Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S936393Ab3DRQi4 (ORCPT ); Thu, 18 Apr 2013 12:38:56 -0400 Received: from co9ehsobe002.messaging.microsoft.com ([207.46.163.25]:34889 "EHLO co9outboundpool.messaging.microsoft.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S936266Ab3DRQix (ORCPT ); Thu, 18 Apr 2013 12:38:53 -0400 X-Forefront-Antispam-Report: CIP:163.181.249.109;KIP:(null);UIP:(null);IPV:NLI;H:ausb3twp02.amd.com;RD:none;EFVD:NLI X-SpamScore: -2 X-BigFish: VPS-2(z551bizbb2dI98dI9371I1432Id799hzz1f42h1fc6h1ee6h1de0h1fdah1202h1e76h1d1ah1d2ahzz8275bhz2dh668h839h947he5bhf0ah1288h12a5h12a9h12bdh137ah13b6h1441h1504h1537h153bh162dh1631h1758h1765h18e1h190ch1946h19b4h19c3h1ad9h1b0ah1889i1155h) X-WSS-ID: 0MLGM8O-02-8VN-02 X-M-MSG: Message-ID: <51702197.4090100@amd.com> Date: Thu, 18 Apr 2013 11:38:47 -0500 From: Suravee Suthikulanit User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:17.0) Gecko/20130307 Thunderbird/17.0.4 MIME-Version: 1.0 To: CC: , , Subject: Re: [PATCH 2/2 V5] iommu/amd: Add logic to decode AMD IOMMU event flag References: <1365609459-3378-1-git-send-email-suravee.suthikulpanit@amd.com> In-Reply-To: <1365609459-3378-1-git-send-email-suravee.suthikulpanit@amd.com> Content-Type: text/plain; charset="ISO-8859-1"; format=flowed Content-Transfer-Encoding: 7bit X-OriginatorOrg: amd.com Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 11425 Lines: 333 Joerg, Do you have any more concerns about this patch? Thank you, Suravee On 4/10/2013 10:57 AM, suravee.suthikulpanit@amd.com wrote: > From: Suravee Suthikulpanit > > Add logic to decode AMD IOMMU event flag based on information from AMD IOMMU specification. > This should simplify debugging IOMMU errors. Also, dump DTE information in some additional > cases. > > Example (default): The flags is now decoded. > AMD-Vi: Event logged [INVALID_DEVICE_REQUEST device=51:00.0 address=0x0000000000000000 flags:G Nx Usr I P W N-Pm Rsv Tl type(0x7)] > AMD-Vi: DTE[0..3]: 600000dfa760b403 0000000000000014 200000dfaba60c11 0000000000000000 > > Example (amd-iommu=verbose): The flags and error types are decoded > AMD-Vi: Event logged [INVALID_DEVICE_REQUEST device=51:00.0 address=0x0000000000000000 flags:G Nx Usr I P W N-Pm Rsv Tl] > AMD-Vi: Error type details: (0x7) Invalid transaction to the system management address range > AMD-Vi: DTE[0..3]: 600000dfa760b403 0000000000000014 200000dfaba60c11 0000000000000000 > > Signed-off-by: Suravee Suthikulpanit > V5: > * Keeping flag decode in 1 line (as requested by Joerg) > * Use "amd-iommu=verbose" for additional printing > V4: > * Change print out format to reduce noise > * Remove "(Note:...)" > V3: > * Move comments to end of line > * Shorten the print out to be within one line > V2: > * Fix printing format to reduce noise > * Use string table instead of switch/case > * Use pr_cont instead of printk > > --- > drivers/iommu/amd_iommu.c | 215 +++++++++++++++++++++++++++++++++++---------- > 1 file changed, 168 