Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1031392AbXEDQ66 (ORCPT ); Fri, 4 May 2007 12:58:58 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1031400AbXEDQ66 (ORCPT ); Fri, 4 May 2007 12:58:58 -0400 Received: from mtagate3.de.ibm.com ([195.212.29.152]:26083 "EHLO mtagate3.de.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1031393AbXEDQ6w (ORCPT ); Fri, 4 May 2007 12:58:52 -0400 Subject: Please pull git390 'for-linus' branch From: Martin Schwidefsky Reply-To: schwidefsky@de.ibm.com To: torvalds@linux-foundation.org Cc: linux-kernel@vger.kernel.org, linux-s390@vger.kernel.org Content-Type: text/plain Organization: IBM Corporation Date: Fri, 04 May 2007 18:59:15 +0200 Message-Id: <1178297955.2640.2.camel@localhost> Mime-Version: 1.0 X-Mailer: Evolution 2.10.1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 49410 Lines: 1492 Please pull from 'for-linus' branch of git://git390.osdl.marist.edu/pub/scm/linux-2.6.git for-linus to receive the following updates: Documentation/feature-removal-schedule.txt | 17 ++ arch/s390/crypto/aes_s390.c | 15 ++- arch/s390/kernel/kprobes.c | 3 +- arch/s390/kernel/setup.c | 111 +++++++++++++- arch/s390/mm/fault.c | 40 ++--- drivers/s390/block/dasd.c | 45 +++++ drivers/s390/block/dasd_eckd.c | 81 +++++++++- drivers/s390/block/dasd_fba.c | 2 +- drivers/s390/block/dasd_int.h | 2 + drivers/s390/char/tape.h | 1 + drivers/s390/char/tape_3590.c | 29 +++- drivers/s390/char/tape_3590.h | 4 + drivers/s390/char/tape_core.c | 3 +- drivers/s390/cio/qdio.c | 240 ++++++++++++++++------------ drivers/s390/cio/qdio.h | 52 +++++-- drivers/s390/net/qeth.h | 3 + drivers/s390/net/qeth_main.c | 77 +++++++++- drivers/s390/net/qeth_mpc.h | 1 + include/asm-s390/ccwdev.h | 6 +- include/asm-s390/elf.h | 7 +- include/asm-s390/kdebug.h | 18 ++- include/asm-s390/kprobes.h | 16 +-- include/asm-s390/lowcore.h | 10 +- 23 files changed, 593 insertions(+), 190 deletions(-) Christoph Hellwig (1): [S390] get rid of kprobes notifier call chain. Cornelia Huck (3): [S390] cio: Deprecate read_dev_chars() and read_conf_data{,_lpm}(). [S390] dasd: New read device characteristics and read configuration data. [S390] qeth: New read configuration data. Jan Glauber (1): [S390] aes-s390 key length. Martin Schwidefsky (2): [S390] Export uaccess as non-gpl symbol. [S390] add hardware capability support (ELF_HWCAP). Michael Holzheu (1): [S390] tape: New read configuration data. Ursula Braun (1): [S390] qdio: make qdio statistics SMP-capable diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt index 5c88ba1..1a9e600 100644 --- a/Documentation/feature-removal-schedule.txt +++ b/Documentation/feature-removal-schedule.txt @@ -314,3 +314,20 @@ Why: Code was merged, then submitter immediately disappeared leaving Who: David S. Miller --------------------------- + +What: read_dev_chars(), read_conf_data{,_lpm}() (s390 common I/O layer) +When: December 2007 +Why: These functions are a leftover from 2.4 times. They have several + problems: + - Duplication of checks that are done in the device driver's + interrupt handler + - common I/O layer can't do device specific error recovery + - device driver can't be notified for conditions happening during + execution of the function + Device drivers should issue the read device characteristics and read + configuration data ccws and do the appropriate error handling + themselves. +Who: Cornelia Huck + +--------------------------- + diff --git a/arch/s390/crypto/aes_s390.c b/arch/s390/crypto/aes_s390.c index 9163635..3660ca6 100644 --- a/arch/s390/crypto/aes_s390.c +++ b/arch/s390/crypto/aes_s390.c @@ -119,7 +119,8 @@ static struct crypto_alg aes_alg = { .cra_name = "aes", .cra_driver_name = "aes-s390", .cra_priority = CRYPT_S390_PRIORITY, - .cra_flags = CRYPTO_ALG_TYPE_CIPHER, + .cra_flags = CRYPTO_ALG_TYPE_CIPHER | + CRYPTO_ALG_NEED_FALLBACK, .cra_blocksize = AES_BLOCK_SIZE, .cra_ctxsize = sizeof(struct s390_aes_ctx), .cra_module = THIS_MODULE, @@ -206,7 +207,8 @@ static struct crypto_alg ecb_aes_alg = { .cra_name = "ecb(aes)", .cra_driver_name = "ecb-aes-s390", .cra_priority = CRYPT_S390_COMPOSITE_PRIORITY, - .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, + .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER | + CRYPTO_ALG_NEED_FALLBACK, .cra_blocksize = AES_BLOCK_SIZE, .cra_ctxsize = sizeof(struct s390_aes_ctx), .cra_type = &crypto_blkcipher_type, @@ -300,7 +302,8 @@ static struct crypto_alg cbc_aes_alg = { .cra_name = "cbc(aes)", .cra_driver_name = "cbc-aes-s390", .cra_priority = CRYPT_S390_COMPOSITE_PRIORITY, - .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, + .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER | + CRYPTO_ALG_NEED_FALLBACK, .cra_blocksize = AES_BLOCK_SIZE, .cra_ctxsize = sizeof(struct s390_aes_ctx), .cra_type = &crypto_blkcipher_type, @@ -333,10 +336,14 @@ static int __init aes_init(void) return -EOPNOTSUPP; /* z9 109 and z9 BC/EC only support 128 bit key length */ - if (keylen_flag == AES_KEYLEN_128) + if (keylen_flag == AES_KEYLEN_128) { + aes_alg.cra_u.cipher.cia_max_keysize = AES_MIN_KEY_SIZE; + ecb_aes_alg.cra_u.blkcipher.max_keysize = AES_MIN_KEY_SIZE; + cbc_aes_alg.cra_u.blkcipher.max_keysize = AES_MIN_KEY_SIZE; printk(KERN_INFO "aes_s390: hardware acceleration only available for" "128 bit keys\n"); + } ret = crypto_register_alg(&aes_alg); if (ret) diff --git a/arch/s390/kernel/kprobes.c b/arch/s390/kernel/kprobes.c index 993f353..23c61f6 100644 --- a/arch/s390/kernel/kprobes.c +++ b/arch/s390/kernel/kprobes.c @@ -516,7 +516,7 @@ out: return 1; } -static int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr) +int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr) { struct kprobe *cur = kprobe_running(); struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); @@ -603,7 +603,6 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, ret = NOTIFY_STOP; break; case