| Message ID | 20210714193542.21857-12-joao.m.martins@oracle.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | mm, sparse-vmemmap: Introduce compound pagemaps | expand |
On Wed, Jul 14, 2021 at 12:36 PM Joao Martins <joao.m.martins@oracle.com> wrote: > > Right now, only static dax regions have a valid @pgmap pointer in its > struct dev_dax. Dynamic dax case however, do not. > > In preparation for device-dax compound pagemap support, make sure that > dev_dax pgmap field is set after it has been allocated and initialized. I think this is ok to fold into the patch that needs it. > > Suggested-by: Dan Williams <dan.j.williams@intel.com> > Signed-off-by: Joao Martins <joao.m.martins@oracle.com> > --- > drivers/dax/device.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/drivers/dax/device.c b/drivers/dax/device.c > index 0b82159b3564..6e348b5f9d45 100644 > --- a/drivers/dax/device.c > +++ b/drivers/dax/device.c > @@ -426,6 +426,8 @@ int dev_dax_probe(struct dev_dax *dev_dax) > } > > pgmap->type = MEMORY_DEVICE_GENERIC; > + dev_dax->pgmap = pgmap; > + > addr = devm_memremap_pages(dev, pgmap); > if (IS_ERR(addr)) > return PTR_ERR(addr); > -- > 2.17.1 >
On 7/28/21 8:30 AM, Dan Williams wrote: > On Wed, Jul 14, 2021 at 12:36 PM Joao Martins <joao.m.martins@oracle.com> wrote: >> >> Right now, only static dax regions have a valid @pgmap pointer in its >> struct dev_dax. Dynamic dax case however, do not. >> >> In preparation for device-dax compound pagemap support, make sure that >> dev_dax pgmap field is set after it has been allocated and initialized. > > I think this is ok to fold into the patch that needs it. OK, I've squashed that in.
On 7/28/21 4:56 PM, Joao Martins wrote: > On 7/28/21 8:30 AM, Dan Williams wrote: >> On Wed, Jul 14, 2021 at 12:36 PM Joao Martins <joao.m.martins@oracle.com> wrote: >>> >>> Right now, only static dax regions have a valid @pgmap pointer in its >>> struct dev_dax. Dynamic dax case however, do not. >>> >>> In preparation for device-dax compound pagemap support, make sure that >>> dev_dax pgmap field is set after it has been allocated and initialized. >> >> I think this is ok to fold into the patch that needs it. > > OK, I've squashed that in. > I am wondering now whether I should un-squash this into a separate patch. It regresses one of the test suites from ndctl for dynamic dax regions. memremap_pages() calls (from a dynamic region) starts hitting rather random splats (below) on the 'root' device (not the dax device children aiui). This fixes it but I am yet 100% if should be this way: diff --git a/drivers/dax/bus.c b/drivers/dax/bus.c index 6cc4da4c713d..2a3a70e62d89 100644 --- a/drivers/dax/bus.c +++ b/drivers/dax/bus.c @@ -363,6 +363,9 @@ void kill_dev_dax(struct dev_dax *dev_dax) kill_dax(dax_dev); unmap_mapping_range(inode->i_mapping, 0, 0, 1); + + if (!is_static(dev_dax->region)) + dev_dax->pgmap = NULL; } EXPORT_SYMBOL_GPL(kill_dev_dax); stacktraces: [ 834.884765] ------------[ cut here ]------------ [ 834.885887] nr_range must be specified [ 834.886461] WARNING: CPU: 3 PID: 3148 at mm/memremap.c:347 memremap_pages+0x54f/0x610 [ 834.887578] Modules linked in: [ 834.887993] CPU: 3 PID: 3148 Comm: lt-daxctl Not tainted 5.14.0-rc3-next-20210729+ #201 [ 834.889131] RIP: 0010:memremap_pages+0x54f/0x610 [ 834.889886] Code: e1 97 00 0f 0b e9 14 fc ff ff 80 3d 11 