insertions(+), 47 deletions(-) > > diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c > index b287ca3..beb61dd 100644 > --- a/drivers/iommu/amd_iommu.c > +++ b/drivers/iommu/amd_iommu.c > @@ -601,13 +601,118 @@ static void amd_iommu_stats_init(void) > * > ****************************************************************************/ > > +struct _event_log_flags { > + u32 gn:1, /* 16 */ > + nx:1, /* 17 */ > + us:1, /* 18 */ > + i:1, /* 19 */ > + pr:1, /* 20 */ > + rw:1, /* 21 */ > + pe:1, /* 22 */ > + rz:1, /* 23 */ > + tr:1, /* 24 */ > + type:3, /* [27:25] */ > + _reserved_:20; /* Reserved */ > +}; > + > +static const char * const _type_field_encodings[] = { > + "Reserved", /* 00 */ > + "Master Abort", /* 01 */ > + "Target Abort", /* 10 */ > + "Data Error", /* 11 */ > +}; > + > +static const char * const _invalid_trnsac_desc[] = { > + "Read request or non-posted write in the interrupt " > + "addres range", /* 000 */ > + "Pretranslated transaction received from an " > + "I/O device that has I=0 or V=0 in DTE", /* 001 */ > + "Port I/O space transaction received from an " > + "I/O device that has IoCtl=00b in DTE", /* 010 */ > + "Posted write to invalid address range", /* 011 */ > + "Invalid read request or non-posted write", /* 100 */ > + "Posted write to the interrupt/EOI range from an " > + "I/O device that has IntCtl=00b in DTE", /* 101 */ > + "Posted write to a reserved interrupt address range", /* 110 */ > + "Invalid transaction to the system management " > + "address range", /* 111 */ > +}; > + > +static const char * const _invalid_trnslt_desc[] = { > + "Translation request received from an I/O device " > + "that has I=0, or has V=0, or has V=1 and " > + "TV=0 in DTE", /* 000 */ > + "Translation request in invalid address range", /* 001 */ > + "Invalid translation request", /* 010 */ > + "Reserved", /* 011 */ > + "Reserved", /* 100 */ > + "Reserved", /* 101 */ > + "Reserved", /* 110 */ > + "Reserved", /* 111 */ > +}; > + > +static void dump_detail_error(struct _event_log_flags *p, int ev_type) > +{ > + u32 err_type = p->type; > + > + pr_err("AMD-Vi: Error type details: (0x%x) ", err_type); > + if ((ev_type == EVENT_TYPE_DEV_TAB_ERR) || > + (ev_type == EVENT_TYPE_PAGE_TAB_ERR) || > + (ev_type == EVENT_TYPE_CMD_HARD_ERR)) { > + if (err_type < ARRAY_SIZE(_type_field_encodings)) { > + pr_cont("%s\n", > + _type_field_encodings[err_type]); > + } > + } else if (ev_type == EVENT_TYPE_INV_DEV_REQ) { > + if (p->tr == 0) { > + if (err_type < ARRAY_SIZE(_invalid_trnslt_desc)) > + pr_cont("%s\n", > + _invalid_trnslt_desc[err_type]); > + } else { > + if (err_type < ARRAY_SIZE(_invalid_trnsac_desc)) > + pr_cont("%s\n", > + _invalid_trnsac_desc[err_type]); > + } > + } > +} > + > +static void dump_flags(int flags, int ev_type) > +{ > + struct _event_log_flags *p = (struct _event_log_flags *) &flags; > + u32 err_type = p->type; > + > + pr_cont(" flags:%s %s %s %s %s %s %s %s %s", > + (p->gn ? "G" : "N"), > + (p->nx ? "Nx" : "Ex"), > + (p->us ? "Usr" : "Sup"), > + (p->i ? "I" : "M"), > + (p->pr ? "P" : "NP"), > + (p->rw ? "W" : "R"), > + (p->pe ? "N-Pm" : "Pm"), > + (p->rz ? "Rsv" : "Ill"), > + (p->tr ? "Tl" : "Ta")); > + > + /* Error type only needed for certain events */ > + if (!amd_iommu_verbose) { > + if ((ev_type == EVENT_TYPE_DEV_TAB_ERR) || > + (ev_type == EVENT_TYPE_PAGE_TAB_ERR) || > + (ev_type == EVENT_TYPE_CMD_HARD_ERR) || > + (ev_type == EVENT_TYPE_INV_DEV_REQ)) > + pr_cont(" type(0x%x)]\n", err_type); > + } else { > + pr_cont("]\n"); > + dump_detail_error(p, ev_type); > + } > +} > + > static void dump_dte_entry(u16 devid) > { > int i; > > + pr_err("AMD-Vi: DTE[0..3]:"); > for (i = 0; i < 4; ++i) > - pr_err("AMD-Vi: DTE[%d]: %016llx\n", i, > - amd_iommu_dev_table[devid].data[i]); > + pr_cont(" %016llx", amd_iommu_dev_table[devid].data[i]); > + pr_cont("\n"); > } > > static void dump_command(unsigned long phys_addr) > @@ -619,81 +724,96 @@ static void dump_command(unsigned long phys_addr) > pr_err("AMD-Vi: CMD[%d]: %08x\n", i, cmd->data[i]); > } > > -static void iommu_print_event(struct amd_iommu *iommu, void *__evt) > +void amd_iommu_print_event(int type, int devid, int domid, > + int flags, u64 address) > { > - int type, devid, domid, flags; > - volatile u32 *event = __evt; > - int count = 0; > - u64 address; > - > -retry: > - type = (event[1] >> EVENT_TYPE_SHIFT) & EVENT_TYPE_MASK; > - devid = (event[0] >> EVENT_DEVID_SHIFT) & EVENT_DEVID_MASK; > - domid = (event[1] >> EVENT_DOMID_SHIFT) & EVENT_DOMID_MASK; > - flags = (event[1] >> EVENT_FLAGS_SHIFT) & EVENT_FLAGS_MASK; > - address = (u64)(((u64)event[3]) << 32) | event[2]; > - > - if (type == 0) { > - /* Did we hit the erratum? */ > - if (++count == LOOP_TIMEOUT) { > - pr_err("AMD-Vi: No event written to event log\n"); > - return; > - } > - udelay(1); > - goto retry; > - } > - > - printk(KERN_ERR "AMD-Vi: Event logged ["); > + pr_err("AMD-Vi: Event logged ["); > > switch (type) { > case EVENT_TYPE_ILL_DEV: > - printk("ILLEGAL_DEV_TABLE_ENTRY device=%02x:%02x.%x " > - "address=0x%016llx flags=0x%04x]\n", > + pr_cont("ILLEGAL_DEV_TABLE_ENTRY device=%02x:%02x.%x " > + "address=0x%016llx", > PCI_BUS(devid), PCI_SLOT(devid), PCI_FUNC(devid), > - address, flags); > + address); > + dump_flags(flags, type); > dump_dte_entry(devid); > break; > case EVENT_TYPE_IO_FAULT: > - printk("IO_PAGE_FAULT device=%02x:%02x.%x " > - "domain=0x%04x address=0x%016llx flags=0x%04x]\n", > + pr_cont("IO_PAGE_FAULT device=%02x:%02x.%x " > + "domain=0x%04x address=0x%016llx", > PCI_BUS(devid), PCI_SLOT(devid), PCI_FUNC(devid), > - domid, address, flags); > + domid, address); > + dump_flags(flags, type); > + dump_dte_entry(devid); > break; > case EVENT_TYPE_DEV_TAB_ERR: > - printk("DEV_TAB_HARDWARE_ERROR device=%02x:%02x.%x " > - "address=0x%016llx flags=0x%04x]\n", > + pr_cont("DEV_TAB_HARDWARE_ERROR device=%02x:%02x.