DIE_TRAP: - case DIE_PAGE_FAULT: /* kprobe_running() needs smp_processor_id() */ preempt_disable(); if (kprobe_running() && diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c index 3dfd098..6bfb088 100644 --- a/arch/s390/kernel/setup.c +++ b/arch/s390/kernel/setup.c @@ -65,7 +65,7 @@ long psw_user_bits = (PSW_BASE_BITS | PSW_MASK_DAT | PSW_ASC_HOME | * User copy operations. */ struct uaccess_ops uaccess; -EXPORT_SYMBOL_GPL(uaccess); +EXPORT_SYMBOL(uaccess); /* * Machine setup.. @@ -74,6 +74,8 @@ unsigned int console_mode = 0; unsigned int console_devno = -1; unsigned int console_irq = -1; unsigned long machine_flags = 0; +unsigned long elf_hwcap = 0; +char elf_platform[ELF_PLATFORM_SIZE]; struct mem_chunk __initdata memory_chunk[MEMORY_CHUNKS]; volatile int __cpu_logical_map[NR_CPUS]; /* logical cpu to cpu address */ @@ -749,6 +751,98 @@ setup_memory(void) #endif } +static __init unsigned int stfl(void) +{ + asm volatile( + " .insn s,0xb2b10000,0(0)\n" /* stfl */ + "0:\n" + EX_TABLE(0b,0b)); + return S390_lowcore.stfl_fac_list; +} + +static __init int stfle(unsigned long long *list, int doublewords) +{ + typedef struct { unsigned long long _[doublewords]; } addrtype; + register unsigned long __nr asm("0") = doublewords - 1; + + asm volatile(".insn s,0xb2b00000,%0" /* stfle */ + : "=m" (*(addrtype *) list), "+d" (__nr) : : "cc"); + return __nr + 1; +} + +/* + * Setup hardware capabilities. + */ +static void __init setup_hwcaps(void) +{ + static const int stfl_bits[6] = { 0, 2, 7, 17, 19, 21 }; + struct cpuinfo_S390 *cpuinfo = &S390_lowcore.cpu_data; + unsigned long long facility_list_extended; + unsigned int facility_list; + int i; + + facility_list = stfl(); + /* + * The store facility list bits numbers as found in the principles + * of operation are numbered with bit 1UL<<31 as number 0 to + * bit 1UL<<0 as number 31. + * Bit 0: instructions named N3, "backported" to esa-mode + * Bit 2: z/Architecture mode is active + * Bit 7: the store-facility-list-extended facility is installed + * Bit 17: the message-security assist is installed + * Bit 19: the long-displacement facility is installed + * Bit 21: the extended-immediate facility is installed + * These get translated to: + * HWCAP_S390_ESAN3 bit 0, HWCAP_S390_ZARCH bit 1, + * HWCAP_S390_STFLE bit 2, HWCAP_S390_MSA bit 3, + * HWCAP_S390_LDISP bit 4, and HWCAP_S390_EIMM bit 5. + */ + for (i = 0; i < 6; i++) + if (facility_list & (1UL << (31 - stfl_bits[i]))) + elf_hwcap |= 1UL << i; + + /* + * Check for additional facilities with store-facility-list-extended. + * stfle stores doublewords (8 byte) with bit 1ULL<<63 as bit 0 + * and 1ULL<<0 as bit 63. Bits 0-31 contain the same information + * as stored by stfl, bits 32-xxx contain additional facilities. + * How many facility words are stored depends on the number of + * doublewords passed to the instruction. The additional facilites + * are: + * Bit 43: decimal floating point facility is installed + * translated to: + * HWCAP_S390_DFP bit 6. + */ + if ((elf_hwcap & (1UL << 2)) && + stfle(&facility_list_extended, 1) > 0) { + if (facility_list_extended & (1ULL << (64 - 43))) + elf_hwcap |= 1UL << 6; + } + + switch (cpuinfo->cpu_id.machine) { + case 0x9672: +#if !defined(CONFIG_64BIT) + default: /* Use "g5" as default for 31 bit kernels. */ +#endif + strcpy(elf_platform, "g5"); + break; + case 0x2064: + case 0x2066: +#if defined(CONFIG_64BIT) + default: /* Use "z900" as default for 64 bit kernels. */ +#endif + strcpy(elf_platform, "z900"); + break; + case 0x2084: + case 0x2086: + strcpy(elf_platform, "z990"); + break; + case 0x2094: + strcpy(elf_platform, "z9-109"); + break; + } +} + /* * Setup function called from init/main.c just after the banner * was printed. @@ -805,6 +899,11 @@ setup_arch(char **cmdline_p) smp_setup_cpu_possible_map(); /* + * Setup capabilities (ELF_HWCAP & ELF_PLATFORM). + */ + setup_hwcaps(); + + /* * Create kernel page tables and switch to virtual addressing. */ paging_init(); @@ -839,8 +938,12 @@ void print_cpu_info(struct cpuinfo_S390 *cpuinfo) static int show_cpuinfo(struct seq_file *m, void *v) { + static const char *hwcap_str[7] = { + "esan3", "zarch", "stfle", "msa", "ldisp", "eimm", "dfp" + }; struct cpuinfo_S390 *cpuinfo; unsigned long n = (unsigned long) v - 1; + int i; s390_adjust_jiffies(); preempt_disable(); @@ -850,7 +953,13 @@ static int show_cpuinfo(struct seq_file *m, void *v) "bogomips per cpu: %lu.%02lu\n", num_online_cpus(), loops_per_jiffy/(500000/HZ), (loops_per_jiffy/(5000/HZ))%100); + seq_puts(m, "features\t: "); + for (i = 0; i < 7; i++) + if (hwcap_str[i] && (elf_hwcap & (1UL << i))) + seq_printf(m, "%s ", hwcap_str[i]); + seq_puts(m, "\n"); } + if (cpu_online(n)) { #ifdef CONFIG_SMP if (smp_processor_id() == n) diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c index 2b76a87..91f705a 100644 --- a/arch/s390/mm/fault.c +++ b/arch/s390/mm/fault.c @@ -52,38 +52,24 @@ extern int sysctl_userprocess_debug; extern void die(const char *,struct pt_regs *,long); #ifdef CONFIG_KPROBES -static ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain); -int register_page_fault_notifier(struct notifier_block *nb) -{ - return atomic_notifier_chain_register(¬ify_page_fault_chain, nb); -} - -int unregister_page_fault_notifier(struct notifier_block *nb) -{ - return atomic_notifier_chain_unregister(¬ify_page_fault_chain, nb); -} - -static int __kprobes __notify_page_fault(struct pt_regs *regs, long err) -{ - struct die_args args = { .str = "page fault", - .trapnr = 14, - .signr = SIGSEGV }; - args.regs = regs; - args.err = err; - return atomic_notifier_call_chain(¬ify_page_fault_chain, - DIE_PAGE_FAULT, &args); -} - static inline int notify_page_fault(struct pt_regs *regs, long err) { - if (unlikely(kprobe_running())) - return __notify_page_fault(regs, err); - return NOTIFY_DONE; + int ret = 