| Message ID | 20210623132226.140341-1-saubhik.mukherjee@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | char: tpm: vtpm_proxy: Fix race in init | expand |
On Wed, Jun 23, 2021 at 06:52:26PM +0530, Saubhik Mukherjee wrote: > vtpm_module_init calls vtpmx_init which calls misc_register. The file > operations callbacks are registered. So, vtpmx_fops_ioctl can execute in > parallel with rest of vtpm_module_init. vtpmx_fops_ioctl calls > vtpmx_ioc_new_dev, which calls vtpm_proxy_create_device, which calls > vtpm_proxy_work_start, which could read uninitialized workqueue. > > To avoid this, create workqueue before vtpmx init. > > Found by Linux Driver Verification project (linuxtesting.org). > > Signed-off-by: Saubhik Mukherjee <saubhik.mukherjee@gmail.com> > --- > drivers/char/tpm/tpm_vtpm_proxy.c | 19 +++++++++---------- > 1 file changed, 9 insertions(+), 10 deletions(-) > > diff --git a/drivers/char/tpm/tpm_vtpm_proxy.c b/drivers/char/tpm/tpm_vtpm_proxy.c > index 91c772e38bb5..225dfa026a8f 100644 > --- a/drivers/char/tpm/tpm_vtpm_proxy.c > +++ b/drivers/char/tpm/tpm_vtpm_proxy.c > @@ -697,23 +697,22 @@ static int __init vtpm_module_init(void) > { > int rc; > > - rc = vtpmx_init(); > - if (rc) { > - pr_err("couldn't create vtpmx device\n"); > - return rc; > - } > - > workqueue = create_workqueue("tpm-vtpm"); > if (!workqueue) { > pr_err("couldn't create workqueue\n"); > - rc = -ENOMEM; > - goto err_vtpmx_cleanup; > + return -ENOMEM; > + } > + > + rc = vtpmx_init(); > + if (rc) { > + pr_err("couldn't create vtpmx device\n"); > + goto err_destroy_workqueue; > } > > return 0; > > -err_vtpmx_cleanup: > - vtpmx_cleanup(); > +err_destroy_workqueue: > + destroy_workqueue(workqueue); > > return rc; > } > -- > 2.30.2 > > Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org> /Jarkko
On Tue, Jun 29, 2021 at 08:27:00PM +0300, Jarkko Sakkinen wrote: > On Wed, Jun 23, 2021 at 06:52:26PM +0530, Saubhik Mukherjee wrote: > > vtpm_module_init calls vtpmx_init which calls misc_register. The file > > operations callbacks are registered. So, vtpmx_fops_ioctl can execute in > > parallel with rest of vtpm_module_init. vtpmx_fops_ioctl calls > > vtpmx_ioc_new_dev, which calls vtpm_proxy_create_device, which calls > > vtpm_proxy_work_start, which could read uninitialized workqueue. > > > > To avoid this, create workqueue before vtpmx init. > > > > Found by Linux Driver Verification project (linuxtesting.org). > > > > Signed-off-by: Saubhik Mukherjee <saubhik.mukherjee@gmail.com> > > --- > > drivers/char/tpm/tpm_vtpm_proxy.c | 19 +++++++++---------- > > 1 file changed, 9 insertions(+), 10 deletions(-) > > > > diff --git a/drivers/char/tpm/tpm_vtpm_proxy.c b/drivers/char/tpm/tpm_vtpm_proxy.c > > index 91c772e38bb5..225dfa026a8f 100644 > > --- a/drivers/char/tpm/tpm_vtpm_proxy.c > > +++ b/drivers/char/tpm/tpm_vtpm_proxy.c > > @@ -697,23 +697,22 @@ static int __init vtpm_module_init(void) > > { > > int rc; > > > > - rc = vtpmx_init(); > > - if (rc) { > > - pr_err("couldn't create vtpmx device\n"); > > - return rc; > > - } > > - > > workqueue = create_workqueue("tpm-vtpm"); > > if (!workqueue) { > > pr_err("couldn't create workqueue\n"); > > - rc = -ENOMEM; > > - goto err_vtpmx_cleanup; > > + return -ENOMEM; > > + } > > + > > + rc = vtpmx_init(); > > + if (rc) { > > + pr_err("couldn't create vtpmx device\n"); > > + goto err_destroy_workqueue; > > } > > > > return 0; > > > > -err_vtpmx_cleanup: > > - vtpmx_cleanup(); > > +err_destroy_workqueue: > > + destroy_workqueue(workqueue); > > > > return rc; > > } > > -- > > 2.30.2 > > > > > > Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org> Taking reviewed-by back. You're lacking fixes tag. Please re-send with one. /Jarkko