%x " > + "address=0x%016llx", > PCI_BUS(devid), PCI_SLOT(devid), PCI_FUNC(devid), > - address, flags); > + address); > + dump_flags(flags, type); > break; > case EVENT_TYPE_PAGE_TAB_ERR: > - printk("PAGE_TAB_HARDWARE_ERROR device=%02x:%02x.%x " > - "domain=0x%04x address=0x%016llx flags=0x%04x]\n", > + pr_cont("PAGE_TAB_HARDWARE_ERROR device=%02x:%02x.%x " > + "domain=0x%04x address=0x%016llx", > PCI_BUS(devid), PCI_SLOT(devid), PCI_FUNC(devid), > - domid, address, flags); > + domid, address); > + dump_flags(flags, type); > break; > case EVENT_TYPE_ILL_CMD: > - printk("ILLEGAL_COMMAND_ERROR address=0x%016llx]\n", address); > + pr_cont("ILLEGAL_COMMAND_ERROR address=0x%016llx\n", > + address); > dump_command(address); > break; > case EVENT_TYPE_CMD_HARD_ERR: > - printk("COMMAND_HARDWARE_ERROR address=0x%016llx " > - "flags=0x%04x]\n", address, flags); > + pr_cont("COMMAND_HARDWARE_ERROR address=0x%016llx", > + address); > + dump_flags(flags, type); > break; > case EVENT_TYPE_IOTLB_INV_TO: > - printk("IOTLB_INV_TIMEOUT device=%02x:%02x.%x " > + pr_cont("IOTLB_INV_TIMEOUT device=%02x:%02x.%x " > "address=0x%016llx]\n", > PCI_BUS(devid), PCI_SLOT(devid), PCI_FUNC(devid), > address); > break; > case EVENT_TYPE_INV_DEV_REQ: > - printk("INVALID_DEVICE_REQUEST device=%02x:%02x.%x " > - "address=0x%016llx flags=0x%04x]\n", > + pr_cont("INVALID_DEVICE_REQUEST device=%02x:%02x.%x " > + "address=0x%016llx", > PCI_BUS(devid), PCI_SLOT(devid), PCI_FUNC(devid), > - address, flags); > + address); > + dump_flags(flags, type); > + dump_dte_entry(devid); > break; > default: > - printk(KERN_ERR "UNKNOWN type=0x%02x]\n", type); > + pr_cont("UNKNOWN type=0x%02x\n", type); > } > +} > + > +static void iommu_handle_event(struct amd_iommu *iommu, void *__evt) > +{ > + int type, devid, domid, flags; > + u32 *event = __evt; > + int count = 0; > + u64 address; > + > +retry: > + type = (event[1] >> EVENT_TYPE_SHIFT) & EVENT_TYPE_MASK; > + devid = (event[0] >> EVENT_DEVID_SHIFT) & EVENT_DEVID_MASK; > + domid = (event[1] >> EVENT_DOMID_SHIFT) & EVENT_DOMID_MASK; > + flags = (event[1] >> EVENT_FLAGS_SHIFT) & EVENT_FLAGS_MASK; > + address = (u64)(((u64)event[3]) << 32) | event[2]; > + > + if (type == 0) { > + /* Did we hit the erratum? */ > + if (++count == LOOP_TIMEOUT) { > + pr_err("AMD-Vi: No event written to event log\n"); > + return; > + } > + udelay(1); > + goto retry; > + } > + > + amd_iommu_print_event(type, devid, domid, flags, address); > > memset(__evt, 0, 4 * sizeof(u32)); > } > @@ -709,7 +829,7 @@ static void iommu_poll_events(struct amd_iommu *iommu) > tail = readl(iommu->mmio_base + MMIO_EVT_TAIL_OFFSET); > > while (head != tail) { > - iommu_print_event(iommu, iommu->evt_buf + head); > + iommu_handle_event(iommu, iommu->evt_buf + head); > head = (head + EVENT_ENTRY_SIZE) % iommu->evt_buf_size; > } > > @@ -3268,6 +3388,7 @@ static int __init alloc_passthrough_domain(void) > > return 0; > } > + > static int amd_iommu_domain_init(struct iommu_domain *dom) > { > struct protection_domain *domain; -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/