From: Benjamin Herrenschmidt <benh@kernel.crashing.org> To: David Miller <davem@davemloft.net> Cc: aik@au1.ibm.com, aik@ozlabs.ru, sowmini.varadhan@oracle.com, anton@au1.ibm.com, paulus@samba.org, sparclinux@vger.kernel.org, linuxppc-dev@lists.ozlabs.org Subject: Re: Generic IOMMU pooled allocator Date: Tue, 24 Mar 2015 02:08:10 +0000 [thread overview] Message-ID: <1427162890.4770.307.camel@kernel.crashing.org> (raw) In-Reply-To: <20150323.214453.255192641139042325.davem@davemloft.net> On Mon, 2015-03-23 at 21:44 -0400, David Miller wrote: > From: Benjamin Herrenschmidt <benh@kernel.crashing.org> > Date: Tue, 24 Mar 2015 09:21:05 +1100 > > > Dave, what's your feeling there ? Does anybody around still have > > some HW that we can test with ? > > I don't see what the actual problem is. > > Even if you use multiple pools, which we should for scalability on > sun4u too, just do the flush when allocation in _any_ pool wraps > around. > > That's still better than not doing the optimization at all. I agree, I wasn't sure it was good enough for you, so was just putting other options on the table. > That is always going to be correct, and you can use a separate > spinlock to make sure only one thread of control does the full > IOMMU flush at a time. That would have to be done inside the of the flush callback but I don't see a big issue there. For the large pool, we don't keep a hint so we don't know it's wrapped, in fact we purposefully don't use a hint to limit fragmentation on it, but then, it should be used rarely enough that flushing always is, I suspect, a good option. Cheers, Ben.
WARNING: multiple messages have this Message-ID (diff)
From: Benjamin Herrenschmidt <benh@kernel.crashing.org> To: David Miller <davem@davemloft.net> Cc: aik@au1.ibm.com, aik@ozlabs.ru, sowmini.varadhan@oracle.com, anton@au1.ibm.com, paulus@samba.org, sparclinux@vger.kernel.org, linuxppc-dev@lists.ozlabs.org Subject: Re: Generic IOMMU pooled allocator Date: Tue, 24 Mar 2015 13:08:10 +1100 [thread overview] Message-ID: <1427162890.4770.307.camel@kernel.crashing.org> (raw) In-Reply-To: <20150323.214453.255192641139042325.davem@davemloft.net> On Mon, 2015-03-23 at 21:44 -0400, David Miller wrote: > From: Benjamin Herrenschmidt <benh@kernel.crashing.org> > Date: Tue, 24 Mar 2015 09:21:05 +1100 > > > Dave, what's your feeling there ? Does anybody around still have > > some HW that we can test with ? > > I don't see what the actual problem is. > > Even if you use multiple pools, which we should for scalability on > sun4u too, just do the flush when allocation in _any_ pool wraps > around. > > That's still better than not doing the optimization at all. I agree, I wasn't sure it was good enough for you, so was just putting other options on the table. > That is always going to be correct, and you can use a separate > spinlock to make sure only one thread of control does the full > IOMMU flush at a time. That would have to be done inside the of the flush callback but I don't see a big issue there. For the large pool, we don't keep a hint so we don't know it's wrapped, in fact we purposefully don't use a hint to limit fragmentation on it, but then, it should be used rarely enough that flushing always is, I suspect, a good option. Cheers, Ben.
