From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To: Umang Jain <umang.jain@ideasonboard.com>
Cc: linux-staging@lists.linux.dev,
Stefan Wahren <stefan.wahren@i2se.com>,
Dan Carpenter <error27@gmail.com>,
Kieran Bingham <kieran.bingham@ideasonboard.com>,
Phil Elwell <phil@raspberrypi.com>,
Dave Stevenson <dave.stevenson@raspberrypi.com>,
Greg KH <greg@kroah.com>
Subject: Re: [PATCH v3 5/6] staging: vc04_services: Drop global variables tracking allocated pages
Date: Fri, 22 Mar 2024 13:44:42 +0200 [thread overview]
Message-ID: <20240322114442.GP18799@pendragon.ideasonboard.com> (raw)
In-Reply-To: <372d242d-1db5-4451-b595-c47b18368656@ideasonboard.com>
On Fri, Mar 22, 2024 at 02:45:11PM +0530, Umang Jain wrote:
> On 22/03/24 1:20 am, Laurent Pinchart wrote:
> > On Thu, Mar 21, 2024 at 04:07:39PM +0530, Umang Jain wrote:
> >> The variables tracking allocated pages fragments in vchiq_arm.c
> >> can be easily moved to struct vchiq_drv_mgmt instead of being global.
> >> This helps us to drop the non-essential global variables in the vchiq
> >> interface.
> >>
> >> However, g_regs is intentionally kept global since it is read/write
> >> in fast paths (for e.g. doorbell IRQs). Introducing it in struct
> >> vchiq_drv_mgmt and getting it in remote_event_signal handler, turns
> >> out to be quite tedious. Add a comment for the same.
> > The fast path argument doesn't really hold in my opinion, I think you
> > can retrieve the variable through vchiq_state in vchiq_doorbell_irq().
> >
> > All callers for remote_event_signal have a vchiq_state pointer, so it
> > shouldn't be that difficult to get rid of that global variable. It
> > doesn't have to be done in this patch, it can be done on top.
> >
> >> Drop the TODO item as well, as it is now totally addressed:
> >>> Get rid of all non essential global structures and create a proper per
> >> device structure
> >
> > Don't you need to also drop the g_state variable ?
>
> Seems essential to me looking at vchiq_get_state() helper
>
> Especially in the vchiq_open() to get a starting state when the char
> device is opened..
Nobody said it would be a 5 minutes job :-) There seems to be an endless
amount of refactoring that vchiq could benefit from. It doesn't have to
be all done in one go of course.
> >> No functional changes intended in this patch.
> >>
> >> Signed-off-by: Umang Jain <umang.jain@ideasonboard.com>
> >> ---
> >> drivers/staging/vc04_services/interface/TODO | 8 ---
> >> .../interface/vchiq_arm/vchiq_arm.c | 57 +++++++++----------
> >> .../interface/vchiq_arm/vchiq_arm.h | 6 ++
> >> 3 files changed, 34 insertions(+), 37 deletions(-)
> >>
> >> diff --git a/drivers/staging/vc04_services/interface/TODO b/drivers/staging/vc04_services/interface/TODO
> >> index 05eb5140d096..15f12b8f213e 100644
> >> --- a/drivers/staging/vc04_services/interface/TODO
> >> +++ b/drivers/staging/vc04_services/interface/TODO
> >> @@ -41,14 +41,6 @@ The code follows the 80 characters limitation yet tends to go 3 or 4 levels of
> >> indentation deep making it very unpleasant to read. This is specially relevant
> >> in the character driver ioctl code and in the core thread functions.
> >>
> >> -* Get rid of all non essential global structures and create a proper per
> >> -device structure
> >> -
> >> -The first thing one generally sees in a probe function is a memory allocation
> >> -for all the device specific data. This structure is then passed all over the
> >> -driver. This is good practice since it makes the driver work regardless of the
> >> -number of devices probed.
> >> -
> >> * Clean up Sparse warnings from __user annotations. See
> >> vchiq_irq_queue_bulk_tx_rx(). Ensure that the address of "&waiter->bulk_waiter"
> >> is never disclosed to userspace.
> >> diff --git a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c
> >> index ba096bcb32c8..c4807b302718 100644
> >> --- a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c
> >> +++ b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c
> >> @@ -140,13 +140,11 @@ struct vchiq_pagelist_info {
> >> unsigned int scatterlist_mapped;
> >> };
> >>
> >> +/*
> >> + * Essential global pointer since read/written on fast paths
> >> + * (doorbell IRQs).
