From: Arnd Bergmann <arnd@arndb.de> To: linux-arm-kernel@lists.infradead.org Cc: Alan Stern <stern@rowland.harvard.edu>, Nicolas Pitre <nico@fluxnic.net>, gregkh@suse.de, linux-usb@vger.kernel.org, lkml <linux-kernel@vger.kernel.org>, Rabin Vincent <rabin@rab.in>, Alexander Holler <holler@ahsoftware.de> Subject: Re: [PATCH] USB: ehci: use packed,aligned(4) instead of removing the packed attribute Date: Sun, 19 Jun 2011 22:02:23 +0200 [thread overview] Message-ID: <201106192202.23989.arnd@arndb.de> (raw) In-Reply-To: <Pine.LNX.4.44L0.1106191453200.14586-100000@netrider.rowland.org> On Sunday 19 June 2011 21:00:01 Alan Stern wrote: > On Sun, 19 Jun 2011, Nicolas Pitre wrote: > > On Thu, 16 Jun 2011, Arnd Bergmann wrote: > > > On Thursday 16 June 2011 22:10:53 Alexander Holler wrote: > > > At least I would be happier without the patch. I'm trying to convince > > > people to not use these attributes unless required because too much > > > harm is done when they are used without understanding the full > > > consequences. I also recommend using __packed as localized as possible, > > > i.e. set it for the members that need it, not the entire struct. > > > > > > I agree that your patch is harmless, it's just the opposite of > > > a cleanup in my opinion. > > > > The question is: does the structure really has to be packed? > > What do you mean? The structure really does need to be allocated > without padding between the fields; is that the same thing? So do a > bunch of other structures that currently have no annotations at all. I guess the issue is that some ABIs actually require a minimum alignment, like the old ARM ABI that you can still use to build the kernel. If a structure is not a multiple of four bytes in size, that ABI will add padding at the end, e.g. in struct s { char c[2]; }; struct t { struct s t1; unsigned short t2[3]; }; On most architectures, struct s will be two bytes in size and one byte aligned, while struct t is eight bytes and two byte aligned. On ARM oABI, struct s ends up with four byte size and alignment while struct t is twelve bytes long. All this is ok for regular structures, but not when they are used to describe memory layout of hardware registers on on-wire packets. > > If it does, then the follow-up question is: is a packing on word > > boundaries sufficient? > > > If the answer is yes in both cases, then having packed,aligned(4) is not > > a frivolity but rather a correctness issue. > > Why so? Current systems work just fine without it. I think Nicolas got it backwards here, adding both packed and aligned(4) would make a structure like the one above consistently incorrect when used to describe a tightly packed hardware structure. In this case, we would have to do struct s { char c[2]; } __packed; struct t { struct s t1; unsigned short t2[3] __aligned(2); } __packed; To tell the compiler that t2 is indeed aligned, while struct t is packed to include no padding around t. I actually recently stumbled over code that gets this wrong, see http://git.kernel.org/?p=linux/kernel/git/arnd/playground.git;a=commit;h=284cef173aafd531a708f48e71a9cc7249fc8a98 > > We can of course provide a > > define in include/linux/compiler-gcc.hto hide the ugliness of it > > somewhat: > > > > #define __packed_32 __attribute__((packed,aligned(4))) > > > > I suspect that the vast majority of the __packed uses in the kernel > > would be better with this __packed_32 instead, the actual need and > > intent would be more clearly expressed, and the generated code in the > > presence of those GCC changes would then be way more efficient and still > > correct. > > What if the intent is that the structure should be 4-byte aligned on > 32-bit systems and 8-byte aligned on 64-bit systems? The compiler > already does this sort of thing automatically, why mess with it? Different issue. Arnd
WARNING: multiple messages have this Message-ID (diff)
From: arnd@arndb.de (Arnd Bergmann) To: linux-arm-kernel@lists.infradead.org Subject: [PATCH] USB: ehci: use packed, aligned(4) instead of removing the packed attribute Date: Sun, 19 Jun 2011 22:02:23 +0200 [thread overview] Message-ID: <201106192202.23989.arnd@arndb.de> (raw) In-Reply-To: <Pine.LNX.4.44L0.1106191453200.14586-100000@netrider.rowland.org> On Sunday 19 June 2011 21:00:01 Alan Stern wrote: > On Sun, 19 Jun 2011, Nicolas Pitre wrote: > > On Thu, 16 Jun 2011, Arnd Bergmann wrote: > > > On Thursday 16 June 2011 22:10:53 Alexander Holler wrote: > > > At least I would be happier without the patch. I'm trying to convince > > > people to not use these attributes unless required because too much > > > harm is done when they are used without understanding the