linux-kernel - Re: [PATCH v3 0/2] Cleanup io.h

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <fa3b2ef5-e2f8-473f-9e9c-c855d70230cf@app.fastmail.com>
Date: Mon, 03 Mar 2025 09:31:06 +0100
From: "Arnd Bergmann" <arnd@...db.de>
To: "Andy Shevchenko" <andriy.shevchenko@...ux.intel.com>,
 "Raag Jadav" <raag.jadav@...el.com>
Cc: "Greg Kroah-Hartman" <gregkh@...uxfoundation.org>,
 "Dave Airlie" <airlied@...il.com>, "Simona Vetter" <simona@...ll.ch>,
 dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 0/2] Cleanup io.h

On Mon, Mar 3, 2025, at 08:48, Andy Shevchenko wrote:
> On Fri, Feb 28, 2025 at 09:08:15PM +0200, Raag Jadav wrote:
>> > > > Ah, you mean devres related?
>> > > 
>> > > Yeah, couldn't find it on Arnd's tree and I'm not sure if this series
>> > > works without it.
>> > 
>> > But err.h is included in the io.h, no? Or did I misunderstand the point?
>> 
>> First patch on the immutable tag moves IOMEM_ERR_PTR() to err.h and here
>> we're dropping err.h from io.h. So without the tag this series will probably
>> break IOMEM_ERR_PTR().
>
> I see, I think it might be due to some other includes that make this happen.
> Whatever, I assume that Arnd heavily tested this anyway, so it's good to go
> independently on the immutable tag..

I see that err.h gets included through at least linux/string.h,
linux/logic_pio.h and (depending on the architecture architectures)
asm/page.h, so there is no huge risk of something breaking here ;-)

Ideally all of the above should be removed here, but to do
it right, the linux/pgtable.h dependency needs to also be replaced
with a smaller pgprot_t specific one, which takes a bit of
cross-platform work as the definition is not in a consistent
place at the moment.

The change below still builds fine, and that likely indicates
that the same headers are still included from either asm/io.h
or asm/pgtable.h on the architectures I'm testing.

    Arnd

----
diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h
index 402020b23423..60a162591072 100644
--- a/include/asm-generic/io.h
+++ b/include/asm-generic/io.h
@@ -7,16 +7,16 @@
 #ifndef __ASM_GENERIC_IO_H
 #define __ASM_GENERIC_IO_H
 
-#include <asm/page.h> /* I/O is all done through memory accesses */
-#include <linux/string.h> /* for memset() and memcpy() */
 #include <linux/sizes.h>
 #include <linux/types.h>
 #include <linux/instruction_pointer.h>
+#include <linux/compiler.h>
 
 #ifdef CONFIG_GENERIC_IOMAP
 #include <asm-generic/iomap.h>
 #endif
 
+#include <asm/byteorder.h>
 #include <asm/mmiowb.h>
 #include <asm-generic/pci_iomap.h>
 
diff --git a/include/linux/logic_pio.h b/include/linux/logic_pio.h
index 8f1a9408302f..a83286bf13ab 100644
--- a/include/linux/logic_pio.h
+++ b/include/linux/logic_pio.h
@@ -8,7 +8,9 @@
 #ifndef __LINUX_LOGIC_PIO_H
 #define __LINUX_LOGIC_PIO_H
 
-#include <linux/fwnode.h>
+#include <linux/types.h>
+
+struct fwnode_handle;
 
 enum {
        LOGIC_PIO_INDIRECT,             /* Indirect IO flag */