0; + + /* kprobe_running() needs smp_processor_id() */ + if (!user_mode(regs)) { + preempt_disable(); + if (kprobe_running() && kprobe_fault_handler(regs, 14)) + ret = 1; + preempt_enable(); + } + + return ret; } #else static inline int notify_page_fault(struct pt_regs *regs, long err) { - return NOTIFY_DONE; + return 0; } #endif @@ -319,7 +305,7 @@ do_exception(struct pt_regs *regs, unsigned long error_code, int write) int space; int si_code; - if (notify_page_fault(regs, error_code) == NOTIFY_STOP) + if (notify_page_fault(regs, error_code)) return; tsk = current; diff --git a/drivers/s390/block/dasd.c b/drivers/s390/block/dasd.c index e71929d..9775210 100644 --- a/drivers/s390/block/dasd.c +++ b/drivers/s390/block/dasd.c @@ -2174,6 +2174,51 @@ dasd_generic_notify(struct ccw_device *cdev, int event) return ret; } +struct dasd_ccw_req * dasd_generic_build_rdc(struct dasd_device *device, + void *rdc_buffer, + int rdc_buffer_size, char *magic) +{ + struct dasd_ccw_req *cqr; + struct ccw1 *ccw; + + cqr = dasd_smalloc_request(magic, 1 /* RDC */, rdc_buffer_size, device); + + if (IS_ERR(cqr)) { + DEV_MESSAGE(KERN_WARNING, device, "%s", + "Could not allocate RDC request"); + return cqr; + } + + ccw = cqr->cpaddr; + ccw->cmd_code = CCW_CMD_RDC; + ccw->cda = (__u32)(addr_t)rdc_buffer; + ccw->count = rdc_buffer_size; + + cqr->device = device; + cqr->expires = 10*HZ; + clear_bit(DASD_CQR_FLAGS_USE_ERP, &cqr->flags); + cqr->retries = 2; + cqr->buildclk = get_clock(); + cqr->status = DASD_CQR_FILLED; + return cqr; +} + + +int dasd_generic_read_dev_chars(struct dasd_device *device, char *magic, + void **rdc_buffer, int rdc_buffer_size) +{ + int ret; + struct dasd_ccw_req *cqr; + + cqr = dasd_generic_build_rdc(device, *rdc_buffer, rdc_buffer_size, + magic); + if (IS_ERR(cqr)) + return PTR_ERR(cqr); + + ret = dasd_sleep_on(cqr); + dasd_sfree_request(cqr, cqr->device); + return ret; +} static int __init dasd_init(void) diff --git a/drivers/s390/block/dasd_eckd.c b/drivers/s390/block/dasd_eckd.c index cecab22..c9583fb 100644 --- a/drivers/s390/block/dasd_eckd.c +++ b/drivers/s390/block/dasd_eckd.c @@ -450,6 +450,81 @@ dasd_eckd_generate_uid(struct dasd_device *device, struct dasd_uid *uid) return 0; } +struct dasd_ccw_req * dasd_eckd_build_rcd_lpm(struct dasd_device *device, + void *rcd_buffer, + struct ciw *ciw, __u8 lpm) +{ + struct dasd_ccw_req *cqr; + struct ccw1 *ccw; + + cqr = dasd_smalloc_request("ECKD", 1 /* RCD */, ciw->count, device); + + if (IS_ERR(cqr)) { + DEV_MESSAGE(KERN_WARNING, device, "%s", + "Could not allocate RCD request"); + return cqr; + } + + ccw = cqr->cpaddr; + ccw->cmd_code = ciw->cmd; + ccw->cda = (__u32)(addr_t)rcd_buffer; + ccw->count = ciw->count; + + cqr->device = device; + cqr->expires = 10*HZ; + cqr->lpm = lpm; + clear_bit(DASD_CQR_FLAGS_USE_ERP, &cqr->flags); + cqr->retries = 2; + cqr->buildclk = get_clock(); + cqr->status = DASD_CQR_FILLED; + return cqr; +} + +static int dasd_eckd_read_conf_lpm(struct dasd_device *device, + void **rcd_buffer, + int *rcd_buffer_size, __u8 lpm) +{ + struct ciw *ciw; + char *rcd_buf = NULL; + int ret; + struct dasd_ccw_req *cqr; + + /* + * scan for RCD command in extended SenseID data + */ + ciw = ccw_device_get_ciw(device->cdev, CIW_TYPE_RCD); + if (!ciw || ciw->cmd == 0) { + ret = -EOPNOTSUPP; + goto out_error; + } + rcd_buf = kzalloc(ciw->count, GFP_KERNEL | GFP_DMA); + if (!rcd_buf) { + ret = -ENOMEM; + goto out_error; + } + cqr = dasd_eckd_build_rcd_lpm(device, rcd_buf, ciw, lpm); + if (IS_ERR(cqr)) { + ret = PTR_ERR(cqr); + goto out_error; + } + ret = dasd_sleep_on(cqr); + /* + * on success we update the user input parms + */ + dasd_sfree_request(cqr, cqr->device); + if (ret) + goto out_error; + + *rcd_buffer_size = ciw->count; + *rcd_buffer = rcd_buf; + return 0; +out_error: + kfree(rcd_buf); + *rcd_buffer = NULL; + *rcd_buffer_size = 0; + return ret; +} + static int dasd_eckd_read_conf(struct dasd_device *device) { @@ -469,8 +544,8 @@ dasd_eckd_read_conf(struct dasd_device *device) /* get configuration data per operational path */ for (lpm = 0x80; lpm; lpm>>= 1) { if (lpm & path_data->opm){ - rc = read_conf_data_lpm(device->cdev, &conf_data, - &conf_len, lpm); + rc = dasd_eckd_read_conf_lpm(device, &conf_data, + &conf_len, lpm); if (rc && rc != -EOPNOTSUPP) { /* -EOPNOTSUPP is ok */ MESSAGE(KERN_WARNING, "Read configuration data returned " @@ -639,7 +714,7 @@ dasd_eckd_check_characteristics(struct dasd_device *device) /* Read Device Characteristics */ rdc_data = (void *) &(private->rdc_data); memset(rdc_data, 0, sizeof(rdc_data)); - rc = read_dev_chars(device->cdev, &rdc_data, 64); + rc = dasd_generic_read_dev_chars(device, "ECKD", &rdc_data, 64); if (rc) DEV_MESSAGE(KERN_WARNING, device, "Read device characteristics returned " diff --git a/drivers/s390/block/dasd_fba.c b/drivers/s390/block/dasd_fba.c index be0909e..da16ead 100644 --- a/drivers/s390/block/dasd_fba.c +++ b/drivers/s390/block/dasd_fba.c @@ -135,7 +135,7 @@ dasd_fba_check_characteristics(struct dasd_device *device) } /* Read Device Characteristics */ rdc_data = (void *) &(private->rdc_data); - rc = read_dev_chars(device->cdev, &rdc_data, 32); + rc = dasd_generic_read_dev_chars(device, "FBA ", &rdc_data, 32); if (rc) { DEV_MESSAGE(KERN_WARNING, device, "Read device characteristics returned error %d", diff --git a/drivers/s390/block/dasd_int.h b/drivers/s390/block/dasd_int.h index a2cc69e..241294c 100644 --- a/drivers/s390/block/dasd_int.h +++ b/drivers/s390/block/dasd_int.h @@ -509,6 +509,8 @@ int dasd_generic_set_online(struct ccw_device *, struct dasd_discipline *); int dasd_generic_set_offline (struct ccw_device *cdev); int dasd_generic_notify(struct ccw_device *, int); +int dasd_generic_read_dev_chars(struct dasd_device *, char *, void **, int); + /* externals in dasd_devmap.c */ extern int dasd_max_devindex; extern int dasd_probeonly; diff --git a/drivers/s390/char/tape.h b/drivers/s390/char/tape.h index bb4ff53..3b52f5c 100644 --- a/drivers/s390/char/tape.h +++ b/drivers/s390/char/tape.h @@ -103,6 +103,7 @@ enum tape_op { TO_CRYPT_OFF, /* Disable encrpytion */ TO_KEKL_SET, /* Set KEK label */ TO_KEKL_QUERY, /* Query KEK label */ + TO_RDC, /* Read device characteristics */ TO_SIZE, /* #entries in tape_op_t */ }; diff --git a/drivers/s390/char/tape_3590.c b/drivers/s390/char/tape_3590.c index 50f5eda..7e2b2ab 100644 --- a/drivers/s390/char/tape_3590.c +++ b/drivers/s390/char/tape_3590.c @@ -788,6 +788,7 @@ tape_3590_done(struct tape_device *device, struct tape_request *request) case TO_SIZE: case TO_KEKL_SET: case TO_KEKL_QUERY: + case TO_RDC: break; } return TAPE_IO_SUCCESS; @@ -1549,6 +1550,26 @@ tape_3590_irq(struct tape_device *device, struct tape_request *request, return TAPE_IO_STOP; } + +static int tape_3590_read_dev_chars(struct tape_device *device, + struct tape_3590_rdc_data *rdc_data) +{ + int rc; + struct tape_request *request; + + request = tape_alloc_request(1, sizeof(*rdc_data)); + if (IS_ERR(request)) + return PTR_ERR(request); + request->op = TO_RDC; + tape_ccw_end(request->cpaddr, CCW_CMD_RDC, sizeof(*rdc_data), + request->cpdata); + rc = tape_do_io(device, request); + if (rc == 0) + memcpy(rdc_data, request->cpdata, sizeof(*rdc_data)); + tape_free_request(request); + return rc; +} + /* * Setup device function */ @@ -1557,7 +1578,7 @@ tape_3590_setup_device(struct tape_device *device) { int rc; struct tape_3590_disc_data *data; - char *rdc_data; + struct tape_3590_rdc_data *rdc_data; DBF_EVENT(6, "3590 device setup\n"); data = kzalloc(sizeof(struct tape_3590_disc_data), GFP_KERNEL | GFP_DMA); @@ -1566,12 +1587,12 @@ tape_3590_setup_device(struct tape_device *device) data->read_back_op = READ_PREVIOUS; device->discdata = data; - rdc_data = kmalloc(64, GFP_KERNEL | GFP_DMA); + rdc_data = kmalloc(sizeof(*rdc_data), GFP_KERNEL | GFP_DMA); if (!rdc_data) { rc = -ENOMEM; goto fail_kmalloc; } - rc = read_dev_chars(device->cdev, (void**)&rdc_data, 64); + rc = tape_3590_read_dev_chars(device, rdc_data); if (rc) { DBF_LH(3, "Read device characteristics failed!\n"); goto fail_kmalloc; @@ -1579,7 +1600,7 @@ tape_3590_setup_device(struct tape_device *device) rc = tape_std_assign(device); if (rc) goto fail_rdc_data; - if (rdc_data[31] == 0x13) { + if (rdc_data->data[31] == 0x13) { PRINT_INFO("Device has crypto support\n"); data->crypt_info.capability |= TAPE390_CRYPT_SUPPORTED_MASK; tape_3592_disable_crypt(device); diff --git a/drivers/s390/char/tape_3590.h b/drivers/s390/char/tape_3590.h index aa51388..4534055 100644 --- a/drivers/s390/char/tape_3590.h +++ b/drivers/s390/char/tape_3590.h @@ -129,6 +129,10 @@ struct tape_3590_med_sense { char pad2[116]; } __attribute__ ((packed)); +struct tape_3590_rdc_data { + char data[64]; +} __attribute__ ((packed)); + /* Datastructures for 3592 encryption support */ struct tape3592_kekl { diff --git a/drivers/s390/char/tape_core.c b/drivers/s390/char/tape_core.c index e2a8a1a..2fae633 100644 --- a/drivers/s390/char/tape_core.c +++ b/drivers/s390/char/tape_core.c @@ -73,7 +73,7 @@ const char *tape_op_verbose[TO_SIZE] = [TO_DIS] = "DIS", [TO_ASSIGN] = "ASS", [TO_UNASSIGN] = "UAS", [TO_CRYPT_ON] = "CON", [TO_CRYPT_OFF] = "COF", [TO_KEKL_SET] = "KLS", - [TO_KEKL_QUERY] = "KLQ", + [TO_KEKL_QUERY] = "KLQ",[TO_RDC] = "RDC", }; static int @@ -911,6 +911,7 @@ __tape_start_request(struct tape_device *device, struct tape_request *request) case TO_ASSIGN: case TO_UNASSIGN: case TO_READ_ATTMSG: + case TO_RDC: if (device->tape_state == TS_INIT) break; if (device->tape_state == TS_UNUSED) diff --git a/drivers/s390/cio/qdio.c b/drivers/s390/cio/qdio.c index 05fac07..cba64e4 100644 --- a/drivers/s390/cio/qdio.c +++ b/drivers/s390/cio/qdio.c @@ -69,7 +69,6 @@ static const char version[] = "QDIO base support version 2"; static int qdio_performance_stats = 0; static int proc_perf_file_registration; -static unsigned long i_p_c, i_p_nc, o_p_c, o_p_nc, ii_p_c, ii_p_nc; static struct qdio_perf_stats perf_stats; static int hydra_thinints; @@ -111,6 +110,31 @@ qdio_min(int a,int b) } /***************** SCRUBBER HELPER ROUTINES **********************/ +#ifdef CONFIG_64BIT +static inline void qdio_perf_stat_inc(atomic64_t *count) +{ + if (qdio_performance_stats) + atomic64_inc(count); +} + +static inline void qdio_perf_stat_dec(atomic64_t *count) +{ + if (qdio_performance_stats) + atomic64_dec(count); +} +#else /* CONFIG_64BIT */ +static inline void qdio_perf_stat_inc(atomic_t *count) +{ + if (qdio_performance_stats) + atomic_inc(count); +} + +static inline void qdio_perf_stat_dec(atomic_t *count) +{ + if (qdio_performance_stats) + atomic_dec(count); +} +#endif /* CONFIG_64BIT */ static inline __u64 qdio_get_micros(void) @@ -277,8 +301,7 @@ qdio_siga_sync(struct qdio_q *q, unsigned int gpr2, QDIO_DBF_TEXT4(0,trace,"sigasync"); QDIO_DBF_HEX4(0,trace,&q,sizeof(void*)); - if (qdio_performance_stats) - perf_stats.siga_syncs++; + qdio_perf_stat_inc(&perf_stats.siga_syncs); cc = do_siga_sync(q->schid, gpr2, gpr3); if (cc) @@ -323,8 +346,7 @@ qdio_siga_output(struct qdio_q *q) __u32 busy_bit; __u64 start_time=0; - if (qdio_performance_stats) - perf_stats.siga_outs++; + qdio_perf_stat_inc(&perf_stats.siga_outs); QDIO_DBF_TEXT4(0,trace,"sigaout"); QDIO_DBF_HEX4(0,trace,&q,sizeof(void*)); @@ -358,8 +380,7 @@ qdio_siga_input(struct qdio_q *q) QDIO_DBF_TEXT4(0,trace,"sigain"); QDIO_DBF_HEX4(0,trace,&q,sizeof(void*)); - if (qdio_performance_stats) - perf_stats.siga_ins++; + qdio_perf_stat_inc(&perf_stats.siga_ins); cc = do_siga_input(q->schid, q->mask); @@ -953,8 +974,7 @@ __qdio_outbound_processing(struct qdio_q *q) if (unlikely(qdio_reserve_q(q))) { qdio_release_q(q); - if (qdio_performance_stats) - o_p_c++; + qdio_perf_stat_inc(&perf_stats.outbound_tl_runs_resched); /* as we're sissies, we'll check next time */ if (likely(!atomic_read(&q->is_in_shutdown))) { qdio_mark_q(q); @@ -962,10 +982,8 @@ __qdio_outbound_processing(struct qdio_q *q) } return; } - if (qdio_performance_stats) { - o_p_nc++; - perf_stats.tl_runs++; - } + qdio_perf_stat_inc(&perf_stats.outbound_tl_runs); + qdio_perf_stat_inc(&perf_stats.tl_runs); /* see comment in qdio_kick_outbound_q */ siga_attempts=atomic_read(&q->busy_siga_counter); @@ -1139,17 +1157,6 @@ qdio_has_inbound_q_moved(struct qdio_q *q) { int i; - static int old_pcis=0; - static int old_thinints=0; - - if (qdio_performance_stats) { - if ((old_pcis==perf_stats.pcis)&& - (old_thinints==perf_stats.thinints)) - perf_stats.start_time_inbound=NOW; - else - old_pcis=perf_stats.pcis; - } - i=qdio_get_inbound_buffer_frontier(q); if ( (i!