full > > > consequences. I also recommend using __packed as localized as possible, > > > i.e. set it for the members that need it, not the entire struct. > > > > > > I agree that your patch is harmless, it's just the opposite of > > > a cleanup in my opinion. > > > > The question is: does the structure really has to be packed? > > What do you mean? The structure really does need to be allocated > without padding between the fields; is that the same thing? So do a > bunch of other structures that currently have no annotations at all. I guess the issue is that some ABIs actually require a minimum alignment, like the old ARM ABI that you can still use to build the kernel. If a structure is not a multiple of four bytes in size, that ABI will add padding at the end, e.g. in struct s { char c[2]; }; struct t { struct s t1; unsigned short t2[3]; }; On most architectures, struct s will be two bytes in size and one byte aligned, while struct t is eight bytes and two byte aligned. On ARM oABI, struct s ends up with four byte size and alignment while struct t is twelve bytes long. All this is ok for regular structures, but not when they are used to describe memory layout of hardware registers on on-wire packets. > > If it does, then the follow-up question is: is a packing on word > > boundaries sufficient? > > > If the answer is yes in both cases, then having packed,aligned(4) is not > > a frivolity but rather a correctness issue. > > Why so? Current systems work just fine without it. I think Nicolas got it backwards here, adding both packed and aligned(4) would make a structure like the one above consistently incorrect when used to describe a tightly packed hardware structure. In this case, we would have to do struct s { char c[2]; } __packed; struct t { struct s t1; unsigned short t2[3] __aligned(2); } __packed; To tell the compiler that t2 is indeed aligned, while struct t is packed to include no padding around t. I actually recently stumbled over code that gets this wrong, see http://git.kernel.org/?p=linux/kernel/git/arnd/playground.git;a=commit;h=284cef173aafd531a708f48e71a9cc7249fc8a98 > > We can of course provide a > > define in include/linux/compiler-gcc.hto hide the ugliness of it > > somewhat: > > > > #define __packed_32 __attribute__((packed,aligned(4))) > > > > I suspect that the vast majority of the __packed uses in the kernel > > would be better with this __packed_32 instead, the actual need and > > intent would be more clearly expressed, and the generated code in the > > presence of those GCC changes would then be way more efficient and still > > correct. > > What if the intent is that the structure should be 4-byte aligned on > 32-bit systems and 8-byte aligned on 64-bit systems? The compiler > already does this sort of thing automatically, why mess with it? Different issue. Arnd
next prev parent reply other threads:[~2011-06-19 20:03 UTC|newest] Thread overview: 124+ messages / expand[flat|nested] mbox.gz Atom feed top 2011-04-27 14:34 [PATCH] echi: remove structure packing from ehci_def Rabin Vincent 2011-04-27 14:34 ` Rabin Vincent 2011-04-27 15:15 ` Sergei Shtylyov 2011-04-27 15:15 ` Sergei Shtylyov 2011-04-27 15:37 ` [PATCHv2] " Rabin Vincent 2011-04-27 15:37 ` Rabin Vincent 2011-06-16 16:17 ` [PATCH] USB: ehci: use packed,aligned(4) instead of removing the packed attribute Alexander Holler 2011-06-16 16:17 ` [PATCH] USB: ehci: use packed, aligned(4) " Alexander Holler 2011-06-16 17:09 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-16 17:09 ` Alan Stern 2011-06-16 17:55 ` Arnd Bergmann 2011-06-16 17:55 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-16 19:25 ` [PATCH] USB: ehci: use packed,aligned(4) " Alexander Holler 2011-06-16 19:25 ` Alexander Holler 2011-06-16 19:46 ` Alan Stern 2011-06-16 19:46 ` Alan Stern 2011-06-16 20:10 ` Alexander Holler 2011-06-16 20:10 ` Alexander Holler 2011-06-16 20:20 ` Arnd Bergmann 2011-06-16 20:20 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-19 15:02 ` [PATCH] USB: ehci: use packed,aligned(4) " Nicolas Pitre 2011-06-19 15:02 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-19 19:00 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-19 19:00 ` Alan Stern 2011-06-19 20:02 ` Arnd Bergmann [this message] 2011-06-19 20:02 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-19 20:11 ` [PATCH] USB: ehci: use packed,aligned(4) " Arnd Bergmann 2011-06-19 20:11 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-19 21:39 ` [PATCH] USB: ehci: use packed,aligned(4) " Nicolas Pitre 2011-06-19 21:39 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-19 21:27 ` [PATCH] USB: ehci: use packed,aligned(4) " Nicolas Pitre 