Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752048AbdHBIED (ORCPT ); Wed, 2 Aug 2017 04:04:03 -0400 Received: from us01smtprelay-2.synopsys.com ([198.182.47.9]:55718 "EHLO smtprelay.synopsys.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751220AbdHBIEB (ORCPT ); Wed, 2 Aug 2017 04:04:01 -0400 From: Alexey Brodkin To: Vineet Gupta CC: "linux-kernel@vger.kernel.org" , "vladimir.kondratiev@intel.com" , "linux-snps-arc@lists.infradead.org" Subject: Re: [PATCH] arc: arcv2: cache: Explicitly set MSB counterpart of region ops addresses Thread-Topic: [PATCH] arc: arcv2: cache: Explicitly set MSB counterpart of region ops addresses Thread-Index: AQHTCqzTK8Zq1C7d5kCTLDPlukEpCKJwSxUAgABJ0QA= Date: Wed, 2 Aug 2017 08:03:57 +0000 Message-ID: <1501661036.6345.15.camel@synopsys.com> References: <20170801095847.6425-1-abrodkin@synopsys.com> <2f6ba6bb-ae07-0ff9-c2ff-0f162eab2ef0@synopsys.com> In-Reply-To: <2f6ba6bb-ae07-0ff9-c2ff-0f162eab2ef0@synopsys.com> Accept-Language: en-US, ru-RU Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.225.15.124] Content-Type: text/plain; charset="utf-8" Content-ID: MIME-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from base64 to 8bit by nfs id v7284Be2003587 Content-Length: 4173 Lines: 105 Hi Vineet, On Wed, 2017-08-02 at 09:09 +0530, Vineet Gupta wrote: > On 08/01/2017 03:29 PM, Alexey Brodkin wrote: > > > > It is necessary to explicitly set both SLC_AUX_RGN_START1 and SLC_AUX_RGN_END1 > > which hold MSB bits of the physical address correspondingly of region start > > and end otherwise SLC region operation is executed in unpredictable manner, > > for example on HSDK platform where PAE40 support exists in hardware > > we saw each and every SLC region op to take seconds (sic!). > > > > Signed-off-by: Alexey Brodkin > > Reported-by: Vladimir Kondratiev > > --- > >   arch/arc/include/asm/cache.h | 2 ++ > >   arch/arc/mm/cache.c          | 8 ++++++-- > >   2 files changed, 8 insertions(+), 2 deletions(-) > > > > diff --git a/arch/arc/include/asm/cache.h b/arch/arc/include/asm/cache.h > > index 35127ad95124..1f3c2f967471 100644 > > --- a/arch/arc/include/asm/cache.h > > +++ b/arch/arc/include/asm/cache.h > > @@ -91,7 +91,9 @@ extern unsigned long perip_base, perip_end; > >   #define ARC_REG_SLC_FLUSH 0x904 > >   #define ARC_REG_SLC_INVALIDATE 0x905 > >   #define ARC_REG_SLC_RGN_START 0x914 > > +#define ARC_REG_SLC_RGN_START1 0x915 > >   #define ARC_REG_SLC_RGN_END 0x916 > > +#define ARC_REG_SLC_RGN_END1 0x917 > >    > >   /* Bit val in SLC_CONTROL */ > >   #define SLC_CTRL_DIS 0x001 > > diff --git a/arch/arc/mm/cache.c b/arch/arc/mm/cache.c > > index b7a1face1584..0b4e2650c5de 100644 > > --- a/arch/arc/mm/cache.c > > +++ b/arch/arc/mm/cache.c > > @@ -580,6 +580,7 @@ noinline void slc_op(phys_addr_t paddr, unsigned long sz, const int op) > >    static DEFINE_SPINLOCK(lock); > >    unsigned long flags; > >    unsigned int ctrl; > > + phys_addr_t end; > >    > >    spin_lock_irqsave(&lock, flags); > >    > > @@ -609,8 +610,11 @@ noinline void slc_op(phys_addr_t paddr, unsigned long sz, const int op) > >     * END needs to be setup before START (latter triggers the operation) > >     * END can't be same as START, so add (l2_line_sz - 1) to sz > >     */ > > - write_aux_reg(ARC_REG_SLC_RGN_END, (paddr + sz + l2_line_sz - 1)); > > - write_aux_reg(ARC_REG_SLC_RGN_START, paddr); > > + end = paddr + sz + l2_line_sz - 1; > > + write_aux_reg(ARC_REG_SLC_RGN_END1, upper_32_bits(end)); > > + write_aux_reg(ARC_REG_SLC_RGN_END, lower_32_bits(end)); > > + write_aux_reg(ARC_REG_SLC_RGN_START1, upper_32_bits(paddr)); > > + write_aux_reg(ARC_REG_SLC_RGN_START, lower_32_bits(paddr)); > > Are these registers present even if PAE is not configured in hardware ? Apparently no :( I did test it on AXS103 but just now realized its latest firmware has PAE40. Retested on older firmware (still with HS 2.1) without PAE and indeed those AUX registers are not implemented: ----------------------->8---------------------- ehci-platform e0040000.ehci: new USB bus registered, assigned bus number 1 Path: (null) CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.13.0-rc1-next-20170718-00001-g6f0be43cd25b-dirty #1 task: 9f02ba80 task.stack: 9f02c000 [ECR   ]: 0x00020000 => Illegal Insn [EFA   ]: 0x8020bba2 [BLINK ]: slc_op+0x1a/0xac [ERET  ]: slc_op+0x4a/0xac [STAT32]: 0x00080802 :   K      BTA: 0x8020bb80  SP: 0x9f02dd88  FP: 0x00000000 LPS: 0x80670da4 LPE: 0x80670db8 LPC: 0x00000000 r00: 0x00000031 r01: 0x00002000 r02: 0x9f38403f r03: 0x00000917 r04: 0x00000031 r05: 0x9f38a5c0 r06: 0x9f0baf0d r07: 0x9f0baf0c r08: 0x00000000 r09: 0x00000000 r10: 0x00000000 r11: 0x80808080 r12: 0x8020bb72 Stack Trace:   slc_op+0x4a/0xac   arc_dma_alloc+0x7c/0xd8   dma_pool_alloc+0x186/0x1d0   ehci_qh_alloc+0x34/0xd4   ehci_setup+0x15c/0x420   ehci_platform_reset+0x48/0x68   usb_add_hcd+0x186/0x624   ehci_platform_probe+0x210/0x514   platform_drv_probe+0x26/0x64   really_probe+0x284/0x348   __driver_attach+0xac/0xd4   bus_for_each_dev+0x38/0x70   bus_add_driver+0xc0/0x180   driver_register+0x50/0xec   do_one_initcall+0x32/0x118   kernel_init_freeable+0x108/0x198 ----------------------->8---------------------- So I'll respin this patch with conditional setup of those regs depending on PAE40 presence. -Alexey