diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2012-05-29 18:32:37 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2012-05-29 18:32:37 -0700 |
commit | 87a5af24e54857e7b15c1f1b0468512ee65c94e3 (patch) | |
tree | ee346852a0fc27f172a5eb57b6e3c7bf111f2fad /drivers/edac/i7core_edac.c | |
parent | 7e5b2db77b05746613516599c916a8cc2e321077 (diff) | |
parent | 0bf09e829dd4b07227ed5a8bc4ac85752a044458 (diff) |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-edac
Pull EDAC internal API changes from Mauro Carvalho Chehab:
"This changeset is the first part of a series of patches that fixes the
EDAC sybsystem. On this set, it changes the Kernel EDAC API in order
to properly represent the Intel i3/i5/i7, Xeon 3xxx/5xxx/7xxx, and
Intel E5-xxxx memory controllers.
The EDAC core used to assume that:
- the DRAM chip select pin is directly accessed by the memory
controller
- when multiple channels are used, they're all filled with the
same type of memory.
None of the above premises is true on Intel memory controllers since
2002, when RAMBUS and FB-DIMMs were introduced, and Advanced Memory
Buffer or by some similar technologies hides the direct access to the
DRAM pins.
So, the existing drivers for those chipsets had to lie to the EDAC
core, in general telling that just one channel is filled. That
produces some hard to understand error messages like:
EDAC MC0: CE row 3, channel 0, label "DIMM1": 1 Unknown error(s): memory read error on FATAL area : cpu=0 Err=0008:00c2 (ch=2), addr = 0xad1f73480 => socket=0, Channel=0(mask=2), rank=1
The location information there (row3 channel 0) is completely bogus:
it has no physical meaning, and are just some random values that the
driver uses to talk with the EDAC core. The error actually happened
at CPU socket 0, channel 0, slot 1, but this is not reported anywhere,
as the EDAC core doesn't know anything about the memory layout. So,
only advanced users that know how the EDAC driver works and that tests
their systems to see how DIMMs are mapped can actually benefit for
such error logs.
This patch series fixes the error report logic, in order to allow the
EDAC to expose the memory architecture used by them to the EDAC core.
So, as the EDAC core now understands how the memory is organized, it
can provide an useful report:
EDAC MC0: CE memory read error on DIMM1 (channel:0 slot:1 page:0x364b1b offset:0x600 grain:32 syndrome:0x0 - count:1 area:DRAM err_code:0001:0090 socket:0 channel_mask:1 rank:4)
The location of the DIMM where the error happened is reported by "MC0"
(cpu socket #0), at "channel:0 slot:1" location, and matches the
physical location of the DIMM.
There are two remaining issues not covered by this patch series:
- The EDAC sysfs API will still report bogus values. So,
userspace tools like edac-utils will still use the bogus data;
- Add a new tracepoint-based way to get the binary information
about the errors.
Those are on a second series of patches (also at -next), but will
probably miss the train for 3.5, due to the slow review process."
Fix up trivial conflict (due to spelling correction of removed code) in
drivers/edac/edac_device.c
* git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-edac: (42 commits)
i7core: fix ranks information at the per-channel struct
i5000: Fix the fatal error handling
i5100_edac: Fix a warning when compiled with 32 bits
i82975x_edac: Test nr_pages earlier to save a few CPU cycles
e752x_edac: provide more info about how DIMMS/ranks are mapped
i5000_edac: Fix the logic that retrieves memory information
i5400_edac: improve debug messages to better represent the filled memory
edac: Cleanup the logs for i7core and sb edac drivers
edac: Initialize the dimm label with the known information
edac: Remove the legacy EDAC ABI
x38_edac: convert driver to use the new edac ABI
tile_edac: convert driver to use the new edac ABI
sb_edac: convert driver to use the new edac ABI
r82600_edac: convert driver to use the new edac ABI
ppc4xx_edac: convert driver to use the new edac ABI
pasemi_edac: convert driver to use the new edac ABI
mv64x60_edac: convert driver to use the new edac ABI
mpc85xx_edac: convert driver to use the new edac ABI
i82975x_edac: convert driver to use the new edac ABI
i82875p_edac: convert driver to use the new edac ABI
...
