Re: [PATCH v3 2/8] s390/cio: introduce DMA pools to cio
On Mon, 3 Jun 2019 16:04:28 +0200 Halil Pasic wrote: > On Mon, 3 Jun 2019 14:09:02 +0200 > Michael Mueller wrote: > > > >> @@ -1059,16 +1168,19 @@ static int __init css_bus_init(void) > > >> if (ret) > > >> goto out_unregister; > > >> ret = register_pm_notifier(&css_power_notifier); > > >> -if (ret) { > > >> -unregister_reboot_notifier(&css_reboot_notifier); > > >> -goto out_unregister; > > >> -} > > >> +if (ret) > > >> +goto out_unregister_rn; > > >> +ret = cio_dma_pool_init(); > > >> +if (ret) > > >> +goto out_unregister_rn; > > > > > > Don't you also need to unregister the pm notifier on failure here? > > > > Mmh, that was the original intention. Thanks! > > I suppose we could also move cio_dma_pool_init() right before the > register_reboot_notifier() call and goto out_unregister on error. > Forget it, then we have to rollback the pool creation if the register stuff fails... Sorry for the noise. Regards, Halil ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Re: [PATCH v3 2/8] s390/cio: introduce DMA pools to cio
On Mon, 3 Jun 2019 14:09:02 +0200 Michael Mueller wrote: > >> @@ -1059,16 +1168,19 @@ static int __init css_bus_init(void) > >>if (ret) > >>goto out_unregister; > >>ret = register_pm_notifier(&css_power_notifier); > >> - if (ret) { > >> - unregister_reboot_notifier(&css_reboot_notifier); > >> - goto out_unregister; > >> - } > >> + if (ret) > >> + goto out_unregister_rn; > >> + ret = cio_dma_pool_init(); > >> + if (ret) > >> + goto out_unregister_rn; > > > > Don't you also need to unregister the pm notifier on failure here? > > Mmh, that was the original intention. Thanks! I suppose we could also move cio_dma_pool_init() right before the register_reboot_notifier() call and goto out_unregister on error. Regards, Halil ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Re: [PATCH v3 2/8] s390/cio: introduce DMA pools to cio
On 03.06.19 15:34, Cornelia Huck wrote: On Mon, 3 Jun 2019 14:57:30 +0200 Halil Pasic wrote: On Mon, 3 Jun 2019 14:09:02 +0200 Michael Mueller wrote: @@ -224,6 +226,8 @@ struct subchannel *css_alloc_subchannel(struct subchannel_id schid, INIT_WORK(&sch->todo_work, css_sch_todo); sch->dev.release = &css_subchannel_release; device_initialize(&sch->dev); It might be helpful to add a comment why you use 31 bit here... @Halil, please let me know what comment you prefere here... How about? /* * The physical addresses of some the dma structures that * can belong to a subchannel need to fit 31 bit width (examples ccw,). */ "e.g. ccw"? + sch->dev.coherent_dma_mask = DMA_BIT_MASK(31); + sch->dev.dma_mask = &sch->dev.coherent_dma_mask; return sch; err: @@ -899,6 +903,8 @@ static int __init setup_css(int nr) dev_set_name(&css->device, "css%x", nr); css->device.groups = cssdev_attr_groups; css->device.release = channel_subsystem_release; ...and 64 bit here. and here. /* * We currently allocate notifier bits with this (using css->device * as the device argument with the DMA API), and are fine with 64 bit * addresses. */ Thanks, that makes things hopefully clearer if we look at it some time in the future ;) Applied both with with requested change. Michael ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Re: [PATCH v3 2/8] s390/cio: introduce DMA pools to cio