On 6/30/21 2:35 AM, Jarkko Sakkinen wrote: > On Tue, Jun 29, 2021 at 08:27:00PM +0300, Jarkko Sakkinen wrote: >> On Wed, Jun 23, 2021 at 06:52:26PM +0530, Saubhik Mukherjee wrote: >>> vtpm_module_init calls vtpmx_init which calls misc_register. The file >>> operations callbacks are registered. So, vtpmx_fops_ioctl can execute in >>> parallel with rest of vtpm_module_init. vtpmx_fops_ioctl calls >>> vtpmx_ioc_new_dev, which calls vtpm_proxy_create_device, which calls >>> vtpm_proxy_work_start, which could read uninitialized workqueue. >>> >>> To avoid this, create workqueue before vtpmx init. >>> >>> Found by Linux Driver Verification project (linuxtesting.org). >>> >>> Signed-off-by: Saubhik Mukherjee <saubhik.mukherjee@gmail.com> >>> --- >>> drivers/char/tpm/tpm_vtpm_proxy.c | 19 +++++++++---------- >>> 1 file changed, 9 insertions(+), 10 deletions(-) >>> >>> diff --git a/drivers/char/tpm/tpm_vtpm_proxy.c b/drivers/char/tpm/tpm_vtpm_proxy.c >>> index 91c772e38bb5..225dfa026a8f 100644 >>> --- a/drivers/char/tpm/tpm_vtpm_proxy.c >>> +++ b/drivers/char/tpm/tpm_vtpm_proxy.c >>> @@ -697,23 +697,22 @@ static int __init vtpm_module_init(void) >>> { >>> int rc; >>> >>> - rc = vtpmx_init(); >>> - if (rc) { >>> - pr_err("couldn't create vtpmx device\n"); >>> - return rc; >>> - } >>> - >>> workqueue = create_workqueue("tpm-vtpm"); >>> if (!workqueue) { >>> pr_err("couldn't create workqueue\n"); >>> - rc = -ENOMEM; >>> - goto err_vtpmx_cleanup; >>> + return -ENOMEM; >>> + } >>> + >>> + rc = vtpmx_init(); >>> + if (rc) { >>> + pr_err("couldn't create vtpmx device\n"); >>> + goto err_destroy_workqueue; >>> } >>> >>> return 0; >>> >>> -err_vtpmx_cleanup: >>> - vtpmx_cleanup(); >>> +err_destroy_workqueue: >>> + destroy_workqueue(workqueue); >>> >>> return rc; >>> } >>> -- >>> 2.30.2 >>> >>> >> >> Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org> > > Taking reviewed-by back. > > You're lacking fixes tag. Please re-send with one. > > /Jarkko > Thank you for noticing. I sent the patch containing the Fixes tag in reply to your last message. (https://lkml.org/lkml/2021/6/30/104)
On Wed, Jun 30, 2021 at 12:54:30PM +0530, Saubhik Mukherjee wrote: > On 6/30/21 2:35 AM, Jarkko Sakkinen wrote: > > On Tue, Jun 29, 2021 at 08:27:00PM +0300, Jarkko Sakkinen wrote: > > > On Wed, Jun 23, 2021 at 06:52:26PM +0530, Saubhik Mukherjee wrote: > > > > vtpm_module_init calls vtpmx_init which calls misc_register. The file > > > > operations callbacks are registered. So, vtpmx_fops_ioctl can execute in > > > > parallel with rest of vtpm_module_init. vtpmx_fops_ioctl calls > > > > vtpmx_ioc_new_dev, which calls vtpm_proxy_create_device, which calls > > > > vtpm_proxy_work_start, which could read uninitialized workqueue. > > > > > > > > To avoid this, create workqueue before vtpmx init. > > > > > > > > Found by