| Message ID | 1564092101-3865-3-git-send-email-jane.chu@oracle.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | mm/memory-failure: Poison read receives SIGKILL instead of SIGBUS issue | expand |
On Thu, Jul 25, 2019 at 04:01:41PM -0600, Jane Chu wrote: > Mmap /dev/dax more than once, then read the poison location using address > from one of the mappings. The other mappings due to not having the page > mapped in will cause SIGKILLs delivered to the process. SIGKILL succeeds > over SIGBUS, so user process looses the opportunity to handle the UE. > > Although one may add MAP_POPULATE to mmap(2) to work around the issue, > MAP_POPULATE makes mapping 128GB of pmem several magnitudes slower, so > isn't always an option. > > Details - > > ndctl inject-error --block=10 --count=1 namespace6.0 > > ./read_poison -x dax6.0 -o 5120 -m 2 > mmaped address 0x7f5bb6600000 > mmaped address 0x7f3cf3600000 > doing local read at address 0x7f3cf3601400 > Killed > > Console messages in instrumented kernel - > > mce: Uncorrected hardware memory error in user-access at edbe201400 > Memory failure: tk->addr = 7f5bb6601000 > Memory failure: address edbe201: call dev_pagemap_mapping_shift > dev_pagemap_mapping_shift: page edbe201: no PUD > Memory failure: tk->size_shift == 0 > Memory failure: Unable to find user space address edbe201 in read_poison > Memory failure: tk->addr = 7f3cf3601000 > Memory failure: address edbe201: call dev_pagemap_mapping_shift > Memory failure: tk->size_shift = 21 > Memory failure: 0xedbe201: forcibly killing read_poison:22434 because of failure to unmap corrupted page > => to deliver SIGKILL > Memory failure: 0xedbe201: Killing read_poison:22434 due to hardware memory corruption > => to deliver SIGBUS > > Signed-off-by: Jane Chu <jane.chu@oracle.com> > Suggested-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> > --- > mm/memory-failure.c | 22 +++++++++++++--------- > 1 file changed, 13 insertions(+), 9 deletions(-) > > diff --git a/mm/memory-failure.c b/mm/memory-failure.c > index 51d5b20..f668c88 100644 > --- a/mm/memory-failure.c > +++ b/mm/memory-failure.c > @@ -199,7 +199,6 @@ struct to_kill { > struct task_struct *tsk; > unsigned long addr; > short size_shift; > - char addr_valid; > }; > > /* > @@ -318,22 +317,27 @@ static void add_to_kill(struct task_struct *tsk, struct page *p, > } > > tk->addr = page_address_in_vma(p, vma); > - tk->addr_valid = 1; > if (is_zone_device_page(p)) > tk->size_shift = dev_pagemap_mapping_shift(p, vma); > else > tk->size_shift = compound_order(compound_head(p)) + PAGE_SHIFT; > > /* > - * In theory we don't have to kill when the page was > - * munmaped. But it could be also a mremap. Since that's > - * likely very rare kill anyways just out of paranoia, but use > - * a SIGKILL because the error is not contained anymore. > + * Send SIGKILL if "tk->addr == -EFAULT". Also, as > + * "tk->size_shift" is always non-zero for !is_zone_device_page(), > + * so "tk->size_shift == 0" effectively checks no mapping on > + * ZONE_DEVICE. Indeed, when a devdax page is mmapped N times > + * to a process' address space, it's possible not all N VMAs > + * contain mappings for the page, but at least one VMA does. > + * Only deliver SIGBUS with payload derived from the VMA that > + * has a mapping for the page. > */ > - if (tk->addr == -EFAULT || tk->size_shift == 0) { > + if (tk->addr == -EFAULT) { ^ (sorry nitpicking...) there's a trailing whitespace. Otherwise looks good to me. Acked-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> > pr_info("Memory failure: Unable to find user space address %lx in %s\n", > page_to_pfn(p), tsk->comm); > - tk->addr_valid = 0; > + } else if (tk->size_shift == 0) { > + kfree(tk); > + return; > } > > get_task_struct(tsk); > @@ -361,7 +365,7 @@ static void kill_procs(struct list_head *to_kill, int forcekill, bool fail, > * make sure the process doesn't catch the > * signal and then access the memory. Just kill it. > */ > - if (fail || tk->addr_valid == 0) { > + if (fail || tk->addr == -EFAULT) { > pr_err("Memory failure: %#lx: forcibly killing %s:%d because of failure to unmap corrupted page\n", > pfn, tk->tsk->comm, tk->tsk->pid); > do_send_sig_info(SIGKILL, SEND_SIG_PRIV, > -- > 1.8.3.1 > >