On Mon, 3 Jun 2019 14:47:06 +0200 Halil Pasic wrote: > On Mon, 3 Jun 2019 13:37:45 +0200 > Cornelia Huck wrote: > > > On Wed, 29 May 2019 14:26:51 +0200 > > Michael Mueller wrote: > > > diff --git a/arch/s390/include/asm/cio.h b/arch/s390/include/asm/cio.h > > > index 1727180e8ca1..43c007d2775a 100644 > > > --- a/arch/s390/include/asm/cio.h > > > +++ b/arch/s390/include/asm/cio.h > > > @@ -328,6 +328,17 @@ static inline u8 pathmask_to_pos(u8 mask) > > > void channel_subsystem_reinit(void); > > > extern void css_schedule_reprobe(void); > > > > > > +extern void *cio_dma_zalloc(size_t size); > > > +extern void cio_dma_free(void *cpu_addr, size_t size); > > > +extern struct device *cio_get_dma_css_dev(void); > > > + > > > +struct gen_pool; > > > > That forward declaration is a bit ugly... > > Can you explain to me what is ugly about it so I can avoid similar > mistakes in the future? > > >I guess the alternative was > > include hell? > > > > What do you mean by include hell? > > I decided to use a forward declaration because the guys that include > "cio.h" are not expected to require the interfaces defined in > linux/genalloc.h. My motivation to do it like this was the principle of > encapsulation. My personal rule-of-thumb is to include the header if it is straightforward enough (e.g. if adding a basic header is enough). If you need to include a header together with all of its friends and family, a forward declaration is probably nicer. And of course, sometimes it is simply needed. ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Re: [PATCH v3 2/8] s390/cio: introduce DMA pools to cio
On Mon, 3 Jun 2019 14:57:30 +0200 Halil Pasic wrote: > On Mon, 3 Jun 2019 14:09:02 +0200 > Michael Mueller wrote: > > > >> @@ -224,6 +226,8 @@ struct subchannel *css_alloc_subchannel(struct > > >> subchannel_id schid, > > >> INIT_WORK(&sch->todo_work, css_sch_todo); > > >> sch->dev.release = &css_subchannel_release; > > >> device_initialize(&sch->dev); > > > > > > It might be helpful to add a comment why you use 31 bit here... > > > > @Halil, please let me know what comment you prefere here... > > > > How about? > > /* > * The physical addresses of some the dma structures that > * can belong to a subchannel need to fit 31 bit width (examples ccw,). > */ "e.g. ccw"? > > > > > > > >> +sch->dev.coherent_dma_mask = DMA_BIT_MASK(31); > > >> +sch->dev.dma_mask = &sch->dev.coherent_dma_mask; > > >> return sch; > > >> > > >> err: > > >> @@ -899,6 +903,8 @@ static int __init setup_css(int nr) > > >> dev_set_name(&css->device, "css%x", nr); > > >> css->device.groups = cssdev_attr_groups; > > >> css->device.release = channel_subsystem_release; > > > > > > ...and 64 bit here. > > > > and here. > > /* > * We currently allocate notifier bits with this (using css->device > * as the device argument with the DMA API), and are fine with 64 bit > * addresses. > */ Thanks, that makes things hopefully clearer if we look at it some time in the future ;) ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Re: [PATCH v3 2/8] s390/cio: introduce DMA pools to cio
On Mon, 3 Jun 2019 14:09:02 +0200 Michael Mueller wrote: > >> @@ -224,6 +226,8 @@ struct subchannel *css_alloc_subchannel(struct > >> subchannel_id schid, > >>INIT_WORK(&sch->todo_work, css_sch_todo); > >>sch->dev.release = &css_subchannel_release; > >>device_initialize(&sch->dev); > > > > It might be helpful to add a comment why you use 31 bit here... > > @Halil, please let me know what comment you prefere here... > How about? /* * The physical addresses of some the dma structures that * can belong to a subchannel need to fit 31 bit width (examples ccw,). */ > > > >> + sch->dev.coherent_dma_mask = DMA_BIT_MASK(31); > >> + sch->dev.dma_mask = &sch->dev.coherent_dma_mask; > >>return sch; > >> > >> err: > >> @@ -899,6 +903,8 @@ static int __init setup_css(int nr) > >>dev_set_name(&css->device, "css%x", nr); > >>css->device.groups = cssdev_attr_groups; > >>css->device.release = channel_subsystem_release; > > > > ...and 64 bit here. > > and here. /* * We currently allocate notifier bits with this (using css->device * as the device argument with the DMA API), and are fine with 64 bit * addresses. */ Regards, Halil ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Re: [PATCH v3 2/8] s390/cio: introduce DMA pools to cio
