From: Michael Neuling <mikey@neuling.org> To: Anton Blanchard <anton@samba.org> Cc: greg@kroah.com, arnd@arndb.de, mpe@ellerman.id.au, benh@kernel.crashing.org, linux-kernel@vger.kernel.org, linuxppc-dev@ozlabs.org, jk@ozlabs.org, imunsie@au.ibm.com, cbe-oss-dev@lists.ozlabs.org Subject: Re: [PATCH 10/15] powerpc/mm: Add hooks for cxl Date: Fri, 26 Sep 2014 21:33:29 +1000 [thread overview] Message-ID: <1411731209.1733.51.camel@ale.ozlabs.ibm.com> (raw) In-Reply-To: <20140926143309.33c0eeac@kryten> On Fri, 2014-09-26 at 14:33 +1000, Anton Blanchard wrote: > > From: Ian Munsie <imunsie@au1.ibm.com> > > > > This add a hook into tlbie() so that we use global invalidations when > > there are cxl contexts active. > > > > Normally cxl snoops broadcast tlbie. cxl can have TLB entries > > invalidated via MMIO, but we aren't doing that yet. So for now we > > are just disabling local tlbies when cxl contexts are active. In > > future we can make tlbie() local mode smarter so that it invalidates > > cxl contexts explicitly when it needs to. > > > > This also adds a hooks for when SLBs are invalidated to ensure any > > corresponding SLBs in cxl are also invalidated at the same time. > > > > Signed-off-by: Ian Munsie <imunsie@au1.ibm.com> > > Signed-off-by: Michael Neuling <mikey@neuling.org> > > > + use_local = local && mmu_has_feature(MMU_FTR_TLBIEL) && !cxl_ctx_in_use(); > > Seems reasonable until we can get the MMIO based optimisation in. > > Will all CAPI cached translations be invalidated before we finish using > a CAPI context? I'm not sure I understand. Can you elaborate? > And conversely, could CAPI cache any translations when a > context isn't active? The kernel invalidates all translations when the file descriptor is closed. So no, unless the PSL was badly behaving and ignoring the invalidations.... but if we can't trust the PSL we're screwed. > I'm mostly concerned that we can't have a > situation where badly behaving userspace could result in a stale > translation. We only map what a user processes maps and we tear it down when the process is teared down (on the file descriptor release). So I think we are ok. Unless there's some lazy teardown you're alluding to that I'm missing? > > > spu_flush_all_slbs(mm); > > #endif > > + cxl_slbia(mm); > > > spu_flush_all_slbs(mm); > > #endif > > + cxl_slbia(mm); > > > spu_flush_all_slbs(mm); > > #endif > > + cxl_slbia(mm); > > > spu_flush_all_slbs(mm); > > #endif > > + cxl_slbia(mm); > > Should we combine the SPU vs CXL callouts into something common - > perhaps copro_flush_all_slbs()? Sounds good. I'll update. Mikey
WARNING: multiple messages have this Message-ID (diff)
From: Michael Neuling <mikey@neuling.org> To: Anton Blanchard <anton@samba.org> Cc: cbe-oss-dev@lists.ozlabs.org, arnd@arndb.de, greg@kroah.com, linux-kernel@vger.kernel.org, imunsie@au.ibm.com, linuxppc-dev@ozlabs.org, jk@ozlabs.org Subject: Re: [PATCH 10/15] powerpc/mm: Add hooks for cxl Date: Fri, 26 Sep 2014 21:33:29 +1000 [thread overview] Message-ID: <1411731209.1733.51.camel@ale.ozlabs.ibm.com> (raw) In-Reply-To: <20140926143309.33c0eeac@kryten> On Fri, 2014-09-26 at 14:33 +1000, Anton Blanchard wrote: > > From: Ian Munsie <imunsie@au1.ibm.com> > >=20 > > This add a hook into tlbie() so that we use global invalidations when > > there are cxl contexts active. > >=20 > > Normally cxl snoops broadcast tlbie. cxl can have TLB entries > > invalidated via MMIO, but we aren't doing that yet. So for now we > > are just disabling local tlbies when cxl contexts are active. In > > future we can make tlbie() local mode smarter so that it invalidates > > cxl contexts explicitly when it needs to. > >=20 > > This also adds a hooks for when SLBs are invalidated to ensure any > > corresponding SLBs in cxl are also invalidated at the same time. > >=20 > > Signed-off-by: Ian Munsie <imunsie@au1.ibm.com> > > Signed-off-by: Michael Neuling <mikey@neuling.org> >=20 > > + use_local =3D local && mmu_has_feature(MMU_FTR_TLBIEL) && !cxl_ctx_in= _use(); >=20 > Seems reasonable until we can get the MMIO based optimisation in. >=20 > Will all CAPI cached translations be invalidated before we finish using > a CAPI context?=20 I'm not sure I understand. Can you elaborate? > And conversely, could CAPI cache any translations when a > context isn't active?=20 The kernel invalidates all translations when the file descriptor is closed. So no, unless the PSL was badly behaving and ignoring the invalidations.... but if we can't trust the PSL we're screwed. > I'm mostly concerned that we can't have a > situation where badly behaving userspace could result in a stale > translation. We only map what a user processes maps and we tear it down when the process is teared down (on the file descriptor release). So I think we are ok. =20 Unless there's some lazy teardown you're alluding to that I'm missing? >=20 > > spu_flush_all_slbs(mm); > > #endif > > + cxl_slbia(mm); >=20 > > spu_flush_all_slbs(mm); > > #endif > > + cxl_slbia(mm); >=20 > > spu_flush_all_slbs(mm); > > #endif > > + cxl_slbia(mm); >=20 > > spu_flush_all_slbs(mm); > > #endif > > + cxl_slbia(mm); >=20 > Should we combine the SPU vs CXL callouts into something common - > perhaps copro_flush_all_slbs()? Sounds good. I'll update. Mikey
next prev parent reply other threads:[~2014-09-26 11:33 UTC|newest] Thread overview: 86+ messages / expand[flat|nested] mbox.gz Atom feed top 2014-09-18 8:26 [PATCH 0/15] POWER8 Coherent Accelerator device driver Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 01/15] powerpc/cell: Move spu_handle_mm_fault() out of cell platform Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-18 10:00 ` Jeremy Kerr 2014-09-18 10:00 ` Jeremy Kerr 2014-09-18 23:26 ` Michael Neuling 2014-09-18 23:26 ` Michael Neuling 2014-09-26 3:57 ` Anton Blanchard 2014-09-26 3:57 ` Anton Blanchard 2014-09-18 8:26 ` [PATCH 02/15] powerpc/cell: Move data segment faulting code " Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-18 10:27 ` Jeremy Kerr 2014-09-18 10:27 ` Jeremy Kerr 2014-09-18 23:45 ` Michael Neuling 2014-09-18 23:45 ` Michael Neuling 2014-09-26 4:05 ` Anton Blanchard 2014-09-26 4:05 ` Anton Blanchard 2014-09-26 11:19 ` Michael Neuling 2014-09-26 11:19 ` Michael Neuling 2014-09-29 8:30 ` Aneesh Kumar K.V 2014-09-29 8:30 ` Aneesh Kumar K.V 2014-09-30 4:40 ` Michael Neuling 2014-09-30 4:40 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 03/15] powerpc/msi: Improve IRQ bitmap allocator Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-19 20:16 ` Scott Wood 2014-09-19 20:16 ` Scott Wood 2014-09-19 20:19 ` Scott Wood 2014-09-19 20:19 ` Scott Wood 2014-09-22 8:26 ` Laurentiu Tudor 2014-09-22 8:26 ` Laurentiu Tudor 2014-09-22 23:50 ` Scott Wood 2014-09-22 23:50 ` Scott Wood 2014-09-22 8:25 ` Laurentiu Tudor 2014-09-22 8:25 ` Laurentiu Tudor 2014-09-22 8:29 ` Laurentiu Tudor 2014-09-22 8:29 ` Laurentiu Tudor 2014-09-22 22:59 ` Michael Neuling 2014-09-22 22:59 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 04/15] powerpc/mm: Export mmu_kernel_ssize and mmu_linear_psize Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 05/15] powerpc/powernv: Split out set MSI IRQ chip code Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-19 6:54 ` Gavin Shan 2014-09-19 6:54 ` Gavin Shan 2014-09-22 4:31 ` Michael Neuling 2014-09-22 4:31 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 06/15] cxl: Add new header for call backs and structs Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 07/15] powerpc/powerpc: Add new PCIe functions for allocating cxl interrupts Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-19 7:09 ` Gavin Shan 2014-09-19 7:09 ` Gavin Shan 2014-09-22 5:01 ` Michael Neuling 2014-09-22 5:01 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 08/15] powerpc/mm: Add new hash_page_mm() Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-29 8:50 ` Aneesh Kumar K.V 2014-09-29 8:50 ` Aneesh Kumar K.V [not found] ` <1412054407.1733.77.camel@ale.ozlabs.ibm.com> 2014-09-30 6:13 ` Michael Neuling 2014-09-30 6:13 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 09/15] powerpc/opal: Add PHB to cxl mode call Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-26 4:35 ` Anton Blanchard 2014-09-26 4:35 ` Anton Blanchard 2014-09-18 8:26 ` [PATCH 10/15] powerpc/mm: Add hooks for cxl Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-26 4:33 ` Anton Blanchard 2014-09-26 4:33 ` Anton Blanchard 2014-09-26 11:33 ` Michael Neuling [this message] 2014-09-26 11:33 ` Michael Neuling 2014-09-26 13:24 ` Anton Blanchard 2014-09-26 13:24 ` Anton Blanchard 2014-09-29 9:10 ` Aneesh Kumar K.V 2014-09-29 9:10 ` Aneesh Kumar K.V 2014-09-18 8:26 ` [PATCH 11/15] cxl: Add base builtin support Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 12/15] cxl: Driver code for powernv PCIe based cards for userspace access Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 13/15] cxl: Userspace header file Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-18 8:26 ` [PATCH 14/15] cxl: Add driver to Kbuild and Makefiles Michael Neuling 2014-09-18 8:26 ` Michael Neuling 2014-09-18 8:27 ` [PATCH 15/15] cxl: Add documentation for userspace APIs Michael Neuling 2014-09-18 8:27 ` Michael Neuling
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=1411731209.1733.51.camel@ale.ozlabs.ibm.com \ --to=mikey@neuling.org \ --cc=anton@samba.org \ --cc=arnd@arndb.de \ --cc=benh@kernel.crashing.org \ --cc=cbe-oss-dev@lists.ozlabs.org \ --cc=greg@kroah.com \ --cc=imunsie@au.ibm.com \ --cc=jk@ozlabs.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linuxppc-dev@ozlabs.org \ --cc=mpe@ellerman.id.au \ /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.