by including a synthetic header file when using plugins, that is regenerated any time the plugins are built. This cannot be included via '-include ...' because Makefiles use the "filter-out" s

[PATCH v2 2/3] randstruct: Force full rebuild when seed changes

2025-05-02 Thread Kees Cook
While the randstruct GCC plugin was being rebuilt if the randstruct seed changed, Clangs build did not notice the change. Include the hash header directly so that it becomes a universal build dependency and full rebuilds will happen if it changes. Since we cannot use "-include ..." as the randstru

[PATCH v2 0/3] Detect changed compiler dependencies for full rebuild

2025-05-02 Thread Kees Cook
v2: - switch from -include to -I with a -D gated include compiler-version.h v1: https://lore.kernel.org/lkml/20250501193839.work.525-k...@kernel.org/ Hi, This is my attempt to introduce dependencies that track the various compiler behaviors that may globally change the build that aren't repre

[PATCH v2 3/3] integer-wrap: Force full rebuild when .scl file changes

2025-05-02 Thread Kees Cook
Since the integer wrapping sanitizer's behavior depends on its associated .scl file, we must force a full rebuild if the file changes. Universally include a synthetic header file that is rebuilt when the .scl file changes, via compiler-version.h, since using "-include ..." is not possible in the ca

Re: [PATCH] um: Merge os-Linux/drivers/ into drivers/

2025-05-02 Thread Johannes Berg
On Fri, 2025-05-02 at 10:08 +0100, Anton Ivanov wrote: > > On 02/05/2025 09:35, Thomas Meyer wrote: > > Hi, > > > > Aren't those drivers under exactly this folder because those drivers will > > work only with when the host OS is Linux? I mean this is all hypothetical > > but when somebody is go

Re: [PATCH] um: Merge os-Linux/drivers/ into drivers/

2025-05-02 Thread Anton Ivanov
On 02/05/2025 09:31, Tiwei Bie wrote: Most of the drivers are stored in drivers/, with only two drivers kept in os-Linux/drivers/. Merge os-Linux/drivers/ into drivers/ to simplify and unify the code organization. Signed-off-by: Tiwei Bie --- arch/um/drivers/Makefile

Re: [PATCH] um: Merge os-Linux/drivers/ into drivers/

2025-05-02 Thread Tiwei Bie
On 2025/5/2 16:58, Anton Ivanov wrote: > On 02/05/2025 09:31, Tiwei Bie wrote: >> Most of the drivers are stored in drivers/, with only two drivers >> kept in os-Linux/drivers/. Merge os-Linux/drivers/ into drivers/ >> to simplify and unify the code organization. >> >> Signed-off-by: Tiwei Bie >>

Re: [PATCH] um: Merge os-Linux/drivers/ into drivers/

2025-05-02 Thread Anton Ivanov
On 02/05/2025 09:35, Thomas Meyer wrote: Hi, Aren't those drivers under exactly this folder because those drivers will work only with when the host OS is Linux? I mean this is all hypothetical but when somebody is going to implement UML under Mac OS, this layout won't work anymore, or do I

[PATCH] um: Merge os-Linux/drivers/ into drivers/

2025-05-02 Thread Tiwei Bie
Most of the drivers are stored in drivers/, with only two drivers kept in os-Linux/drivers/. Merge os-Linux/drivers/ into drivers/ to simplify and unify the code organization. Signed-off-by: Tiwei Bie --- arch/um/drivers/Makefile | 4 arch/um/{os-Linux => }/drivers/et