On Mon, 3 Jun 2019 13:37:45 +0200 Cornelia Huck wrote: > On Wed, 29 May 2019 14:26:51 +0200 > Michael Mueller wrote: > > > From: Halil Pasic > > > > To support protected virtualization cio will need to make sure the > > memory used for communication with the hypervisor is DMA memory. > > > > Let us introduce one global pool for cio. > > > > Our DMA pools are implemented as a gen_pool backed with DMA pages. The > > idea is to avoid each allocation effectively wasting a page, as we > > typically allocate much less than PAGE_SIZE. > > > > Signed-off-by: Halil Pasic > > Reviewed-by: Sebastian Ott > > Signed-off-by: Michael Mueller > > --- > > arch/s390/Kconfig | 1 + > > arch/s390/include/asm/cio.h | 11 > > drivers/s390/cio/css.c | 120 > > ++-- > > 3 files changed, 128 insertions(+), 4 deletions(-) > > (...) > > > diff --git a/arch/s390/include/asm/cio.h b/arch/s390/include/asm/cio.h > > index 1727180e8ca1..43c007d2775a 100644 > > --- a/arch/s390/include/asm/cio.h > > +++ b/arch/s390/include/asm/cio.h > > @@ -328,6 +328,17 @@ static inline u8 pathmask_to_pos(u8 mask) > > void channel_subsystem_reinit(void); > > extern void css_schedule_reprobe(void); > > > > +extern void *cio_dma_zalloc(size_t size); > > +extern void cio_dma_free(void *cpu_addr, size_t size); > > +extern struct device *cio_get_dma_css_dev(void); > > + > > +struct gen_pool; > > That forward declaration is a bit ugly... Can you explain to me what is ugly about it so I can avoid similar mistakes in the future? >I guess the alternative was > include hell? > What do you mean by include hell? I decided to use a forward declaration because the guys that include "cio.h" are not expected to require the interfaces defined in linux/genalloc.h. My motivation to do it like this was the principle of encapsulation. Regards, Halil ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Re: [PATCH v3 2/8] s390/cio: introduce DMA pools to cio
On 03.06.19 13:37, Cornelia Huck wrote: On Wed, 29 May 2019 14:26:51 +0200 Michael Mueller wrote: From: Halil Pasic To support protected virtualization cio will need to make sure the memory used for communication with the hypervisor is DMA memory. Let us introduce one global pool for cio. Our DMA pools are implemented as a gen_pool backed with DMA pages. The idea is to avoid each allocation effectively wasting a page, as we typically allocate much less than PAGE_SIZE. Signed-off-by: Halil Pasic Reviewed-by: Sebastian Ott Signed-off-by: Michael Mueller --- arch/s390/Kconfig | 1 + arch/s390/include/asm/cio.h | 11 drivers/s390/cio/css.c | 120 ++-- 3 files changed, 128 insertions(+), 4 deletions(-) (...) diff --git a/arch/s390/include/asm/cio.h b/arch/s390/include/asm/cio.h index 1727180e8ca1..43c007d2775a 100644 --- a/arch/s390/include/asm/cio.h +++ b/arch/s390/include/asm/cio.h @@ -328,6 +328,17 @@ static inline u8 pathmask_to_pos(u8 mask) void channel_subsystem_reinit(void); extern void css_schedule_reprobe(void); +extern void *cio_dma_zalloc(size_t size); +extern void cio_dma_free(void *cpu_addr, size_t size); +extern struct device *cio_get_dma_css_dev(void); + +struct gen_pool; That forward