=GET_SAVED_FRONTIER(q)) || (q->error_status_flags&QDIO_STATUS_LOOK_FOR_ERROR) ) { @@ -1337,10 +1344,7 @@ qdio_kick_inbound_handler(struct qdio_q *q) q->siga_error=0; q->error_status_flags=0; - if (qdio_performance_stats) { - perf_stats.inbound_time+=NOW-perf_stats.start_time_inbound; - perf_stats.inbound_cnt++; - } + qdio_perf_stat_inc(&perf_stats.inbound_cnt); } static void @@ -1360,8 +1364,7 @@ __tiqdio_inbound_processing(struct qdio_q *q, int spare_ind_was_set) */ if (unlikely(qdio_reserve_q(q))) { qdio_release_q(q); - if (qdio_performance_stats) - ii_p_c++; + qdio_perf_stat_inc(&perf_stats.inbound_thin_tl_runs_resched); /* * as we might just be about to stop polling, we make * sure that we check again at least once more @@ -1369,8 +1372,7 @@ __tiqdio_inbound_processing(struct qdio_q *q, int spare_ind_was_set) tiqdio_sched_tl(); return; } - if (qdio_performance_stats) - ii_p_nc++; + qdio_perf_stat_inc(&perf_stats.inbound_thin_tl_runs); if (unlikely(atomic_read(&q->is_in_shutdown))) { qdio_unmark_q(q); goto out; @@ -1412,8 +1414,7 @@ __tiqdio_inbound_processing(struct qdio_q *q, int spare_ind_was_set) for (i=0;ino_output_qs;i++) { oq = irq_ptr->output_qs[i]; if (!qdio_is_outbound_q_done(oq)) { - if (qdio_performance_stats) - perf_stats.tl_runs--; + qdio_perf_stat_dec(&perf_stats.tl_runs); __qdio_outbound_processing(oq); } } @@ -1452,8 +1453,7 @@ __qdio_inbound_processing(struct qdio_q *q) if (unlikely(qdio_reserve_q(q))) { qdio_release_q(q); - if (qdio_performance_stats) - i_p_c++; + qdio_perf_stat_inc(&perf_stats.inbound_tl_runs_resched); /* as we're sissies, we'll check next time */ if (likely(!atomic_read(&q->is_in_shutdown))) { qdio_mark_q(q); @@ -1461,10 +1461,8 @@ __qdio_inbound_processing(struct qdio_q *q) } return; } - if (qdio_performance_stats) { - i_p_nc++; - perf_stats.tl_runs++; - } + qdio_perf_stat_inc(&perf_stats.inbound_tl_runs); + qdio_perf_stat_inc(&perf_stats.tl_runs); again: if (qdio_has_inbound_q_moved(q)) { @@ -1510,8 +1508,7 @@ tiqdio_reset_processing_state(struct qdio_q *q, int q_laps) if (unlikely(qdio_reserve_q(q))) { qdio_release_q(q); - if (qdio_performance_stats) - ii_p_c++; + qdio_perf_stat_inc(&perf_stats.inbound_thin_tl_runs_resched); /* * as we might just be about to stop polling, we make * sure that we check again at least once more @@ -1602,8 +1599,7 @@ tiqdio_tl(unsigned long data) { QDIO_DBF_TEXT4(0,trace,"iqdio_tl"); - if (qdio_performance_stats) - perf_stats.tl_runs++; + qdio_perf_stat_inc(&perf_stats.tl_runs); tiqdio_inbound_checks(); } @@ -1914,10 +1910,7 @@ tiqdio_thinint_handler(void) { QDIO_DBF_TEXT4(0,trace,"thin_int"); - if (qdio_performance_stats) { - perf_stats.thinints++; - perf_stats.start_time_inbound=NOW; - } + qdio_perf_stat_inc(&perf_stats.thinints); /* SVS only when needed: * issue SVS to benefit from iqdio interrupt avoidance @@ -1972,17 +1965,12 @@ qdio_handle_pci(struct qdio_irq *irq_ptr) int i; struct qdio_q *q; - if (qdio_performance_stats) { - perf_stats.pcis++; - perf_stats.start_time_inbound=NOW; - } + qdio_perf_stat_inc(&perf_stats.pcis); for (i=0;ino_input_qs;i++) { q=irq_ptr->input_qs[i]; if (q->is_input_q&QDIO_FLAG_NO_INPUT_INTERRUPT_CONTEXT) qdio_mark_q(q); else { - if (qdio_performance_stats) - perf_stats.tl_runs--; __qdio_inbound_processing(q); } } @@ -1992,8 +1980,7 @@ qdio_handle_pci(struct qdio_irq *irq_ptr) q=irq_ptr->output_qs[i]; if (qdio_is_outbound_q_done(q)) continue; - if (qdio_performance_stats) - perf_stats.tl_runs--; + qdio_perf_stat_dec(&perf_stats.tl_runs); if (!irq_ptr->sync_done_on_outb_pcis) SYNC_MEMORY; __qdio_outbound_processing(q); @@ -3463,18 +3450,12 @@ do_qdio_handle_outbound(struct qdio_q *q, unsigned int callflags, struct qdio_irq *irq = (struct qdio_irq *) q->irq_ptr; /* This is the outbound handling of queues */ - if (qdio_performance_stats) - perf_stats.start_time_outbound=NOW; - qdio_do_qdio_fill_output(q,qidx,count,buffers); used_elements=atomic_add_return(count, &q->number_of_buffers_used) - count; if (callflags&QDIO_FLAG_DONT_SIGA) { - if (qdio_performance_stats) { - perf_stats.outbound_time+=NOW-perf_stats.start_time_outbound; - perf_stats.outbound_cnt++; - } + qdio_perf_stat_inc(&perf_stats.outbound_cnt); return; } if (q->is_iqdio_q) { @@ -3504,8 +3485,7 @@ do_qdio_handle_outbound(struct qdio_q *q, unsigned int callflags, qdio_kick_outbound_q(q); } else { QDIO_DBF_TEXT3(0,trace, "fast-req"); - if (qdio_performance_stats) - perf_stats.fast_reqs++; + qdio_perf_stat_inc(&perf_stats.fast_reqs); } } /* @@ -3516,10 +3496,7 @@ do_qdio_handle_outbound(struct qdio_q *q, unsigned int callflags, __qdio_outbound_processing(q); } - if (qdio_performance_stats) { - perf_stats.outbound_time+=NOW-perf_stats.start_time_outbound; - perf_stats.outbound_cnt++; - } + qdio_perf_stat_inc(&perf_stats.outbound_cnt); } /* count must be 1 in iqdio */ @@ -3589,33 +3566,67 @@ qdio_perf_procfile_read(char *buffer, char **buffer_location, off_t offset, return 0; #define _OUTP_IT(x...) c+=sprintf(buffer+c,x) - _OUTP_IT("i_p_nc/c=%lu/%lu\n",i_p_nc,i_p_c); - _OUTP_IT("ii_p_nc/c=%lu/%lu\n",ii_p_nc,ii_p_c); - _OUTP_IT("o_p_nc/c=%lu/%lu\n",o_p_nc,o_p_c); - _OUTP_IT("Number