> >> + */
> >> static void __iomem *g_regs;
> >> -static unsigned int g_fragments_size;
> >> -static char *g_fragments_base;
> >> -static char *g_free_fragments;
> >> -static struct semaphore g_free_fragments_sema;
> >> -
> >> -static DEFINE_SEMAPHORE(g_free_fragments_mutex, 1);
> >
> > Why is this code using a semaphore as a mutex ?? It should be replaced
> > with a real mutex. This can be done on top of this patch.
> >
> >>
> >> static int
> >> vchiq_blocking_bulk_transfer(struct vchiq_instance *instance, unsigned int handle, void *data,
> >> @@ -377,20 +375,20 @@ create_pagelist(struct vchiq_instance *instance, char *buf, char __user *ubuf,
> >> (drv_mgmt->pinfo->cache_line_size - 1)))) {
> >> char *fragments;
> >>
> >> - if (down_interruptible(&g_free_fragments_sema)) {
> >> + if (down_interruptible(&drv_mgmt->free_fragments_sema)) {
> >> cleanup_pagelistinfo(instance, pagelistinfo);
> >> return NULL;
> >> }
> >>
> >> - WARN_ON(!g_free_fragments);
> >> + WARN_ON(!drv_mgmt->free_fragments);
> >>
> >> - down(&g_free_fragments_mutex);
> >> - fragments = g_free_fragments;
> >> + down(&drv_mgmt->free_fragments_mutex);
> >> + fragments = drv_mgmt->free_fragments;
> >> WARN_ON(!fragments);
> >> - g_free_fragments = *(char **)g_free_fragments;
> >> - up(&g_free_fragments_mutex);
> >> + drv_mgmt->free_fragments = *(char **)drv_mgmt->free_fragments;
> >> + up(&drv_mgmt->free_fragments_mutex);
> >> pagelist->type = PAGELIST_READ_WITH_FRAGMENTS +
> >> - (fragments - g_fragments_base) / g_fragments_size;
> >> + (fragments - drv_mgmt->fragments_base) / drv_mgmt->fragments_size;
> >> }
> >>
> >> return pagelistinfo;
> >> @@ -420,10 +418,10 @@ free_pagelist(struct vchiq_instance *instance, struct vchiq_pagelist_info *pagel
> >> pagelistinfo->scatterlist_mapped = 0;
> >>
> >> /* Deal with any partial cache lines (fragments) */
> >> - if (pagelist->type >= PAGELIST_READ_WITH_FRAGMENTS && g_fragments_base) {
> >> - char *fragments = g_fragments_base +
> >> + if (pagelist->type >= PAGELIST_READ_WITH_FRAGMENTS && drv_mgmt->fragments_base) {
> >> + char *fragments = drv_mgmt->fragments_base +
> >> (pagelist->type - PAGELIST_READ_WITH_FRAGMENTS) *
> >> - g_fragments_size;
> >> + drv_mgmt->fragments_size;
> >> int head_bytes, tail_bytes;
> >>
> >> head_bytes = (drv_mgmt->pinfo->cache_line_size - pagelist->offset) &
> >> @@ -448,11 +446,11 @@ free_pagelist(struct vchiq_instance *instance, struct vchiq_pagelist_info *pagel
> >> fragments + drv_mgmt->pinfo->cache_line_size,
> >> tail_bytes);
> >>
> >> - down(&g_free_fragments_mutex);
> >> - *(char **)fragments = g_free_fragments;
> >> - g_free_fragments = fragments;
> >> - up(&g_free_fragments_mutex);
> >> - up(&g_free_fragments_sema);
> >> + down(&drv_mgmt->free_fragments_mutex);
> >> + *(char **)fragments = drv_mgmt->free_fragments;
> >> + drv_mgmt->free_fragments = fragments;
> >> + up(&drv_mgmt->free_fragments_mutex);
> >> + up(&drv_mgmt->free_fragments_sema);
> >> }
> >>
> >> /* Need to mark all the pages dirty. */
> >> @@ -488,11 +486,11 @@ static int vchiq_platform_init(struct platform_device *pdev, struct vchiq_state
> >> if (err < 0)
> >> return err;
> >>
> >> - g_fragments_size = 2 * drv_mgmt->pinfo->cache_line_size;
> >> + drv_mgmt->fragments_size = 2 * drv_mgmt->pinfo->cache_line_size;
> >>
> >> /* Allocate space for the channels in coherent memory */
> >> slot_mem_size = PAGE_ALIGN(TOTAL_SLOTS * VCHIQ_SLOT_SIZE);
> >> - frag_mem_size = PAGE_ALIGN(g_fragments_size * MAX_FRAGMENTS);
> >> + frag_mem_size = PAGE_ALIGN(drv_mgmt->fragments_size * MAX_FRAGMENTS);
> >>
> >> slot_mem = dmam_alloc_coherent(dev, slot_mem_size + frag_mem_size,