Diffstat (limited to 'drivers/edac/i7core_edac.c')
-rw-r--r-- | drivers/edac/i7core_edac.c | 270 |
1 files changed, 78 insertions, 192 deletions
diff --git a/drivers/edac/i7core_edac.c b/drivers/edac/i7core_edac.c index 7f1dfcc4e597..d27778f65a5d 100644 --- a/drivers/edac/i7core_edac.c +++ b/drivers/edac/i7core_edac.c @@ -221,7 +221,9 @@ struct i7core_inject { }; struct i7core_channel { - u32 ranks; + bool is_3dimms_present; + bool is_single_4rank; + bool has_4rank; u32 dimms; }; @@ -257,7 +259,6 @@ struct i7core_pvt { struct i7core_channel channel[NUM_CHANS]; int ce_count_available; - int csrow_map[NUM_CHANS][MAX_DIMMS]; /* ECC corrected errors counts per udimm */ unsigned long udimm_ce_count[MAX_DIMMS]; @@ -492,116 +493,15 @@ static void free_i7core_dev(struct i7core_dev *i7core_dev) /**************************************************************************** Memory check routines ****************************************************************************/ -static struct pci_dev *get_pdev_slot_func(u8 socket, unsigned slot, - unsigned func) -{ - struct i7core_dev *i7core_dev = get_i7core_dev(socket); - int i; - - if (!i7core_dev) - return NULL; - - for (i = 0; i < i7core_dev->n_devs; i++) { - if (!i7core_dev->pdev[i]) - continue; - - if (PCI_SLOT(i7core_dev->pdev[i]->devfn) == slot && - PCI_FUNC(i7core_dev->pdev[i]->devfn) == func) { - return i7core_dev->pdev[i]; - } - } - - return NULL; -} - -/** - * i7core_get_active_channels() - gets the number of channels and csrows - * @socket: Quick Path Interconnect socket - * @channels: Number of channels that will be returned - * @csrows: Number of csrows found - * - * Since EDAC core needs to know in advance the number of available channels - * and csrows, in order to allocate memory for csrows/channels, it is needed - * to run two similar steps. At the first step, implemented on this function, - * it checks the number of csrows/channels present at one socket. - * this is used in order to properly allocate the size of mci components. - * - * It should be noticed that none of the current available datasheets explain - * or even mention how csrows are seen by the memory controller. So, we need - * to add a fake description for csrows. - * So, this driver is attributing one DIMM memory for one csrow. - */ -static int i7core_get_active_channels(const u8 socket, unsigned *channels, - unsigned *csrows) -{ - struct pci_dev *pdev = NULL; - int i, j; - u32 status, control; - - *channels = 0; - *csrows = 0; - - pdev = get_pdev_slot_func(socket, 3, 0); - if (!pdev) { - i7core_printk(KERN_ERR, "Couldn't find socket %d fn 3.0!!!\n", - socket); - return -ENODEV; - } - - /* Device 3 function 0 reads */ - pci_read_config_dword(pdev, MC_STATUS, &status); - pci_read_config_dword(pdev, MC_CONTROL, &control); - - for (i = 0; i < NUM_CHANS; i++) { - u32 dimm_dod[3]; - /* Check if the channel is active */ - if (!(control & (1 << (8 + i)))) - continue; - - /* Check if the channel is disabled */ - if (status & (1 << i)) - continue; - - pdev = get_pdev_slot_func(socket, i + 4, 1); - if (!pdev) { - i7core_printk(KERN_ERR, "Couldn't find socket %d " - "fn %d.%d!!!\n", - socket, i + 4, 1); - return -ENODEV; - } - /* Devices 4-6 function 1 */ - pci_read_config_dword(pdev, - MC_DOD_CH_DIMM0, &dimm_dod[0]); - pci_read_config_dword(pdev, - MC_DOD_CH_DIMM1, &dimm_dod[1]); - pci_read_config_dword(pdev, - MC_DOD_CH_DIMM2, &dimm_dod[2]); - (*channels)++; - - for (j = 0; j < 3; j++) { - if (!DIMM_PRESENT(dimm_dod[j])) - continue; - (*csrows)++; - } - } - - debugf0("Number of active channels on socket %d: %d\n", - socket, *channels); - - return 0; -} - -static int get_dimm_config(const struct mem_ctl_info *mci) +static int get_dimm_config(struct mem_ctl_info *mci) { struct i7core_pvt *pvt = mci->pvt_info; - struct csrow_info *csr; struct pci_dev *pdev; int i, j; - int csrow = 0; - unsigned long last_page = 0; enum edac_type mode; enum mem_type mtype; + struct dimm_info *dimm; /* Get data from the MC register, function 0 */ pdev = pvt->pci_mcr[0]; @@ -657,21 +557,20 @@ static int get_dimm_config(const struct mem_ctl_info *mci) pci_read_config_dword(pvt->pci_ch[i][0], MC_CHANNEL_DIMM_INIT_PARAMS, &data); - pvt->channel[i].ranks = (data & QUAD_RANK_PRESENT) ? - 4 : 2; + + if (data & THREE_DIMMS_PRESENT) + pvt->channel[i].is_3dimms_present = true; + + if (data & SINGLE_QUAD_RANK_PRESENT) + pvt->channel[i].is_single_4rank = true; + + if (data & QUAD_RANK_PRESENT) + pvt->channel[i].has_4rank = true; if (data & REGISTERED_DIMM) mtype = MEM_RDDR3; else mtype = MEM_DDR3; -#if 0 - if (data & THREE_DIMMS_PRESENT) - pvt->channel[i].dimms = 3; - else if (data & SINGLE_QUAD_RANK_PRESENT) - pvt->channel[i].dimms = 1; - else - pvt->channel[i].dimms = 2; -#endif /* Devices 4-6 function 1 */ pci_read_config_dword(pvt->pci_ch[i][1], @@ -682,11 +581,13 @@ static int get_dimm_config(const struct mem_ctl_info *mci) MC_DOD_CH_DIMM2, &dimm_dod[2]); debugf0("Ch%d phy rd%d, wr%d (0x%08x): " - "%d ranks, %cDIMMs\n", + "%s%s%s%cDIMMs\n", i, RDLCH(pvt->info.ch_map, i), WRLCH(pvt->info.ch_map, i), data, - pvt->channel[i].ranks, + pvt->channel[i].is_3dimms_present ? "3DIMMS " : "", + pvt->channel[i].is_3dimms_present ? "SINGLE_4R " : "", + pvt->channel[i].has_4rank ? "HAS_4R " : "", (data & REGISTERED_DIMM) ? 'R' : 'U'); for (j = 0; j < 3; j++) { @@ -696,6 +597,8 @@ static int get_dimm_config(const struct mem_ctl_info *mci) if (!DIMM_PRESENT(dimm_dod[j])) continue; + dimm = EDAC_DIMM_PTR(mci->layers, mci->dimms, mci->n_layers, + i, j, 0); banks = numbank(MC_DOD_NUMBANK(dimm_dod[j])); ranks = numrank(MC_DOD_NUMRANK(dimm_dod[j])); rows = numrow(MC_DOD_NUMROW(dimm_dod[j])); @@ -704,8 +607,6 @@ static int get_dimm_config(const struct mem_ctl_info *mci) /* DDR3 has 8 I/O banks */ size = (rows * cols * banks * ranks) >> (20 - 3); - pvt->channel[i].dimms++; - debugf0("\tdimm %d %d Mb offset: %x, " "bank: %d, rank: %d, row: %#x, col: %#x\n", j, size, @@ -714,44 +615,28 @@ static int get_dimm_config(const struct mem_ctl_info *mci) npages = MiB_TO_PAGES(size); - csr = &mci->csrows[csrow]; - csr->first_page = last_page + 1; - last_page += npages; - csr->last_page = last_page; - csr->nr_pages = npages; - - csr->page_mask = 0; - csr->grain = 8; - csr->csrow_idx = csrow; - csr->nr_channels = 1; - - csr->channels[0].chan_idx = i; - csr->channels[0].ce_count = 0; - - pvt->csrow_map[i][j] = csrow; + dimm->nr_pages = npages; switch (banks) { case 4: - csr->dtype = DEV_X4; + dimm->dtype = DEV_X4; break; case 8: - csr->dtype = DEV_X8; + dimm->dtype = DEV_X8; break; case 16: - csr->dtype = DEV_X16; + dimm->dtype = DEV_X16; break; default: - csr->dtype = DEV_UNKNOWN; + dimm->dtype = DEV_UNKNOWN; } - csr->edac_mode = mode; - csr->mtype = mtype; - snprintf(csr->channels[0].label, - sizeof(csr->channels[0].label), - "CPU#%uChannel#%u_DIMM#%u", - pvt->i7core_dev->socket, i, j); - - csrow++; + snprintf(dimm->label, sizeof(dimm->label), + "CPU#%uChannel#%u_DIMM#%u", + pvt->i7core_dev->socket, i, j); + dimm->grain = 8; + dimm->edac_mode = mode; + dimm->mtype = mtype; } pci_read_config_dword(pdev, MC_SAG_CH_0, &value[0]); @@ -1567,22 +1452,16 @@ error: /**************************************************************************** Error check routines ****************************************************************************/ -static void i7core_rdimm_update_csrow(struct mem_ctl_info *mci, +static void i7core_rdimm_update_errcount(struct mem_ctl_info *mci, const int chan, const int dimm, const int add) { - char *msg; - struct i7core_pvt *pvt = mci->pvt_info; - int row = pvt->csrow_map[chan][dimm], i; + int