of tasklet runs (total) : %lu\n", - perf_stats.tl_runs); +#ifdef CONFIG_64BIT + _OUTP_IT("Number of tasklet runs (total) : %li\n", + (long)atomic64_read(&perf_stats.tl_runs)); + _OUTP_IT("Inbound tasklet runs tried/retried : %li/%li\n", + (long)atomic64_read(&perf_stats.inbound_tl_runs), + (long)atomic64_read(&perf_stats.inbound_tl_runs_resched)); + _OUTP_IT("Inbound-thin tasklet runs tried/retried : %li/%li\n", + (long)atomic64_read(&perf_stats.inbound_thin_tl_runs), + (long)atomic64_read(&perf_stats.inbound_thin_tl_runs_resched)); + _OUTP_IT("Outbound tasklet runs tried/retried : %li/%li\n", + (long)atomic64_read(&perf_stats.outbound_tl_runs), + (long)atomic64_read(&perf_stats.outbound_tl_runs_resched)); _OUTP_IT("\n"); - _OUTP_IT("Number of SIGA sync's issued : %lu\n", - perf_stats.siga_syncs); - _OUTP_IT("Number of SIGA in's issued : %lu\n", - perf_stats.siga_ins); - _OUTP_IT("Number of SIGA out's issued : %lu\n", - perf_stats.siga_outs); - _OUTP_IT("Number of PCIs caught : %lu\n", - perf_stats.pcis); - _OUTP_IT("Number of adapter interrupts caught : %lu\n", - perf_stats.thinints); - _OUTP_IT("Number of fast requeues (outg. SBALs w/o SIGA) : %lu\n", - perf_stats.fast_reqs); + _OUTP_IT("Number of SIGA sync's issued : %li\n", + (long)atomic64_read(&perf_stats.siga_syncs)); + _OUTP_IT("Number of SIGA in's issued : %li\n", + (long)atomic64_read(&perf_stats.siga_ins)); + _OUTP_IT("Number of SIGA out's issued : %li\n", + (long)atomic64_read(&perf_stats.siga_outs)); + _OUTP_IT("Number of PCIs caught : %li\n", + (long)atomic64_read(&perf_stats.pcis)); + _OUTP_IT("Number of adapter interrupts caught : %li\n", + (long)atomic64_read(&perf_stats.thinints)); + _OUTP_IT("Number of fast requeues (outg. SBALs w/o SIGA) : %li\n", + (long)atomic64_read(&perf_stats.fast_reqs)); _OUTP_IT("\n"); - _OUTP_IT("Total time of all inbound actions (us) incl. UL : %lu\n", - perf_stats.inbound_time); - _OUTP_IT("Number of inbound transfers : %lu\n", - perf_stats.inbound_cnt); - _OUTP_IT("Total time of all outbound do_QDIOs (us) : %lu\n", - perf_stats.outbound_time); - _OUTP_IT("Number of do_QDIOs outbound : %lu\n", - perf_stats.outbound_cnt); + _OUTP_IT("Number of inbound transfers : %li\n", + (long)atomic64_read(&perf_stats.inbound_cnt)); + _OUTP_IT("Number of do_QDIOs outbound : %li\n", + (long)atomic64_read(&perf_stats.outbound_cnt)); +#else /* CONFIG_64BIT */ + _OUTP_IT("Number of tasklet runs (total) : %i\n", + atomic_read(&perf_stats.tl_runs)); + _OUTP_IT("Inbound tasklet runs tried/retried : %i/%i\n", + atomic_read(&perf_stats.inbound_tl_runs), + atomic_read(&perf_stats.inbound_tl_runs_resched)); + _OUTP_IT("Inbound-thin tasklet runs tried/retried : %i/%i\n", + atomic_read(&perf_stats.inbound_thin_tl_runs), + atomic_read(&perf_stats.inbound_thin_tl_runs_resched)); + _OUTP_IT("Outbound tasklet runs tried/retried : %i/%i\n", + atomic_read(&perf_stats.outbound_tl_runs), + atomic_read(&perf_stats.outbound_tl_runs_resched)); + _OUTP_IT("\n"); + _OUTP_IT("Number of SIGA sync's issued : %i\n", + atomic_read(&perf_stats.siga_syncs)); + _OUTP_IT("Number of SIGA in's issued : %i\n", + atomic_read(&perf_stats.siga_ins)); + _OUTP_IT("Number of SIGA out's issued : %i\n", + atomic_read(&perf_stats.siga_outs)); + _OUTP_IT("Number of PCIs caught : %i\n", + atomic_read(&perf_stats.pcis)); + _OUTP_IT("Number of adapter interrupts caught : %i\n", + atomic_read(&perf_stats.thinints)); + _OUTP_IT("Number of fast requeues (outg. SBALs w/o SIGA) : %i\n", + atomic_read(&perf_stats.fast_reqs)); + _OUTP_IT("\n"); + _OUTP_IT("Number of inbound transfers : %i\n", + atomic_read(&perf_stats.inbound_cnt)); + _OUTP_IT("Number of do_QDIOs outbound : %i\n", + atomic_read(&perf_stats.outbound_cnt)); +#endif /* CONFIG_64BIT */ _OUTP_IT("\n"); return c; @@ -3642,8 +3653,6 @@ qdio_add_procfs_entry(void) static void qdio_remove_procfs_entry(void) { - perf_stats.tl_runs=0; - if (!proc_perf_file_registration) /* means if it went ok earlier */ remove_proc_entry(QDIO_PERF,&proc_root); } @@ -3671,13 +3680,38 @@ qdio_performance_stats_store(struct bus_type *bus, const char *buf, size_t count qdio_performance_stats = i; if (i==0) { /* reset perf. stat. info */ - i_p_nc = 0; - i_p_c = 0; - ii_p_nc = 0; - ii_p_c = 0; - o_p_nc = 0; - o_p_c = 0; - memset(&perf_stats, 0, sizeof(struct qdio_perf_stats)); +#ifdef CONFIG_64BIT + atomic64_set(&perf_stats.tl_runs, 0); + atomic64_set(&perf_stats.outbound_tl_runs, 0); + atomic64_set(&perf_stats.inbound_tl_runs, 0); + atomic64_set(&perf_stats.inbound_tl_runs_resched, 0); + atomic64_set(&perf_stats.inbound_thin_tl_runs, 0); + atomic64_set(&perf_stats.inbound_thin_tl_runs_resched, + 0); + atomic64_set(&perf_stats.siga_outs, 0); + atomic64_set(&perf_stats.siga_ins, 0); + atomic64_set(&perf_stats.siga_syncs, 0); + atomic64_set(&perf_stats.pcis, 0); + atomic64_set(&perf_stats.thinints, 0); + atomic64_set(&perf_stats.fast_reqs, 0); + atomic64_set(&perf_stats.outbound_cnt, 0); + atomic64_set(&perf_stats.inbound_cnt, 0); +#else /* CONFIG_64BIT */ + atomic_set(&perf_stats.tl_runs, 0); + atomic_set(&perf_stats.outbound_tl_runs, 0); + atomic_set(&perf_stats.inbound_tl_runs, 0); + atomic_set(&perf_stats.inbound_tl_runs_resched, 0); + atomic_set(&perf_stats.inbound_thin_tl_runs, 0); + atomic_set(&perf_stats.inbound_thin_tl_runs_resched, 0); + atomic_set(&perf_stats.siga_outs, 0); + atomic_set(&perf_stats.siga_ins, 0); + atomic_set(&perf_stats.siga_syncs, 0); + atomic_set(&perf_stats.pcis, 0); + atomic_set(&perf_stats.thinints, 0); + atomic_set(&perf_stats.fast_reqs, 0); + atomic_set(&perf_stats.outbound_cnt, 0); + atomic_set(&perf_stats.inbound_cnt, 0); +#endif /* CONFIG_64BIT */ } } else { QDIO_PRINT_WARN("QDIO performance_stats: write 0 or 1 to this file!