Linux Driver Verification project (linuxtesting.org). > > > > > > > > Signed-off-by: Saubhik Mukherjee <saubhik.mukherjee@gmail.com> > > > > --- > > > > drivers/char/tpm/tpm_vtpm_proxy.c | 19 +++++++++---------- > > > > 1 file changed, 9 insertions(+), 10 deletions(-) > > > > > > > > diff --git a/drivers/char/tpm/tpm_vtpm_proxy.c b/drivers/char/tpm/tpm_vtpm_proxy.c > > > > index 91c772e38bb5..225dfa026a8f 100644 > > > > --- a/drivers/char/tpm/tpm_vtpm_proxy.c > > > > +++ b/drivers/char/tpm/tpm_vtpm_proxy.c > > > > @@ -697,23 +697,22 @@ static int __init vtpm_module_init(void) > > > > { > > > > int rc; > > > > - rc = vtpmx_init(); > > > > - if (rc) { > > > > - pr_err("couldn't create vtpmx device\n"); > > > > - return rc; > > > > - } > > > > - > > > > workqueue = create_workqueue("tpm-vtpm"); > > > > if (!workqueue) { > > > > pr_err("couldn't create workqueue\n"); > > > > - rc = -ENOMEM; > > > > - goto err_vtpmx_cleanup; > > > > + return -ENOMEM; > > > > + } > > > > + > > > > + rc = vtpmx_init(); > > > > + if (rc) { > > > > + pr_err("couldn't create vtpmx device\n"); > > > > + goto err_destroy_workqueue; > > > > } > > > > return 0; > > > > -err_vtpmx_cleanup: > > > > - vtpmx_cleanup(); > > > > +err_destroy_workqueue: > > > > + destroy_workqueue(workqueue); > > > > return rc; > > > > } > > > > -- > > > > 2.30.2 > > > > > > > > > > > > > > Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org> > > > > Taking reviewed-by back. > > > > You're lacking fixes tag. Please re-send with one. > > > > /Jarkko > > > > Thank you for noticing. I sent the patch containing the Fixes tag in reply > to your last message. (https://lkml.org/lkml/2021/6/30/104) Please do not do that. Instead, version your patches (git format-patch -vX) and send them as separate threads. It's all documented: https://www.kernel.org/doc/html/v5.13/process/submitting-patches.html /Jarkko
diff --git a/drivers/char/tpm/tpm_vtpm_proxy.c b/drivers/char/tpm/tpm_vtpm_proxy.c index 91c772e38bb5..225dfa026a8f 100644 --- a/drivers/char/tpm/tpm_vtpm_proxy.c +++ b/drivers/char/tpm/tpm_vtpm_proxy.c @@ -697,23 +697,22 @@ static int __init vtpm_module_init(void) { int rc; - rc = vtpmx_init(); - if (rc) { - pr_err("couldn't create vtpmx device\n"); - return rc; - } - workqueue = create_workqueue("tpm-vtpm"); if (!workqueue) { pr_err("couldn't create workqueue\n"); - rc = -ENOMEM; - goto err_vtpmx_cleanup; + return -ENOMEM; + } + + rc = vtpmx_init(); + if (rc) { + pr_err("couldn't create vtpmx device\n"); + goto err_destroy_workqueue; } return 0; -err_vtpmx_cleanup: - vtpmx_cleanup(); +err_destroy_workqueue: + destroy_workqueue(workqueue); return rc; }
vtpm_module_init calls vtpmx_init which calls misc_register. The file operations callbacks are registered. So, vtpmx_fops_ioctl can execute in parallel with rest of vtpm_module_init. vtpmx_fops_ioctl calls vtpmx_ioc_new_dev, which calls vtpm_proxy_create_device, which calls vtpm_proxy_work_start, which could read uninitialized workqueue. To avoid this, create workqueue before vtpmx init. Found by Linux Driver Verification project (linuxtesting.org). Signed-off-by: Saubhik Mukherjee <saubhik.mukherjee@gmail.com> --- drivers/char/tpm/tpm_vtpm_proxy.c | 19 +++++++++---------- 1 file changed, 9 insertions(+), 10 deletions(-)