next prev parent reply other threads:[~2015-03-24 2:08 UTC|newest] Thread overview: 71+ messages / expand[flat|nested] mbox.gz Atom feed top 2015-03-19 2:25 Generic IOMMU pooled allocator David Miller 2015-03-19 2:25 ` David Miller 2015-03-19 2:46 ` Benjamin Herrenschmidt 2015-03-19 2:46 ` Benjamin Herrenschmidt 2015-03-19 2:50 ` David Miller 2015-03-19 2:50 ` David Miller 2015-03-19 3:01 ` Benjamin Herrenschmidt 2015-03-19 3:01 ` Benjamin Herrenschmidt 2015-03-19 5:27 ` Alexey Kardashevskiy 2015-03-19 5:27 ` Alexey Kardashevskiy 2015-03-19 13:34 ` Sowmini Varadhan 2015-03-19 13:34 ` Sowmini Varadhan 2015-03-22 19:27 ` Sowmini Varadhan 2015-03-22 19:27 ` Sowmini Varadhan 2015-03-23 16:29 ` David Miller 2015-03-23 16:29 ` David Miller 2015-03-23 16:54 ` Sowmini Varadhan 2015-03-23 16:54 ` Sowmini Varadhan 2015-03-23 19:05 ` David Miller 2015-03-23 19:05 ` David Miller 2015-03-23 19:09 ` Sowmini Varadhan 2015-03-23 19:09 ` Sowmini Varadhan 2015-03-23 22:21 ` Benjamin Herrenschmidt 2015-03-23 22:21 ` Benjamin Herrenschmidt 2015-03-23 23:08 ` Sowmini Varadhan 2015-03-23 23:08 ` Sowmini Varadhan 2015-03-23 23:29 ` chase rayfield 2015-03-24 0:47 ` Benjamin Herrenschmidt 2015-03-24 0:47 ` Benjamin Herrenschmidt 2015-03-24 1:11 ` Sowmini Varadhan 2015-03-24 1:11 ` Sowmini Varadhan 2015-03-24 1:44 ` David Miller 2015-03-24 1:44 ` David Miller 2015-03-24 1:57 ` Sowmini Varadhan 2015-03-24 1:57 ` Sowmini Varadhan 2015-03-24 2:08 ` Benjamin Herrenschmidt [this message] 2015-03-24 2:08 ` Benjamin Herrenschmidt 2015-03-24 2:15 ` David Miller 2015-03-24 2:15 ` David Miller 2015-03-26 0:43 ` cascardo 2015-03-26 0:43 ` cascardo 2015-03-26 0:49 ` Benjamin Herrenschmidt 2015-03-26 0:49 ` Benjamin Herrenschmidt 2015-03-26 10:56 ` Sowmini Varadhan 2015-03-26 10:56 ` Sowmini Varadhan 2015-03-26 22:51 ` David Miller 2015-03-26 23:00 ` David Miller 2015-03-26 23:51 ` Benjamin Herrenschmidt 2015-03-26 23:51 ` Benjamin Herrenschmidt 2015-03-23 22:36 ` Benjamin Herrenschmidt 2015-03-23 22:36 ` Benjamin Herrenschmidt 2015-03-23 23:19 ` Sowmini Varadhan 2015-03-23 23:19 ` Sowmini Varadhan 2015-03-24 0:48 ` Benjamin Herrenschmidt 2015-03-24 0:48 ` Benjamin Herrenschmidt 2015-03-23 22:25 ` Benjamin Herrenschmidt 2015-03-23 22:25 ` Benjamin Herrenschmidt 2015-03-22 19:36 ` Arnd Bergmann 2015-03-22 19:36 ` Arnd Bergmann 2015-03-22 22:02 ` Benjamin Herrenschmidt 2015-03-22 22:02 ` Benjamin Herrenschmidt 2015-03-22 22:07 ` Sowmini Varadhan 2015-03-22 22:07 ` Sowmini Varadhan 2015-03-22 22:22 ` Benjamin Herrenschmidt 2015-03-22 22:22 ` Benjamin Herrenschmidt 2015-03-23 6:04 ` Arnd Bergmann 2015-03-23 6:04 ` Arnd Bergmann 2015-03-23 11:04 ` Benjamin Herrenschmidt 2015-03-23 11:04 ` Benjamin Herrenschmidt 2015-03-23 18:45 ` Arnd Bergmann 2015-03-23 18:45 ` Arnd Bergmann
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=1427162890.4770.307.camel@kernel.crashing.org \ --to=benh@kernel.crashing.org \ --cc=aik@au1.ibm.com \ --cc=aik@ozlabs.ru \ --cc=anton@au1.ibm.com \ --cc=davem@davemloft.net \ --cc=linuxppc-dev@lists.ozlabs.org \ --cc=paulus@samba.org \ --cc=sowmini.varadhan@oracle.com \ --cc=sparclinux@vger.kernel.org \ /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.