99 79 01 00 0f 85 3e fc ff ff 48 c7 c7 e8 78 5e 82 c6 05 fd 98 79 01 01 e8 f8 e0 97 00 <0f> 0b 48 c7 c0 ea ff ff ff e9 2f fb ff ff 48 c7 c7 b0 a7 d1 82 e8 [ 834.892491] RSP: 0018:ffffc90008fb3b40 EFLAGS: 00010286 [ 834.893249] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000027 [ 834.894250] RDX: 0000000000000000 RSI: ffff88900fcd7610 RDI: ffff88900fcd7618 [ 834.895283] RBP: ffffc90008fb3b98 R08: 0000000000000000 R09: c0000000ffffefff [ 834.896286] R10: ffffc90008fb3930 R11: ffffc90008fb3928 R12: 8000000000000063 [ 834.897279] R13: ffff889042856430 R14: 0000000000000001 R15: ffff889042856400 [ 834.898279] FS: 00007f83ddfd8780(0000) GS:ffff88900fcc0000(0000) knlGS:0000000000000000 [ 834.899423] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 834.900194] CR2: 00007ffdb895c148 CR3: 0000000106bfc006 CR4: 0000000000170ee0 [ 834.901195] Call Trace: [ 834.901651] devm_memremap_pages+0x22/0x70 [ 834.902244] dev_dax_probe+0x175/0x2b0 [ 834.902797] dax_bus_probe+0x73/0xa0 [ 834.903285] really_probe+0xcf/0x3a0 [ 834.903801] __driver_probe_device+0xb3/0x130 [ 834.904363] driver_probe_device+0x24/0x90 [ 834.904967] __driver_attach+0xa1/0x170 [ 834.905525] ? __device_attach_driver+0xe0/0xe0 [ 834.906164] ? __device_attach_driver+0xe0/0xe0 [ 834.906818] bus_for_each_dev+0x74/0xb0 [ 834.907321] driver_attach+0x1e/0x20 [ 834.907875] do_id_store+0x1db/0x210 [ 834.908352] new_id_store+0x13/0x20 [ 834.908865] drv_attr_store+0x27/0x40 [ 834.909398] sysfs_kf_write+0x3b/0x50 [ 834.909968] kernfs_fop_write_iter+0x128/0x1b0 [ 834.910635] new_sync_write+0x117/0x1b0 [ 834.911139] vfs_write+0x181/0x250 [ 834.911644] ? do_sys_openat2+0x1d7/0x300 [ 834.912216] ksys_write+0x61/0xe0 [ 834.912698] __x64_sys_write+0x1a/0x20 [ 834.913242] do_syscall_64+0x3a/0x80 [ 834.913772] entry_SYSCALL_64_after_hwframe+0x44/0xae [ 834.914491] RIP: 0033:0x7f83dd4cda00 [ 834.915010] Code: 73 01 c3 48 8b 0d 70 74 2d 00 f7 d8 64 89 01 48 83 c8 ff c3 66 0f 1f 44 00 00 83 3d bd d5 2d 00 00 75 10 b8 01 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 31 c3 48 83 ec 08 e8 7e cc 01 00 48 89 04 24 [ 834.917543] RSP: 002b:00007ffdb895f308 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 834.918627] RAX: ffffffffffffffda RBX: 0000000000000004 RCX: 00007f83dd4cda00 [ 834.919650] RDX: 0000000000000007 RSI: 00000000020c639d RDI: 0000000000000004 [ 834.920658] RBP: 00000000020c639d R08: 0000000000000000 R09: 00000000020c45c0 [ 834.921649] R10: 2f7861645f656369 R11: 0000000000000246 R12: 0000000000000007 [ 834.922670] R13: 0000000000000001 R14: 00000000020c71b0 R15: 00000000020cc2e0 [ 834.923650] ---[ end trace 61ea8560dbeb3a89 ]--- [ 835.941573] Hotplug memory [0x0-0x0] exceeds maximum addressable range [0x0-0x3fffffffffff] [ 835.944269] ------------[ cut here ]------------ [ 835.945412] kernel BUG at mm/memory_hotplug.c:316! [ 835.946574] invalid opcode: 0000 [#1] SMP PTI [ 835.947622] CPU: 6 PID: 3593 Comm: lt-daxctl Tainted: G W 5.14.0-rc3-next-20210729+ #201 [ 835.949845] RIP: 0010:__add_pages+0x100/0x150 [ 835.950899] Code: ba 89 55 d4 e8 4d 32 00 00 8b 55 d4 48 83 c4 08 89 d0 5b 41 5c 41 5d 41 5e 41 5f 5d c3 31 d2 eb e0 0f 0b ba ea ff ff ff eb e2 <0f> 0b 80 3d e6 43 dc 00 00 ba ea ff ff ff 75 d2 48 c7 c7 b8 35 5e [ 835.953662] RSP: 0018:ffffc900094bbab0 EFLAGS: 00010246 [ 835.954336] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000 [ 835.955276] RDX: 0000000000000000 RSI: ffff88900fd97610 RDI: ffff88900fd97610 [ 835.956251] RBP: ffffc900094bbae0 R08: 0000000000000000 R09: c0000000ffffefff [ 835.957214] R10: ffffc900094bb888 R11: ffffc900094bb880 R12: 0000000000000000 [ 835.958154] R13: 0000000000000000 R14: ffffc900094bbb50 R15: 0000000000000000 [ 835.959097] FS: 00007fd63c55e780(0000) GS:ffff88900fd80000(0000) knlGS:0000000000000000 [ 835.960163] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 835.960960] CR2: 00007fffc4d7abd8 CR3: 00000001126f2001 CR4: 0000000000170ee0 [ 835.961903] Call Trace: [ 835.962220] add_pages+0x17/0x70 [ 835.962696] arch_add_memory+0x45/0x50 [ 835.963175] memremap_pages+0x2a8/0x610 [ 835.963726] devm_memremap_pages+0x22/0x70 [ 835.964223] dev_dax_probe+0x175/0x2b0 [ 835.964778] dax_bus_probe+0x73/0xa0 [ 835.965215] really_probe+0xcf/0x3a0 [ 835.965717] __driver_probe_device+0xb3/0x130 [ 835.966248] driver_probe_device+0x24/0x90 [ 835.966835] __driver_attach+0xa1/0x170 [ 835.967340] ? __device_attach_driver+0xe0/0xe0 [ 835.967956] ? __device_attach_driver+0xe0/0xe0 [ 835.968589] bus_for_each_dev+0x74/0xb0 [ 835.969090] driver_attach+0x1e/0x20 [ 835.969605] do_id_store+0x1db/0x210 [ 835.970057] new_id_store+0x13/0x20 [ 835.970536] drv_attr_store+0x27/0x40 [ 835.971023] sysfs_kf_write+0x3b/0x50 [ 835.971529] kernfs_fop_write_iter+0x128/0x1b0 [ 835.972141] new_sync_write+0x117/0x1b0 [ 835.972697] vfs_write+0x181/0x250 [ 835.973130] ? do_sys_openat2+0x1d7/0x300 [ 835.973698] ksys_write+0x61/0xe0 [ 835.974117] __x64_sys_write+0x1a/0x20 [ 835.974654] do_syscall_64+0x3a/0x80 [ 835.975105] entry_SYSCALL_64_after_hwframe+0x44/0xae [ 835.975802] RIP: 0033:0x7fd63ba53a00 [ 835.976257] Code: 73 01 c3 48 8b 0d 70 74 2d 00 f7 d8 64 89 01 48 83 c8 ff c3 66 0f 1f 44 00 00 83 3d bd d5 2d 00 00 75 10 b8 01 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 31 c3 48 83 ec 08 e8 7e cc 01 00 48 89 04 24 [ 835.978754] RSP: 002b:00007fffc4d7dd98 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 835.979765] RAX: ffffffffffffffda RBX: 0000000000000004 RCX: 00007fd63ba53a00 [ 835.980761] RDX: 0000000000000007 RSI: 000000000081939d RDI: 0000000000000004 [ 835.981714] RBP: 000000000081939d R08: 0000000000000000 R09: 00000000008175c0 [ 835.982671] R10: 2f7861645f656369 R11: 0000000000000246 R12: 0000000000000007 [ 835.983647] R13: 0000000000000001 R14: 000000000081a1b0 R15: 000000000081f2e0 [ 835.984593] Modules linked in: [ 835.985041] ---[ end trace 61ea8560dbeb3a8a ]--- [ 835.985789] RIP: 0010:__add_pages+0x100/0x150
diff --git a/drivers/dax/device.c b/drivers/dax/device.c index 0b82159b3564..6e348b5f9d45 100644 --- a/drivers/dax/device.c +++ b/drivers/dax/device.c @@ -426,6 +426,8 @@ int dev_dax_probe(struct dev_dax *dev_dax) } pgmap->type = MEMORY_DEVICE_GENERIC; + dev_dax->pgmap = pgmap; + addr = devm_memremap_pages(dev, pgmap); if (IS_ERR(addr)) return PTR_ERR(addr);
Right now, only static dax regions have a valid @pgmap pointer in its struct dev_dax. Dynamic dax case however, do not. In preparation for device-dax compound pagemap support, make sure that dev_dax pgmap field is set after it has been allocated and initialized. Suggested-by: Dan Williams <dan.j.williams@intel.com> Signed-off-by: Joao Martins <joao.m.martins@oracle.com> --- drivers/dax/device.c | 2 ++ 1 file changed, 2 insertions(+)