declaration is a bit ugly... I guess the alternative was include hell? That's an easy one. #include +#include #include +void *cio_gp_dma_zalloc(struct gen_pool *gp_dma, struct device *dma_dev, + size_t size); +void cio_gp_dma_free(struct gen_pool *gp_dma, void *cpu_addr, size_t size); +void cio_gp_dma_destroy(struct gen_pool *gp_dma, struct device *dma_dev); +struct gen_pool *cio_gp_dma_create(struct device *dma_dev, int nr_pages); + /* Function from drivers/s390/cio/chsc.c */ int chsc_sstpc(void *page, unsigned int op, u16 ctrl, u64 *clock_delta); int chsc_sstpi(void *page, void *result, size_t size); diff --git a/drivers/s390/cio/css.c b/drivers/s390/cio/css.c index aea502922646..b97618497848 100644 --- a/drivers/s390/cio/css.c +++ b/drivers/s390/cio/css.c @@ -20,6 +20,8 @@ #include #include #include +#include +#include #include #include @@ -224,6 +226,8 @@ struct subchannel *css_alloc_subchannel(struct subchannel_id schid, INIT_WORK(&sch->todo_work, css_sch_todo); sch->dev.release = &css_subchannel_release; device_initialize(&sch->dev); It might be helpful to add a comment why you use 31 bit here... @Halil, please let me know what comment you prefere here... + sch->dev.coherent_dma_mask = DMA_BIT_MASK(31); + sch->dev.dma_mask = &sch->dev.coherent_dma_mask; return sch; err: @@ -899,6 +903,8 @@ static int __init setup_css(int nr) dev_set_name(&css->device, "css%x", nr); css->device.groups = cssdev_attr_groups; css->device.release = channel_subsystem_release; ...and 64 bit here. and here. + css->device.coherent_dma_mask = DMA_BIT_MASK(64); + css->device.dma_mask = &css->device.coherent_dma_mask; mutex_init(&css->mutex); css->cssid = chsc_get_cssid(nr); (...) @@ -1059,16 +1168,19 @@ static int __init css_bus_init(void) if (ret) goto out_unregister; ret = register_pm_notifier(&css_power_notifier); - if (ret) { - unregister_reboot_notifier(&css_reboot_notifier); - goto out_unregister; - } + if (ret) + goto out_unregister_rn; + ret = cio_dma_pool_init(); + if (ret) + goto out_unregister_rn; Don't you also need to unregister the pm notifier on failure here? Mmh, that was the original intention. Thanks! Other than that, I noticed only cosmetic issues; seems reasonable to me. css_init_done = 1; /* Enable default isc for I/O subchannels. */ isc_register(IO_SCH_ISC); return 0; +out_unregister_rn: + unregister_reboot_notifier(&css_reboot_notifier); out_unregister: while (i-- > 0) { struct channel_subsystem *css = channel_subsystems[i]; Thanks, Michael ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Re: [PATCH v3 2/8] s390/cio: introduce DMA pools to cio
On Wed, 29 May 2019 14:26:51 +0200 Michael Mueller wrote: > From: Halil Pasic > > To support protected virtualization cio will need to make sure the > memory used for communication with the hypervisor is DMA memory. > > Let us introduce one global pool for cio. > > Our DMA pools are implemented as a gen_pool backed with DMA pages. The > idea is to avoid each allocation effectively wasting a page, as we > typically allocate much less than PAGE_SIZE. > > Signed-off-by: Halil Pasic > Reviewed-by: Sebastian Ott > Signed-off-by: Michael Mueller > --- > arch/s390/Kconfig | 1 + > arch/s390/include/asm/cio.h | 11 > drivers/s390/cio/css.c | 120 > ++-- > 3 files changed, 128 insertions(+), 4 deletions(-) (...) > diff --git a/arch/s390/include/asm/cio.h b/arch/s390/include/asm/cio.h > index 1727180e8ca1..43c007d2775a 100644 > --- a/arch/s390/include/asm/cio.h > +++ b/arch/s390/include/asm/cio.h > @@ -328,6 +328,17 @@ static inline u8 pathmask_to_pos(u8 mask) > void