2011-06-19 21:27 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-20 15:03 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-20 15:03 ` Alan Stern 2011-06-20 16:16 ` Nicolas Pitre 2011-06-20 16:16 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-20 16:48 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-20 16:48 ` Alan Stern 2011-06-20 16:58 ` Arnd Bergmann 2011-06-20 16:58 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-20 19:02 ` Russell King - ARM Linux 2011-06-20 19:02 ` Russell King - ARM Linux 2011-06-20 19:20 ` Nicolas Pitre 2011-06-20 19:20 ` Nicolas Pitre 2011-06-20 19:29 ` Nicolas Pitre 2011-06-20 19:29 ` Nicolas Pitre 2011-06-20 17:10 ` [PATCH] USB: ehci: use packed,aligned(4) " Nicolas Pitre 2011-06-20 17:10 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-20 17:35 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-20 17:35 ` Alan Stern 2011-06-20 18:48 ` Russell King - ARM Linux 2011-06-20 18:48 ` Russell King - ARM Linux 2011-06-20 20:26 ` Arnd Bergmann 2011-06-20 20:26 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-20 20:50 ` Nicolas Pitre 2011-06-20 20:50 ` Nicolas Pitre 2011-06-20 20:55 ` [PATCH] USB: ehci: use packed,aligned(4) " Russell King - ARM Linux 2011-06-20 20:55 ` Russell King - ARM Linux 2011-06-20 21:23 ` Arnd Bergmann 2011-06-20 21:23 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-20 22:23 ` [PATCH] USB: ehci: use packed,aligned(4) " Nicolas Pitre 2011-06-20 22:23 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-21 11:25 ` [PATCH] USB: ehci: use packed,aligned(4) " Arnd Bergmann 2011-06-21 11:25 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-25 1:25 ` Nicolas Pitre 2011-06-25 8:09 ` Arnd Bergmann 2011-06-28 18:51 ` Nicolas Pitre 2011-06-29 10:56 ` Arnd Bergmann 2011-06-20 19:14 ` [PATCH] USB: ehci: use packed,aligned(4) " Nicolas Pitre 2011-06-20 19:14 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-20 19:32 ` Russell King - ARM Linux 2011-06-20 19:32 ` Russell King - ARM Linux 2011-06-20 20:14 ` Arnd Bergmann 2011-06-20 20:14 ` Arnd Bergmann 2011-06-20 20:42 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-20 20:42 ` Alan Stern 2011-06-20 22:36 ` Nicolas Pitre 2011-06-20 22:36 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-21 15:06 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-21 15:06 ` Alan Stern 2011-06-20 17:39 ` Alexander Holler 2011-06-20 17:39 ` Alexander Holler 2011-06-20 18:39 ` Alan Stern 2011-06-20 18:39 ` Alan Stern 2011-06-20 18:46 ` Alexander Holler 2011-06-20 18:46 ` Alexander Holler 2011-06-20 18:57 ` Alan Stern 2011-06-20 18:57 ` Alan Stern 2011-06-20 19:56 ` Nicolas Pitre 2011-06-20 19:56 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-20 21:04 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-20 21:04 ` Alan Stern 2011-06-20 22:31 ` Nicolas Pitre 2011-06-20 22:31 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-21 14:58 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-21 14:58 ` Alan Stern 2011-06-21 20:41 ` Nicolas Pitre 2011-06-21 20:41 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-22 6:23 ` [PATCH] USB: ehci: use packed,aligned(4) " Alexander Holler 2011-06-22 6:23 ` Alexander Holler 2011-06-20 20:09 ` Arnd Bergmann 2011-06-20 20:09 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-20 21:05 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-20 21:05 ` Alan Stern 2011-06-20 20:07 ` Arnd Bergmann 2011-06-20 20:07 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-20 20:28 ` [PATCH] USB: ehci: use packed,aligned(4) " Nicolas Pitre 2011-06-20 20:28 ` [PATCH] USB: ehci: use packed, aligned(4) " Nicolas Pitre 2011-06-20 20:39 ` Arnd Bergmann 2011-06-20 20:39 ` Arnd Bergmann 2011-06-20 21:03 ` Nicolas Pitre 2011-06-20 21:03 ` Nicolas Pitre 2011-06-23 9:47 ` Alexander Holler 2011-06-23 9:47 ` Alexander Holler 2011-06-23 14:25 ` Alan Stern 2011-06-23 14:25 ` Alan Stern 2011-06-24 11:40 ` Alexander Holler 2011-06-24 11:40 ` Alexander Holler 2011-06-20 16:26 ` [PATCH] USB: ehci: use packed,aligned(4) " Arnd Bergmann 2011-06-20 16:26 ` [PATCH] USB: ehci: use packed, aligned(4) " Arnd Bergmann 2011-06-16 20:30 ` [PATCH] USB: ehci: use packed,aligned(4) " Alan Stern 2011-06-16 20:30 ` Alan Stern 2011-06-16 18:16 ` Alexander Holler 2011-06-16 18:16 ` Alexander Holler
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=201106192202.23989.arnd@arndb.de \ --to=arnd@arndb.de \ --cc=gregkh@suse.de \ --cc=holler@ahsoftware.de \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-usb@vger.kernel.org \ --cc=nico@fluxnic.net \ --cc=rabin@rab.in \ --cc=stern@rowland.harvard.edu \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.