i; for (i = 0; i < add; i++) { - msg = kasprintf(GFP_KERNEL, "Corrected error " - "(Socket=%d channel=%d dimm=%d)", - pvt->i7core_dev->socket, chan, dimm); - - edac_mc_handle_fbd_ce(mci, row, 0, msg); - kfree (msg); + edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, 0, 0, 0, + chan, dimm, -1, "error", "", NULL); } } @@ -1623,11 +1502,11 @@ static void i7core_rdimm_update_ce_count(struct mem_ctl_info *mci, /*updated the edac core */ if (add0 != 0) - i7core_rdimm_update_csrow(mci, chan, 0, add0); + i7core_rdimm_update_errcount(mci, chan, 0, add0); if (add1 != 0) - i7core_rdimm_update_csrow(mci, chan, 1, add1); + i7core_rdimm_update_errcount(mci, chan, 1, add1); if (add2 != 0) - i7core_rdimm_update_csrow(mci, chan, 2, add2); + i7core_rdimm_update_errcount(mci, chan, 2, add2); } @@ -1747,20 +1626,30 @@ static void i7core_mce_output_error(struct mem_ctl_info *mci, const struct mce *m) { struct i7core_pvt *pvt = mci->pvt_info; - char *type, *optype, *err, *msg; + char *type, *optype, *err, msg[80]; + enum hw_event_mc_err_type tp_event; unsigned long error = m->status & 0x1ff0000l; + bool uncorrected_error = m->mcgstatus & 1ll << 61; + bool ripv = m->mcgstatus & 1; u32 optypenum = (m->status >> 4) & 0x07; u32 core_err_cnt = (m->status >> 38) & 0x7fff; u32 dimm = (m->misc >> 16) & 0x3; u32 channel = (m->misc >> 18) & 0x3; u32 syndrome = m->misc >> 32; u32 errnum = find_first_bit(&error, 32); - int csrow; - if (m->mcgstatus & 1) - type = "FATAL"; - else - type = "NON_FATAL"; + if (uncorrected_error) { + if (ripv) { + type = "FATAL"; + tp_event = HW_EVENT_ERR_FATAL; + } else { + type = "NON_FATAL"; + tp_event = HW_EVENT_ERR_UNCORRECTED; + } + } else { + type = "CORRECTED"; + tp_event = HW_EVENT_ERR_CORRECTED; + } switch (optypenum) { case 0: @@ -1815,27 +1704,20 @@ static void i7core_mce_output_error(struct mem_ctl_info *mci, err = "unknown"; } - /* FIXME: should convert addr into bank and rank information */ - msg = kasprintf(GFP_ATOMIC, - "%s (addr = 0x%08llx, cpu=%d, Dimm=%d, Channel=%d, " - "syndrome=0x%08x, count=%d, Err=%08llx:%08llx (%s: %s))\n", - type, (long long) m->addr, m->cpu, dimm, channel, - syndrome, core_err_cnt, (long long)m->status, - (long long)m->misc, optype, err); - - debugf0("%s", msg); - - csrow = pvt->csrow_map[channel][dimm]; + snprintf(msg, sizeof(msg), "count=%d %s", core_err_cnt, optype); - /* Call the helper to output message */ - if (m->mcgstatus & 1) - edac_mc_handle_fbd_ue(mci, csrow, 0, - 0 /* FIXME: should be channel here */, msg); - else if (!pvt->is_registered) - edac_mc_handle_fbd_ce(mci, csrow, - 0 /* FIXME: should be channel here */, msg); - - kfree(msg); + /* + * Call the helper to output message + * FIXME: what to do if core_err_cnt > 1? Currently, it generates + * only one event + */ + if (uncorrected_error || !pvt->is_registered) + edac_mc_handle_error(tp_event, mci, + m->addr >> PAGE_SHIFT, + m->addr & ~PAGE_MASK, + syndrome, + channel, dimm, -1, + err, msg, m); } /* @@ -2252,15 +2134,19 @@ static int i7core_register_mci(struct i7core_dev *i7core_dev) { struct mem_ctl_info *mci; struct i7core_pvt *pvt; - int rc, channels, csrows; - - /* Check the number of active and not disabled channels */ - rc = i7core_get_active_channels(i7core_dev->socket, &channels, &csrows); - if (unlikely(rc < 0)) - return rc; + int rc; + struct edac_mc_layer layers[2]; /* allocate a new MC control structure */ - mci = edac_mc_alloc(sizeof(*pvt), csrows, channels, i7core_dev->socket); + + layers[0].type = EDAC_MC_LAYER_CHANNEL; + layers[0].size = NUM_CHANS; + layers[0].is_virt_csrow = false; + layers[1].type = EDAC_MC_LAYER_SLOT; + layers[1].size = MAX_DIMMS; + layers[1].is_virt_csrow = true; + mci = edac_mc_alloc(i7core_dev->socket, ARRAY_SIZE(layers), layers, + sizeof(*pvt)); if (unlikely(!mci)) return -ENOMEM; |