channel_subsystem_reinit(void); > extern void css_schedule_reprobe(void); > > +extern void *cio_dma_zalloc(size_t size); > +extern void cio_dma_free(void *cpu_addr, size_t size); > +extern struct device *cio_get_dma_css_dev(void); > + > +struct gen_pool; That forward declaration is a bit ugly... I guess the alternative was include hell? > +void *cio_gp_dma_zalloc(struct gen_pool *gp_dma, struct device *dma_dev, > + size_t size); > +void cio_gp_dma_free(struct gen_pool *gp_dma, void *cpu_addr, size_t size); > +void cio_gp_dma_destroy(struct gen_pool *gp_dma, struct device *dma_dev); > +struct gen_pool *cio_gp_dma_create(struct device *dma_dev, int nr_pages); > + > /* Function from drivers/s390/cio/chsc.c */ > int chsc_sstpc(void *page, unsigned int op, u16 ctrl, u64 *clock_delta); > int chsc_sstpi(void *page, void *result, size_t size); > diff --git a/drivers/s390/cio/css.c b/drivers/s390/cio/css.c > index aea502922646..b97618497848 100644 > --- a/drivers/s390/cio/css.c > +++ b/drivers/s390/cio/css.c > @@ -20,6 +20,8 @@ > #include > #include > #include > +#include > +#include > #include > #include > > @@ -224,6 +226,8 @@ struct subchannel *css_alloc_subchannel(struct > subchannel_id schid, > INIT_WORK(&sch->todo_work, css_sch_todo); > sch->dev.release = &css_subchannel_release; > device_initialize(&sch->dev); It might be helpful to add a comment why you use 31 bit here... > + sch->dev.coherent_dma_mask = DMA_BIT_MASK(31); > + sch->dev.dma_mask = &sch->dev.coherent_dma_mask; > return sch; > > err: > @@ -899,6 +903,8 @@ static int __init setup_css(int nr) > dev_set_name(&css->device, "css%x", nr); > css->device.groups = cssdev_attr_groups; > css->device.release = channel_subsystem_release; ...and 64 bit here. > + css->device.coherent_dma_mask = DMA_BIT_MASK(64); > + css->device.dma_mask = &css->device.coherent_dma_mask; > > mutex_init(&css->mutex); > css->cssid = chsc_get_cssid(nr); (...) > @@ -1059,16 +1168,19 @@ static int __init css_bus_init(void) > if (ret) > goto out_unregister; > ret = register_pm_notifier(&css_power_notifier); > - if (ret) { > - unregister_reboot_notifier(&css_reboot_notifier); > - goto out_unregister; > - } > + if (ret) > + goto out_unregister_rn; > + ret = cio_dma_pool_init(); > + if (ret) > + goto out_unregister_rn; Don't you also need to unregister the pm notifier on failure here? Other than that, I noticed only cosmetic issues; seems reasonable to me. > css_init_done = 1; > > /* Enable default isc for I/O subchannels. */ > isc_register(IO_SCH_ISC); > > return 0; > +out_unregister_rn: > + unregister_reboot_notifier(&css_reboot_notifier); > out_unregister: > while (i-- > 0) { > struct channel_subsystem *css = channel_subsystems[i]; ___ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization
[PATCH v3 2/8] s390/cio: introduce DMA pools to cio
From: Halil Pasic To support protected virtualization cio will need to make sure the memory used for communication with the hypervisor is DMA memory. Let us introduce one global pool for cio. Our DMA pools are implemented as a gen_pool backed with DMA pages. The idea is to avoid each allocation effectively wasting a page, as we typically allocate much less than PAGE_SIZE. Signed-off-by: Halil Pasic Reviewed-by: Sebastian Ott Signed-off-by: Michael Mueller --- arch/s390/Kconfig | 1 + arch/s390/include/asm/cio.h | 11 drivers/s390/cio/css.c | 120 ++-- 3 files changed, 128 insertions(+), 4 deletions(-) diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig index 88d8355b7bf7..2a245b56db8b 100644 --- a/arch/s390/Kconfig +++ b/arch/s390/Kconfig @@ -191,6 +191,7 @@ config