On Thu, Jul 25, 2019 at 04:01:41PM -0600, Jane Chu wrote: > Mmap /dev/dax more than once, then read the poison location using address > from one of the mappings. The other mappings due to not having the page > mapped in will cause SIGKILLs delivered to the process. SIGKILL succeeds > over SIGBUS, so user process looses the opportunity to handle the UE. > > Although one may add MAP_POPULATE to mmap(2) to work around the issue, > MAP_POPULATE makes mapping 128GB of pmem several magnitudes slower, so > isn't always an option. > > Details - > > ndctl inject-error --block=10 --count=1 namespace6.0 > > ./read_poison -x dax6.0 -o 5120 -m 2 > mmaped address 0x7f5bb6600000 > mmaped address 0x7f3cf3600000 > doing local read at address 0x7f3cf3601400 > Killed > > Console messages in instrumented kernel - > > mce: Uncorrected hardware memory error in user-access at edbe201400 > Memory failure: tk->addr = 7f5bb6601000 > Memory failure: address edbe201: call dev_pagemap_mapping_shift > dev_pagemap_mapping_shift: page edbe201: no PUD > Memory failure: tk->size_shift == 0 > Memory failure: Unable to find user space address edbe201 in read_poison > Memory failure: tk->addr = 7f3cf3601000 > Memory failure: address edbe201: call dev_pagemap_mapping_shift > Memory failure: tk->size_shift = 21 > Memory failure: 0xedbe201: forcibly killing read_poison:22434 because of failure to unmap corrupted page > => to deliver SIGKILL > Memory failure: 0xedbe201: Killing read_poison:22434 due to hardware memory corruption > => to deliver SIGBUS > > Signed-off-by: Jane Chu <jane.chu@oracle.com> > Suggested-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Thanks for the fix. Acked-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
diff --git a/mm/memory-failure.c b/mm/memory-failure.c index 51d5b20..f668c88 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -199,7 +199,6 @@ struct to_kill { struct task_struct *tsk; unsigned long addr; short size_shift; - char addr_valid; }; /* @@ -318,22 +317,27 @@ static void add_to_kill(struct task_struct *tsk, struct page *p, } tk->addr = page_address_in_vma(p, vma); - tk->addr_valid = 1; if (is_zone_device_page(p)) tk->size_shift = dev_pagemap_mapping_shift(p, vma); else tk->size_shift = compound_order(compound_head(p)) + PAGE_SHIFT; /* - * In theory we don't have to kill when the page was - * munmaped. But it could be also a mremap. Since that's - * likely very rare kill anyways just out of paranoia, but use - * a SIGKILL because the error is not contained anymore. + * Send SIGKILL if "tk->addr == -EFAULT". Also, as + * "tk->size_shift" is always non-zero for !is_zone_device_page(), + * so "tk->size_shift == 0" effectively checks no mapping on + * ZONE_DEVICE. Indeed, when a devdax page is mmapped N times + * to a process' address space, it's possible not all N VMAs + * contain mappings for the page, but at least one VMA does. + * Only deliver SIGBUS with payload derived from the VMA that + * has a mapping for the page. */ - if (tk->addr == -EFAULT || tk->size_shift == 0) { + if (tk->addr == -EFAULT) { pr_info("Memory failure: Unable to find user space address %lx in %s\n", page_to_pfn(p), tsk->comm); - tk->addr_valid = 0; + } else if (tk->size_shift == 0) { + kfree(tk); + return; } get_task_struct(tsk); @@ -361,7 +365,7 @@ static void kill_procs(struct list_head *to_kill, int forcekill, bool fail, * make sure the process doesn't catch the * signal and then access the memory. Just kill it. */ - if (fail || tk->addr_valid == 0) { + if (fail || tk->addr == -EFAULT) { pr_err("Memory failure: %#lx: forcibly killing %s:%d because of failure to unmap corrupted page\n", pfn, tk->tsk->comm, tk->tsk->pid); do_send_sig_info(SIGKILL, SEND_SIG_PRIV,
Mmap /dev/dax more than once, then read the poison location using address from one of the mappings. The other mappings due to not having the page mapped in will cause SIGKILLs delivered to the process. SIGKILL succeeds over SIGBUS, so user process looses the opportunity to handle the UE. Although one may add MAP_POPULATE to mmap(2) to work around the issue, MAP_POPULATE makes mapping 128GB of pmem several magnitudes slower, so isn't always an option. Details - ndctl inject-error --block=10 --count=1 namespace6.0 ./read_poison -x dax6.0 -o 5120 -m 2 mmaped address 0x7f5bb6600000 mmaped address 0x7f3cf3600000 doing local read at address 0x7f3cf3601400 Killed Console messages in instrumented kernel - mce: Uncorrected hardware memory error in user-access at edbe201400 Memory failure: tk->addr = 7f5bb6601000 Memory failure: address edbe201: call dev_pagemap_mapping_shift dev_pagemap_mapping_shift: page edbe201: no PUD Memory failure: tk->size_shift == 0 Memory failure: Unable to find user space address edbe201 in read_poison Memory failure: tk->addr = 7f3cf3601000 Memory failure: address edbe201: call dev_pagemap_mapping_shift Memory failure: tk->size_shift = 21 Memory failure: 0xedbe201: forcibly killing read_poison:22434 because of failure to unmap corrupted page => to deliver SIGKILL Memory failure: 0xedbe201: Killing read_poison:22434 due to hardware memory corruption => to deliver SIGBUS Signed-off-by: Jane Chu <jane.chu@oracle.com> Suggested-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> --- mm/memory-failure.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-)