\n"); diff --git a/drivers/s390/cio/qdio.h b/drivers/s390/cio/qdio.h index ec9af72..2895392 100644 --- a/drivers/s390/cio/qdio.h +++ b/drivers/s390/cio/qdio.h @@ -406,21 +406,43 @@ do_clear_global_summary(void) #define CHSC_FLAG_SIGA_SYNC_DONE_ON_OUTB_PCIS 0x04 struct qdio_perf_stats { - unsigned long tl_runs; - - unsigned long siga_outs; - unsigned long siga_ins; - unsigned long siga_syncs; - unsigned long pcis; - unsigned long thinints; - unsigned long fast_reqs; - - __u64 start_time_outbound; - unsigned long outbound_cnt; - unsigned long outbound_time; - __u64 start_time_inbound; - unsigned long inbound_cnt; - unsigned long inbound_time; +#ifdef CONFIG_64BIT + atomic64_t tl_runs; + atomic64_t outbound_tl_runs; + atomic64_t outbound_tl_runs_resched; + atomic64_t inbound_tl_runs; + atomic64_t inbound_tl_runs_resched; + atomic64_t inbound_thin_tl_runs; + atomic64_t inbound_thin_tl_runs_resched; + + atomic64_t siga_outs; + atomic64_t siga_ins; + atomic64_t siga_syncs; + atomic64_t pcis; + atomic64_t thinints; + atomic64_t fast_reqs; + + atomic64_t outbound_cnt; + atomic64_t inbound_cnt; +#else /* CONFIG_64BIT */ + atomic_t tl_runs; + atomic_t outbound_tl_runs; + atomic_t outbound_tl_runs_resched; + atomic_t inbound_tl_runs; + atomic_t inbound_tl_runs_resched; + atomic_t inbound_thin_tl_runs; + atomic_t inbound_thin_tl_runs_resched; + + atomic_t siga_outs; + atomic_t siga_ins; + atomic_t siga_syncs; + atomic_t pcis; + atomic_t thinints; + atomic_t fast_reqs; + + atomic_t outbound_cnt; + atomic_t inbound_cnt; +#endif /* CONFIG_64BIT */ }; /* unlikely as the later the better */ diff --git a/drivers/s390/net/qeth.h b/drivers/s390/net/qeth.h index 84b108d..b34eb82 100644 --- a/drivers/s390/net/qeth.h +++ b/drivers/s390/net/qeth.h @@ -288,6 +288,7 @@ qeth_is_ipa_enabled(struct qeth_ipa_info *ipa, enum qeth_ipa_funcs func) */ #define IF_NAME_LEN 16 #define QETH_TX_TIMEOUT 100 * HZ +#define QETH_RCD_TIMEOUT 60 * HZ #define QETH_HEADER_SIZE 32 #define MAX_PORTNO 15 #define QETH_FAKE_LL_LEN_ETH ETH_HLEN @@ -582,6 +583,8 @@ enum qeth_channel_states { CH_STATE_ACTIVATING, CH_STATE_HALTED, CH_STATE_STOPPED, + CH_STATE_RCD, + CH_STATE_RCD_DONE, }; /** * card state machine diff --git a/drivers/s390/net/qeth_main.c b/drivers/s390/net/qeth_main.c index ad7792d..6fd8870 100644 --- a/drivers/s390/net/qeth_main.c +++ b/drivers/s390/net/qeth_main.c @@ -315,7 +315,8 @@ qeth_alloc_card(void) } static long -__qeth_check_irb_error(struct ccw_device *cdev, struct irb *irb) +__qeth_check_irb_error(struct ccw_device *cdev, unsigned long intparm, + struct irb *irb) { if (!IS_ERR(irb)) return 0; @@ -330,6 +331,14 @@ __qeth_check_irb_error(struct ccw_device *cdev, struct irb *irb) PRINT_WARN("timeout on device %s\n", cdev->dev.bus_id); QETH_DBF_TEXT(trace, 2, "ckirberr"); QETH_DBF_TEXT_(trace, 2, " rc%d", -ETIMEDOUT); + if (intparm == QETH_RCD_PARM) { + struct qeth_card *card = CARD_FROM_CDEV(cdev); + + if (card && (card->data.ccwdev == cdev)) { + card->data.state = CH_STATE_DOWN; + wake_up(&card->wait_q); + } + } break; default: PRINT_WARN("unknown error %ld on device %s\n", PTR_ERR(irb), @@ -401,7 +410,7 @@ qeth_irq(struct ccw_device *cdev, unsigned long intparm, struct irb *irb) QETH_DBF_TEXT(trace,5,"irq"); - if (__qeth_check_irb_error(cdev, irb)) + if (__qeth_check_irb_error(cdev, intparm, irb)) return; cstat = irb->scsw.cstat; dstat = irb->scsw.dstat; @@ -429,7 +438,8 @@ qeth_irq(struct ccw_device *cdev, unsigned long intparm, struct irb *irb) channel->state = CH_STATE_HALTED; /*let's wake up immediately on data channel*/ - if ((channel == &card->data) && (intparm != 0)) + if ((channel == &card->data) && (intparm != 0) && + (intparm != QETH_RCD_PARM)) goto out; if (intparm == QETH_CLEAR_CHANNEL_PARM) { @@ -453,6 +463,10 @@ qeth_irq(struct ccw_device *cdev, unsigned long intparm, struct irb *irb) HEXDUMP16(WARN,"irb: ",irb); HEXDUMP16(WARN,"sense data: ",irb->ecw); } + if (intparm == QETH_RCD_PARM) { + channel->state = CH_STATE_DOWN; + goto out; + } rc = qeth_get_problem(cdev,irb); if (rc) { qeth_schedule_recovery(card); @@ -460,6 +474,10 @@ qeth_irq(struct ccw_device *cdev, unsigned long intparm, struct irb *irb) } } + if (intparm == QETH_RCD_PARM) { + channel->state = CH_STATE_RCD_DONE; + goto out; + } if (intparm) { buffer = (struct qeth_cmd_buffer *) __va((addr_t)intparm); buffer->state = BUF_STATE_PROCESSED; @@ -1204,6 +1222,54 @@ qeth_probe_device(struct ccwgroup_device *gdev) } +static int qeth_read_conf_data(struct qeth_card *card, void **buffer, + int *length) +{ + struct ciw *ciw; + char *rcd_buf; + int ret; + struct qeth_channel *channel = &card->data; + unsigned long flags; + + /* + * scan for RCD command in extended SenseID data + */ + ciw = ccw_device_get_ciw(channel->ccwdev, CIW_TYPE_RCD); + if (!ciw || ciw->cmd == 0) + return -EOPNOTSUPP; + rcd_buf = kzalloc(ciw->count, GFP_KERNEL | GFP_DMA); + if (!rcd_buf) + return -ENOMEM; + + channel->ccw.cmd_code = ciw->cmd; + channel->ccw.cda = (__u32) __pa (rcd_buf); + channel->ccw.count = ciw->count; + channel->ccw.flags = CCW_FLAG_SLI; + channel->state = CH_STATE_RCD; + spin_lock_irqsave(get_ccwdev_lock(channel->ccwdev), flags); + ret = ccw_device_start_timeout(channel->ccwdev, &channel->ccw, + QETH_RCD_PARM, LPM_ANYPATH, 0, + QETH_RCD_TIMEOUT); + spin_unlock_irqrestore(get_ccwdev_lock(channel->ccwdev), flags); + if (!ret) + wait_event(card->wait_q, + (channel->state == CH_STATE_RCD_DONE || + channel->state == CH_STATE_DOWN)); + if (channel->state == CH_STATE_DOWN) + ret = -EIO; + else + channel->state = CH_STATE_DOWN; + if (ret) { + kfree(rcd_buf); + *buffer = NULL; + *length = 0; + } else { + *length = ciw->count; + *buffer = rcd_buf; + } + return ret; +} + static int qeth_get_unitaddr(struct qeth_card *card) { @@ -1212,9 +1278,9 @@ qeth_get_unitaddr(struct qeth_card *card) int rc; QETH_DBF_TEXT(setup, 2, "getunit"); - rc = read_conf_data(CARD_DDEV(card), (void **) &prcd, &length); + rc = qeth_read_conf_data(card, (void **) &prcd, &length); if (rc) { - PRINT_ERR("read_conf_data for device %s returned %i\n", + PRINT_ERR("qeth_read_conf_data for device %s returned %i\n", CARD_DDEV_ID(card), rc); return rc; } @@ -1223,6 +1289,7 @@ qeth_get_unitaddr(struct qeth_card *card) card->info.cula = prcd[63]; card->info.guestlan = ((prcd[0x10] == _ascebc['V']) && (prcd[0x11] == _ascebc['M'])); + kfree(prcd); return 0; } diff --git a/drivers/s390/net/qeth_mpc.h b/drivers/s390/net/qeth_mpc.h index 0477c47..d74bc43 100644 --- a/drivers/s390/net/qeth_mpc.h +++ b/drivers/s390/net/qeth_mpc.h @@ -37,6 +37,7 @@ extern unsigned char IPA_PDU_HEADER[]; #define QETH_CLEAR_CHANNEL_PARM -10 #define QETH_HALT_CHANNEL_PARM -11 +#define QETH_RCD_PARM -12 /*****************************************************************************/ /* IP Assist related definitions */ diff --git a/include/asm-s390/ccwdev.h b/include/asm-s390/ccwdev.h index cfc8153..6795ece 100644 --- a/include/asm-s390/ccwdev.h +++ b/include/asm-s390/ccwdev.h @@ -164,9 +164,9 @@ extern int ccw_device_resume(struct ccw_device *); extern int ccw_device_halt(struct ccw_device *, unsigned long); extern int ccw_device_clear(struct ccw_device *, unsigned long); -extern int read_dev_chars(struct ccw_device *cdev, void **buffer, int length); -extern int read_conf_data(struct ccw_device *cdev, void **buffer, int *length); -extern int read_conf_data_lpm(struct ccw_device *cdev, void **buffer, +extern int __deprecated read_dev_chars(struct ccw_device *cdev, void **buffer, int length); +extern int __deprecated read_conf_data(struct ccw_device *cdev, void **buffer, int *length); +extern int __deprecated read_conf_data_lpm(struct ccw_device *cdev, void **buffer, int *length, __u8 lpm); extern int ccw_device_set_online(struct ccw_device *cdev); diff --git a/include/asm-s390/elf.h b/include/asm-s390/elf.h index c0d629d..91d0632 100644 --- a/include/asm-s390/elf.h +++ b/include/asm-s390/elf.h @@ -188,7 +188,8 @@ static inline int dump_task_fpu(struct task_struct *tsk, elf_fpregset_t *fpregs) /* This yields a mask that user programs can use to figure out what instruction set this CPU supports. */ -#define ELF_HWCAP (0) +extern unsigned long elf_hwcap; +#define ELF_HWCAP (elf_hwcap) /* This yields a string that ld.so will use to load implementation specific libraries for optimization. This is more specific in @@ -197,7 +198,9 @@ static inline int dump_task_fpu(struct task_struct *tsk, elf_fpregset_t *fpregs) For the moment, we have only optimizations for the Intel generations, but that could change... */ -#define ELF_PLATFORM (NULL) +#define ELF_PLATFORM_SIZE 8 +extern char elf_platform[]; +#define ELF_PLATFORM (elf_platform) #ifndef __s390x__ #define SET_PERSONALITY(ex, ibcs2) set_personality((ibcs2)?PER_SVR4:PER_LINUX) diff --git a/include/asm-s390/kdebug.h b/include/asm-s390/kdebug.h index 1b50f89..d2d7ad2 100644 --- a/include/asm-s390/kdebug.h +++ b/include/asm-s390/kdebug.h @@ -22,8 +22,21 @@ struct die_args { */ extern int register_die_notifier(struct notifier_block *); extern int unregister_die_notifier(struct notifier_block *); -extern int register_page_fault_notifier(struct notifier_block *); -extern int unregister_page_fault_notifier(struct notifier_block *); + +/* + * These are only here because kprobes.c wants them to implement a + * blatant layering violation. Will hopefully go away soon once all + * architectures are updated. + */ +static inline int register_page_fault_notifier(struct notifier_block *nb) +{ + return 0; +} +static inline int unregister_page_fault_notifier(struct notifier_block *nb) +{ + return 0; +} + extern struct atomic_notifier_head s390die_chain; enum die_val { @@ -39,7 +52,6 @@ enum die_val { DIE_GPF, DIE_CALL, DIE_NMI_IPI, - DIE_PAGE_FAULT, }; static inline int notify_die(enum die_val val, const char *str, diff --git a/include/asm-s390/kprobes.h b/include/asm-s390/kprobes.h index b847ff0..830fe4c 100644 --- a/include/asm-s390/kprobes.h +++ b/include/asm-s390/kprobes.h @@ -97,18 +97,10 @@ void kretprobe_trampoline(void); int is_prohibited_opcode(kprobe_opcode_t *instruction); void get_instruction_type(struct arch_specific_insn *ainsn); +int kprobe_fault_handler(struct pt_regs *regs, int trapnr); +int kprobe_exceptions_notify(struct notifier_block *self, + unsigned long val, void *data); + #define flush_insn_slot(p) do { } while (0) #endif /* _ASM_S390_KPROBES_H */ - -#ifdef CONFIG_KPROBES - -extern int kprobe_exceptions_notify(struct notifier_block *self, - unsigned long val, void *data); -#else /* !CONFIG_KPROBES */ -static inline int kprobe_exceptions_notify(struct notifier_block *self, - unsigned long val, void *data) -{ - return 0; -} -#endif diff --git a/include/asm-s390/lowcore.h b/include/asm-s390/lowcore.h index ffc9788..801a6fd 100644 --- a/include/asm-s390/lowcore.h +++ b/include/asm-s390/lowcore.h @@ -229,17 +229,19 @@ struct _lowcore __u16 subchannel_nr; /* 0x0ba */ __u32 io_int_parm; /* 0x0bc */ __u32 io_int_word; /* 0x0c0 */ - __u8 pad3[0xD4-0xC4]; /* 0x0c4 */ + __u8 pad3[0xc8-0xc4]; /* 0x0c4 */ + __u32 stfl_fac_list; /* 0x0c8 */ + __u8 pad4[0xd4-0xcc]; /* 0x0cc */ __u32 extended_save_area_addr; /* 0x0d4 */ __u32 cpu_timer_save_area[2]; /* 0x0d8 */ __u32 clock_comp_save_area[2]; /* 0x0e0 */ __u32 mcck_interruption_code[2]; /* 0x0e8 */ - __u8 pad4[0xf4-0xf0]; /* 0x0f0 */ + __u8 pad5[0xf4-0xf0]; /* 0x0f0 */ __u32 external_damage_code; /* 0x0f4 */ __u32 failing_storage_address; /* 0x0f8 */ - __u8 pad5[0x100-0xfc]; /* 0x0fc */ + __u8 pad6[0x100-0xfc]; /* 0x0fc */ __u32 st_status_fixed_logout[4];/* 0x100 */ - __u8 pad6[0x120-0x110]; /* 0x110 */ + __u8 pad7[0x120-0x110]; /* 0x110 */ __u32 access_regs_save_area[16];/* 0x120 */ __u32 floating_pt_save_area[8]; /* 0x160 */ __u32 gpregs_save_area[16]; /* 0x180 */ - 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/