| Message ID | 20190624150758.6695-25-rrichter@marvell.com |
|---|---|
| State | New |
| Headers | show |
| Series | EDAC, mc, ghes: Fixes and updates to improve memory error reporting | expand |
On 24/06/2019 16:09, Robert Richter wrote: > James Morse: "I'm all for removing/warning-its-broken it when > ghes_edac is in use." Thanks for taking that out of context. The very next word was 'but': http://lore.kernel.org/r/c08290d8-3690-efa9-3bc7-37f8b1fdbfd4@arm.com followed by details of the user-space that is still using this. > Let's just disable legacy API for the ghes driver on arm64. Though, I > don't agree with it as there still could be some userland tools that Not could. Are. Someone went and found them for you. > use this interface that cannot be used any longer after a transition > from x86 to arm64.
On 26.06.19 10:33:39, James Morse wrote: > On 24/06/2019 16:09, Robert Richter wrote: > > James Morse: "I'm all for removing/warning-its-broken it when > > ghes_edac is in use." > > Thanks for taking that out of context. The very next word was 'but': > http://lore.kernel.org/r/c08290d8-3690-efa9-3bc7-37f8b1fdbfd4@arm.com > > followed by details of the user-space that is still using this. > > > > Let's just disable legacy API for the ghes driver on arm64. Though, I > > don't agree with it as there still could be some userland tools that > > Not could. Are. Someone went and found them for you. > > > > use this interface that cannot be used any longer after a transition > > from x86 to arm64. Ok, thanks for clarifying. -Robert
diff --git a/drivers/edac/edac_mc.c b/drivers/edac/edac_mc.c index 9383a1179b83..d7a120d7e304 100644 --- a/drivers/edac/edac_mc.c +++ b/drivers/edac/edac_mc.c @@ -276,16 +276,11 @@ void *edac_align_ptr(void **p, unsigned size, int n_elems) return (void *)(((unsigned long)ptr) + align - r); } -static void _edac_mc_free(struct mem_ctl_info *mci) +static void _edac_mc_free_legacy(struct mem_ctl_info *mci) { - int i, chn, row; + int chn, row; struct csrow_info *csr; - if (mci->dimms) { - for (i = 0; i < mci->tot_dimms; i++) - kfree(mci->dimms[i]); - kfree(mci->dimms); - } if (mci->csrows) { for (row = 0; row < mci->nr_csrows; row++) { csr = mci->csrows[row]; @@ -300,9 +295,34 @@ static void _edac_mc_free(struct mem_ctl_info *mci) } kfree(mci->csrows); } +} + +static void _edac_mc_free(struct mem_ctl_info *mci) +{ + int i; + + if (mci->dimms) { + for (i = 0; i < mci->tot_dimms; i++) + kfree(mci->dimms[i]); + kfree(mci->dimms); + } + + _edac_mc_free_legacy(mci); + kfree(mci); } +void edac_mc_disable_legacy_api(struct mem_ctl_info *mci) +{ +#ifdef CONFIG_EDAC_LEGACY_SYSFS + _edac_mc_free_legacy(mci); + mci->nr_csrows = 0; + mci->num_cschannel = 0; + mci->csrows = NULL; +#endif +} +EXPORT_SYMBOL_GPL(edac_mc_disable_legacy_api); + struct mem_ctl_info *__edac_mc_alloc(unsigned mc_num, unsigned dimm_num, unsigned n_layers, diff --git a/drivers/edac/edac_mc.h b/drivers/edac/edac_mc.h index e8215847f853..d1bacc5f47f5 100644 --- a/drivers/edac/edac_mc.h +++ b/drivers/edac/edac_mc.h @@ -270,6 +270,13 @@ void edac_mc_handle_error(const enum hw_event_mc_err_type type, const char *msg, const char *other_detail); +/** + * edac_mc_disable_legacy_api() - Disable legacy sysfs API + * + * @mci: a struct mem_ctl_info pointer + */ +void edac_mc_disable_legacy_api(struct mem_ctl_info *mci); + /* * edac misc APIs */ diff --git a/drivers/edac/ghes_edac.c b/drivers/edac/ghes_edac.c index 714623204232..804e48773f66 100644 --- a/drivers/edac/ghes_edac.c +++ b/drivers/edac/ghes_edac.c @@ -780,7 +780,9 @@ ghes_edac_register_one(int nid, struct ghes *ghes, struct device *parent) mci->dev_name = dev_name(dev); mem_info_prepare_mci(mci); - +#ifdef CONFIG_ARM64 + edac_mc_disable_legacy_api(mci); +#endif rc = edac_mc_add_mc(mci); if (rc < 0) { pr_err("Can't register at EDAC core (%d)\n", rc);
James Morse: "I'm all for removing/warning-its-broken it when ghes_edac is in use." Let's just disable legacy API for the ghes driver on arm64. Though, I don't agree with it as there still could be some userland tools that use this interface that cannot be used any longer after a transition from x86 to arm64. I leave that decision up to James. Signed-off-by: Robert Richter <rrichter@marvell.com> --- drivers/edac/edac_mc.c | 34 +++++++++++++++++++++++++++------- drivers/edac/edac_mc.h | 7 +++++++ drivers/edac/ghes_edac.c | 4 +++- 3 files changed, 37 insertions(+), 8 deletions(-) -- 2.20.1