S390 select ARCH_HAS_SCALED_CPUTIME select HAVE_NMI select SWIOTLB + select GENERIC_ALLOCATOR config SCHED_OMIT_FRAME_POINTER diff --git a/arch/s390/include/asm/cio.h b/arch/s390/include/asm/cio.h index 1727180e8ca1..43c007d2775a 100644 --- a/arch/s390/include/asm/cio.h +++ b/arch/s390/include/asm/cio.h @@ -328,6 +328,17 @@ static inline u8 pathmask_to_pos(u8 mask) void channel_subsystem_reinit(void); extern void css_schedule_reprobe(void); +extern void *cio_dma_zalloc(size_t size); +extern void cio_dma_free(void *cpu_addr, size_t size); +extern struct device *cio_get_dma_css_dev(void); + +struct gen_pool; +void *cio_gp_dma_zalloc(struct gen_pool *gp_dma, struct device *dma_dev, + size_t size); +void cio_gp_dma_free(struct gen_pool *gp_dma, void *cpu_addr, size_t size); +void cio_gp_dma_destroy(struct gen_pool *gp_dma, struct device *dma_dev); +struct gen_pool *cio_gp_dma_create(struct device *dma_dev, int nr_pages); + /* Function from drivers/s390/cio/chsc.c */ int chsc_sstpc(void *page, unsigned int op, u16 ctrl, u64 *clock_delta); int chsc_sstpi(void *page, void *result, size_t size); diff --git a/drivers/s390/cio/css.c b/drivers/s390/cio/css.c index aea502922646..b97618497848 100644 --- a/drivers/s390/cio/css.c +++ b/drivers/s390/cio/css.c @@ -20,6 +20,8 @@ #include #include #include +#include +#include #include #include @@ -224,6 +226,8 @@ struct subchannel *css_alloc_subchannel(struct subchannel_id schid, INIT_WORK(&sch->todo_work, css_sch_todo); sch->dev.release = &css_subchannel_release; device_initialize(&sch->dev); + sch->dev.coherent_dma_mask = DMA_BIT_MASK(31); + sch->dev.dma_mask = &sch->dev.coherent_dma_mask; return sch; err: @@ -899,6 +903,8 @@ static int __init setup_css(int nr) dev_set_name(&css->device, "css%x", nr); css->device.groups = cssdev_attr_groups; css->device.release = channel_subsystem_release; + css->device.coherent_dma_mask = DMA_BIT_MASK(64); + css->device.dma_mask = &css->device.coherent_dma_mask; mutex_init(&css->mutex); css->cssid = chsc_get_cssid(nr); @@ -1018,6 +1024,109 @@ static struct notifier_block css_power_notifier = { .notifier_call = css_power_event, }; +#define CIO_DMA_GFP (GFP_KERNEL | __GFP_ZERO) +static struct gen_pool *cio_dma_pool; + +/* Currently cio supports only a single css */ +struct device *cio_get_dma_css_dev(void) +{ + return &channel_subsystems[0]->device; +} + +struct gen_pool *cio_gp_dma_create(struct device *dma_dev, int nr_pages) +{ + struct gen_pool *gp_dma; + void *cpu_addr; + dma_addr_t dma_addr; + int i; + + gp_dma = gen_pool_create(3, -1); + if (!gp_dma) + return NULL; + for (i = 0; i < nr_pages; ++i) { + cpu_addr = dma_alloc_coherent(dma_dev, PAGE_SIZE, &dma_addr, + CIO_DMA_GFP); + if (!cpu_addr) + return gp_dma; + gen_pool_add_virt(gp_dma, (unsigned long) cpu_addr, + dma_addr, PAGE_SIZE, -1); + } + return gp_dma; +} + +static void __gp_dma_free_dma(struct gen_pool *pool, + struct gen_pool_chunk *chunk, void *data) +{ + size_t chunk_size = chunk->end_addr - chunk->start_addr + 1; + + dma_free_coherent((struct device *) data, chunk_size, +(void *) chunk->start_addr, +(dma_addr_t) chunk->phys_addr); +} + +void cio_gp_dma_destroy(struct gen_pool *gp_dma, struct device *dma_dev) +{ + if (!gp_dma) + return; + /* this is qite ugly but no better idea */ + gen_pool_for_each_chunk(gp_dma, __gp_dma_free_dma, dma_dev); + gen_pool_destroy(gp_dma); +} + +static int cio_dma_pool_init(void) +{ + /* No need to free up the resources: compiled in */ + cio_dma_pool = cio_gp_dma_create(cio_get_dma_css_dev(), 1); + if (!cio_dma_pool) + return -ENOMEM; + return 0; +} +