> >> &slot_phys, GFP_KERNEL);
> >> @@ -512,15 +510,16 @@ static int vchiq_platform_init(struct platform_device *pdev, struct vchiq_state
> >> vchiq_slot_zero->platform_data[VCHIQ_PLATFORM_FRAGMENTS_COUNT_IDX] =
> >> MAX_FRAGMENTS;
> >>
> >> - g_fragments_base = (char *)slot_mem + slot_mem_size;
> >> + drv_mgmt->fragments_base = (char *)slot_mem + slot_mem_size;
> >>
> >> - g_free_fragments = g_fragments_base;
> >> + drv_mgmt->free_fragments = drv_mgmt->fragments_base;
> >> for (i = 0; i < (MAX_FRAGMENTS - 1); i++) {
> >> - *(char **)&g_fragments_base[i * g_fragments_size] =
> >> - &g_fragments_base[(i + 1) * g_fragments_size];
> >> + *(char **)&drv_mgmt->fragments_base[i * drv_mgmt->fragments_size] =
> >> + &drv_mgmt->fragments_base[(i + 1) * drv_mgmt->fragments_size];
> >> }
> >> - *(char **)&g_fragments_base[i * g_fragments_size] = NULL;
> >> - sema_init(&g_free_fragments_sema, MAX_FRAGMENTS);
> >> + *(char **)&drv_mgmt->fragments_base[i * drv_mgmt->fragments_size] = NULL;
> >> + sema_init(&drv_mgmt->free_fragments_sema, MAX_FRAGMENTS);
> >> + sema_init(&drv_mgmt->free_fragments_mutex, 1);
> >>
> >> err = vchiq_init_state(state, vchiq_slot_zero, dev);
> >> if (err)
> >> diff --git a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.h b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.h
> >> index 1190fab2efc4..1134187da6ab 100644
> >> --- a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.h
> >> +++ b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.h
> >> @@ -41,6 +41,12 @@ struct vchiq_drv_mgmt {
> >> bool connected;
> >> int num_deferred_callbacks;
> >> void (*deferred_callback[VCHIQ_DRV_MAX_CALLBACKS])(void);
> >> +
> >> + struct semaphore free_fragments_sema;
> >> + struct semaphore free_fragments_mutex;
> >> + char *fragments_base;
> >> + char *free_fragments;
> >> + unsigned int fragments_size;
> >> };
> >>
> >> struct user_service {
--
Regards,
Laurent Pinchart
next prev parent reply other threads:[~2024-03-22 11:44 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-03-21 10:37 [PATCH v3 0/6] staging: vc04_services: Drop non-essential global members Umang Jain
2024-03-21 10:37 ` [PATCH v3 1/6] staging: vc04_services: Drop g_once_init global variable Umang Jain
2024-03-21 10:37 ` [PATCH v3 2/6] staging: vc04_services: vchiq_arm: Split driver static and runtime data Umang Jain
2024-03-21 18:46 ` Laurent Pinchart
2024-03-21 10:37 ` [PATCH v3 3/6] staging: vc04_services: vchiq_arm: Drop g_cache_line_size Umang Jain
2024-03-21 19:04 ` Laurent Pinchart
2024-03-21 19:10 ` Laurent Pinchart
2024-03-21 10:37 ` [PATCH v3 4/6] staging: vc04_services: Drop global members for tracking connections Umang Jain
2024-03-21 19:12 ` Laurent Pinchart
2024-03-22 4:55 ` Umang Jain
2024-03-22 8:17 ` Laurent Pinchart
2024-03-21 10:37 ` [PATCH v3 5/6] staging: vc04_services: Drop global variables tracking allocated pages Umang Jain
2024-03-21 19:50 ` Laurent Pinchart
2024-03-22 9:15 ` Umang Jain
2024-03-22 11:44 ` Laurent Pinchart [this message]
2024-03-23 13:58 ` Umang Jain
2024-03-21 10:37 ` [PATCH v3 6/6] staging: vc04_services: Drop completed TODO item Umang Jain
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240322114442.GP18799@pendragon.ideasonboard.com \
--to=laurent.pinchart@ideasonboard.com \
--cc=dave.stevenson@raspberrypi.com \
--cc=error27@gmail.com \
--cc=greg@kroah.com \
--cc=kieran.bingham@ideasonboard.com \
--cc=linux-staging@lists.linux.dev \
--cc=phil@raspberrypi.com \
--cc=stefan.wahren@i2se.com \
--cc=umang.jain@ideasonboard.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox